We strive to be thought leaders in the CPG tech industry, and present monthly 20-30 minute webinars sharing our insights.

Threats to AI Development - Webinar

Observa – Threats to AI Development

Observa Title

Hello and welcome to Observa’s webinar on threats to AI. I first wanted to just kind of let you know who I am.

Your Speaker

My name is Erik Chelstad and I’m a CTO and co-founder of Observa. I’ve been working in AI and technology for over 20 years. I think one of the highlights I like to think of at least that’s a highlight is that I wrote a lot of smart algorithms, they’re embedded on almost every plane in the US that has more than 13 seats and they are intended to keep aircraft from colliding with each other, also to make slight course adjustments when flying internationally over large bodies of water in order to avoid collisions but also save fuel.


So normally we talk a lot about Observa, kind of go into some things here. But basically, you know Observa, we do some things you know they are because you’ve signed up for this and I want to say that we use AI for image recognition. We also use it for a retail recommendation engine. But we’re really not here to talk about Observa per say today. We’re going to talk more about AI in general. We’re going to do some fun technical going to cover some fun technical ground in the next twelve minutes.

AI is Threatening Humans

So, I want to set the stage here with what’s going on with AI. We’ve been seeing things in the news. If you listen to people like Elon Musk you’ll know that he signed a petition along with a couple hundred other organizations and over three thousand individuals such as the lead of Google’s AI, the founders of Deep Mind, so some of the big power names in AI and they’ve signed this pledge to prevent killer robots that they won’t work on it and they won’t do anything. So, we’re talking about killer robots.

The late Stephen Hawking said, “Must we learn how to prepare for and avoid the potential risks, AI could be the worst event in the history of our civilization.” Google came out with a product called Duplex that simulates a human calling the person. And when that demo came out a few months ago it was so freaky to everybody that California immediately made a new law that goes into effect next year that now robots or bots have to announce themselves when they call. So that was really fast. Think about the number of recorded calls you get versus the bot calls you get. So, robots and AI are replacing all the workers is another fear. Bill Gates wants to tax robots and slow down automation. So, what we’re really looking at here is there’s a lot of kind of fear being talked about in the press with regard to AI.

Hope and Hype

So why is there so much fear? To me there’s a lot of possibility. Our imaginations are kind of running wild right now with all the potential things that we could do. I like to bring up the Gartner hype cycle that you can see on the screen. And if you look at that – and my eyes are going to stray off the cameras as I actually look at this – because there are things like machine learning and autonomous vehicles right up at the top of this peak of inflated expectations. So, what we’re seeing is that we’re very enchanted with this as humans and businesses right now.

What’s Causing Human Resistance?

So why would humans stall AI? So again, we’ve got the stage set where we’re very excited about it. We’re very fearful about it and what’s going to stop it from becoming the really good things that we want it to become? So that’s kind of the idea that we’re going to be talking about.

Moral and Ethical Dilemmas

So, I was going to start right off with moral and ethical dilemmas. We have a lot going on here with regards to AI. So, the big question is can we trust AI to act like a human? And this isn’t the bad humans, this is the human when you say act like a nice person, treat someone like you want to be treated – those types of things that we all have in our collective memories. So, I think an example of a lemonade stand and since I had these cute little kids behind bars, think about a lemonade stand. If you were to prosecute the law by the letter of the law a lemonade stand is definitely illegal. They do not have a health permit. They do not have a business license. They’re probably not paying their taxes on it. They’re probably violating some work laws like hiring themselves under the age of 15 or whatever it is in your state. But if we just left this up to a computer algorithm, how would the computer know to have compassion and treat these kids that they’re having a little fun thing they’re trying to do that they’re not really creating this giant business? So, a human wouldn’t know but a beat cop would. I think of the kind of the inability to deal with the unexpected.

A really visceral example of this is the self-driving car idea. And what happens when that self-driving car goes into town and there’s a different colored fire truck. So, it’s what does it what is it supposed to do? I think another really great example of kind of the moral and ethical thing is you look at the graph on the screen, that is a graph of a study that shows what percentage of people would save their dog’s life over another human life. And this is actually broken down by men and women. I thought that was interesting just because kind of goes to show that all humans, even from the same societies, might have a little bit of a different view on this. So almost half of women surveyed would save their dog before they saved a foreign tourist. Now how are we going to train and codify that into a computer? These are some of the more ethical dilemmas that we’re going to be facing. And again, if we decide that it’s just too complicated make that decision, then we’re going to shut down development on AI instead of continuing.

Lack of Understanding.

I think another big area is just a lack of understanding. By that I mean we don’t understand what’s going on. Again, to go with a really visceral example would you want a robot operating on your unborn child? And I think of that because it’s a very vulnerable segment of society. So just kind of that thought experiment, I’m not going down the whole Blade Runner, path but the idea of what does this robot doing, what regulations did it pass, where’s the software tested, is it used in the latest version of software? We kind of understand when it’s a doctor that we can say yes, this doctor went to Harvard and they did their apprenticeship in Chicago but what did the AI have to do. So, are we going to be able to understand that and feel comfortable with whatever it is that the AI’s has been tested to do?

Tech Obstacles – We Aren’t There Yet.

Now we just go to kind of tech obstacles. So, there are technical obstacles that are there. We’re not necessarily as far along as we think we are with AI as in there are lots and lots of hard problems to solve. We’re starting to bump up against the limits of some of the CPU’s use. We’ve all started using graphical processing units instead of computational processing units, so GPU’s, so we’ve switched over in the last couple of years. The storage of all the data that we need to collect and train things on we’re starting to hit limits on things. And so, we’re just going to be bumping up against things and developing new things because of it. But the concern here is that if tech isn’t keeping up with the dreams will the popular media and people get disillusioned by how we’re not moving fast enough in AI and so we’ll stop getting engineers excited and people will stop moving into these areas. That would be a concern. It could slow down and be a threat to AI.

Sorry, I Can’t Do That Dave

I think a really good example is you know we talk about can a self-driving car recognize a yellow fire truck versus a red fire truck. I think last week Tesla announced that their auto pilot will soon be able to recognize emergency vehicles. I think that probably comes as a shock to a lot of people that it didn’t already. We now have these vehicles out driving on the road. There’s hundreds of them out there if not thousands doing autonomous driving. And now the autonomous driving is in very limited circumstances. You’re only supposed to use it on a freeway and there’s other rules, but you’re never supposed to take your hands off the wheel that kind of thing. But I think most people will be shocked to know that if a fire truck or an ambulance pulled up beside your self-driving car right now it wouldn’t know what to do, it wouldn’t do anything different. So again, this is that set of expectations where we are aren’t as far along as we think we are.

And just to that point still situational awareness as well as saying it apparently is difficult but so is actual situational awareness. This is a pretty advanced thing that computers are not ready for. We have years and years of experience. By the time you’re able to drive you have a lot of experience being in a car. So, this is where the human part is way better than what we have right now. Again, we’re developing fast, but I mean think about how a car might react to a sheet of plywood flying out of a truck in front of it. Has it been trained on that situation? Not necessarily. Probably not. Does it feel fear? Probably not. So, we’re not sure how it might react. It might think it’s a leaf and just drive right through it.

Black Swan

So, talking about these kind of like what if situations like that bring up the idea of a Black Swan. This has to do more with a book about unexpected events being expected as opposed to killer ballerinas. I just want to be clear on that. So, I think of the black swan events that are going to happen – so if we go back that self-driving car what if that self-driving car crashes into a school bus? What if your personal stock market mutual fund retirement management AI loses money while the stock market is gaining. These are things that are going to erode your trust and our public trust in the AI systems. These are pretty much inevitable events. But how do we react, and do we again say OK I’ve had it with AI. Done and cut it off. Or do we go in with more of a managed set of expectations?

Inflated Expectations

Kind of looking at those inflated expectations we have our personal side. We have the public side but then there’s investors. The investors are putting a lot of money right now into this. It’s where a lot, I think, the majority of venture capital is going in the US right now is in AI startups and things that are fueled by AI, but there is a lot of really hard work that still has to happen as we’ve been talking about. So are investors thinking that they’re going to get a really quick return and then if they don’t they’re going to be unhappy about it, they’re going to start pulling their funding from this technology and start putting it someplace else so the money’s going to dry out. This is a potential situation that could happen – another threat to AI.

Religious – Building Gods

I think another one with regards to human or humanity as something I’m going to go and ahead and talk about a couple of things now that we shouldn’t really talk about at work, right? This is going to be religion and politics, but there are potential we’re going to bump into some religious objections with AI about as far as how we’re trying to build lifeforms or turn ourselves into beings that are too godlike. And I think especially this may come into play when some – as I was talking about we’re going to be bumping up against the computational and storage limits of traditional computers – we may find that it makes sense to go in to sort of biological computation which is going to have us manipulating DNA and creating different types of life forms. And so, we may be bumping up against some other ethical and religious issues there. So, we have to be careful not to tread too heavily and just to keep people aware and educated about what’s going on with the technology.

Politics – We Are All the Same

Politically I think this is going to become a very politicized technology. The governments are going to want to keep a lot of the tech internal because it can be used for both offensive and defensive capabilities. I think we’ve seen that in the last couple of years with China saying that they were going to keep all of their AI for themselves and become a superpower of AI and then recently announcing that they were going to be more open about sharing. In the 90s the US had a pretty hard embargo – I think it’s a lot of it’s still there – on cryptography. That was a technology that came out and enabled people to have secure communications with each other. But it was seen, I think the phrase was Weapons Grade Cryptography – these are the types of things that governments are going to want to control because it has a massive impact on their societies.

Politics – We Are All Different

So that’s kind of how I think governments are the same. And then when you look at how governments can be very different we look at – I think the EU right now is leading the charge with privacy laws on the digital scene. And we look at something that is being developed perhaps in AI that’s looking at some people’s proclivity to steal something. China might be developing that, but the ability for a governmental or quasi-governmental agency to collect information about an individual in China is quite different than that same agency in the EU. So, I think this is where we are going to find some differences. There’s also going to be risk tolerance where some cultures may be much more open to trying things out and not being upset when some of those black swans happen.

Intense Competition

Finally, I think another thing with regards to large organizations is companies and the intense competition that’s happening with AI and with what’s going on. So, the companies that that own the current state have invested very very heavily in this. IBM and Google and Microsoft – they’ve spent a lot of money getting their AI to where it is and they’re going to try to prevent other companies from taking their dominant positions away from them. And so, they have lots of tools to do this whether they’re just buying the companies and integrating it into their ecosystem or perhaps working with the government to create and regulations. For instance, a licensing situation where it costs a million dollars to get a self-driving car AI license. It totally makes sense for it to go through rigorous testing but only the large companies will be able to afford that potentially. So that is something else that could keep this technology in its nascent form.

Malicious Circumventing

And I think really finally I want to talk about this potential for this kind of maliciousness. This children’s drawing of a stop sign with go painted over it is an example where the computer we can create things that the computer will misinterpret. That’s known as adversarial networks or it’s basically the idea here is that you look at a stop sign, and you see it as a stop. If somebody was to suddenly paint the word go in invisible ink basically on that stop sign you wouldn’t even notice it, but a computer might. Depends on what kind of camera. So, we could fool a computer into making a stop sign it would think a stop sign was a draft and just drive right through it.

Another example of this kind of headline, you know about Google rewriting its search engine rankings algorithms this is something that’s constantly happening it’s kind of a battle of arms race of sorts where Google is trying to prevent fake news from happening but the people that want to get the fake news out are constantly battling and rewriting their code to go around it. So, people will be trying to take advantage of AI pretty consistently whether it’s to break into your bank account or steal your self-driving car or whatever it is they might be doing.

Moving Forward

So, kind of moving forward, I think it’s important that we understand as people that are interested in this as was practitioners that there’s many pitfalls ahead of us. There’s both real pitfalls and imagined ones. We really need to deal with those and we need to understand that innovation has issues along the way. It’s never going to be a perfect path. There’s going to be bad things as well as unexpected good things. But we need to be realistic. We can’t let our bright-eyed optimism allow withholding common sense or being overly disappointed.


I think a really powerful quote here is from Bill Gates, I mentioned him at the beginning about wanting to tax robots and such. But in this case, he’s saying be careful about what you feel about these incoming technologies because we’ve got this thing that is really relevant right now when we look at self-driving cars. You know everyone’s saying that’s going to happen tomorrow. But I think a longer horizon of a decade is a little bit more likely. Same thing, I don’t know how many years ago it was that Amazon promised they would have drone deliveries by 2018. And even though they have some test ones that worked they haven’t put them in full swing production because they ran into other barriers.

So just being aware that technology for us on a professional level stay educated do not over commit your resources to these new technologies that AI is coming and also keep an eye on it, treat it like a new blade and make sure that you’re not committing all of your trust to something right away. Personally, I think make sure that before you promise anything or before you implement this really test things and vet them thoroughly because it can be personally damaging if you just suddenly throw all your money into an AI trading fund for your retirement. Just be aware that again like anything test it. Test the waters and make sure you understand what’s happening there. So that’s all I want to present today about AI, the world of AI, and threats to AI. So, let’s keep us going. And enjoy the rest of 2018. We’ll see in 2019. Personally, and my business partner who will also be back next week to talk more about retail. So, thanks again and have a great rest of the week.

Observa x Ricoh Webinar: What Robots Mean For the Future of Work - Webinar Transcript

Observa Ricoh Webinar

Erik: Hello and good morning or good afternoon depending on where you are. And welcome to the webinar presented by Observa and Ricoh about what robots mean for the future of work. There are two of us presenting today.

Your Speakers

And the first person I guess I will say since I’m talking is myself. I am Erik Chelstad. I’m the CTO and Co-Founder of Observa. I am by background an electrical and computer engineer. So, that makes me fairly nerdy. I’ve spent a lot of my career doing control systems and that involves both people and things like airplanes and other vehicles. The other presenter is Denise Chan and I will let her introduce herself.

Denise: Hi everyone. Good morning. It’s great to be here today. I’m really excited about this topic. I am the product marketing manager at Ricoh Innovations and a full staff market with experience mainly working with B2B, SaaS companies selling to both the developer and marketing space.

Who we are

Erik: Alright. And so to briefly to talk about our two companies, Denise if you’d like to continue with Ricoh.

Denise: Yeah. So, Ricoh Innovations – we are the research and venture center and open innovation center and we are a part of Ricoh Company Limited, and as part of Ricoh Company Limited we are really focused on creating new and also researching new emerging technology lines for Ricoh Company, and these involve anything from computer vision to IoT, to edge computing and machine learning technologies. And that’s why we’ve dove a lot into AI technologies and how we’ve used that to improve on and really optimize the workplace, both in the office and outside of the office. And we’re really excited to talk a little bit more about some of those learnings today.

Erik: Thanks Denise. And as Erik at Observa, we go out and we provide insights and real time actions for retail and brands involved in retail. Basically, we utilize technology, both human-enabling technology and AI. And so, that’s briefly it. And I’d like to go right into what our topic is.

Types of Robots

So, we’re talking about today basically robots and the future of work. And a lot of this, I think it’s important to look at the different types of robots that we are often talking about now. So, kind of broken up into the idea of an abstract robot or a physical robot. And this is just a way of looking at this. The abstract robots are the ones that – I think one of the things you’ve been hearing about is Google Duplex kind of made a splash a couple of months ago when people were unable to determine that it was a robot calling them. So, this is one of those things that might change the way we interact with robots given that phones are very prevalent, and they are part of our culture and robots might start calling you and you won’t be able to know. The other types of these robots, they are things that you might already have in your house, whether it’s Alexa or Google Home. And then of course, we’re familiar with a type of bot, whether a chat bot’s helping us online or getting into our social media and perceiving people there. And then of course the physical robots that again some of these may be in your home, such as a Roomba, and then other are more industrial and part of things such as auto factories or I think a great example is the autopilot on an airplane. It’s a very prevalent thing that is used quite a bit.

Gartner Hype Cycle for Emerging Technologies, 2018

Kind of where are we with this robots, again some of these things you already interact with in your life whether it’s taking an airplane flight or having your home cleaned. And I think it’s just interesting to look at something like Gartner’s Hype Cycle where we are looking at what’s happening with emerging technologies right now. And on this scale of things at the very top of this peak this year you’re seeing things like smart robots, virtual assistants and a ton of small mobile robots. Those are kind of at the peak inflated expectations meaning that for most of these it’s five to ten years before they are really a part of our mainstream lives. And so, some of these you’re getting the early adopters going out and doing things with virtual assistants and smart robots – smart robots being the ones that can interact with us and work side by side with humans much better.

Types of Work

So, that’s kind of it with the robots, but what types of work are available? And what’s being done? And so, I’d just thought that I’d break it up into a few pieces here. So, it’s kind of like the idea of hired work. And this is what we think of as our normal day-to-day full-time jobs and then the paid – not that hired isn’t paid – but you’re just paid for your work. So more like a piecework. But this is what we think of when we often think of the gig economy or sort of a task-based economy. The difference there being that the gig economy is something that you’re doing that’s going to take a while, you’re going to have to do something, perhaps bring some skills into it, and the task-based is something that is going to be very simple like looking at a picture on your phone and classifying or circling a cat or a face. So, when you look at this

Attributes of Work

So, when you look at this you kind of see the attributes of the work. And so, I start at the top and if you look at the very right of this box you’ll see that what is the payout here. It think it’s sorted by that. Where if you’re doing hired job, if you have a salary or if you’re doing contract work like painting people’s houses, your payout is going to be high. This is what you can pay your mortgage with. And then it kind of goes down from there where gig work might be that you’re making your car payments. And some extreme examples, some people are using gig work as their full-time jobs. And then it even goes down to task-based work where you might circle that cat in a picture 500 times and then be able to buy a latte that week. So, there’s these different types of things. It’s also good to know I highlighted the things in yellow that were really showing up as basically the companies or the things that are organizing that work, they handle trust, organization, guarantee of your work, and follow-up. So, those are being handled by somebody else when you start looking at the gig economy or the task-based work.

So, what do you bring to these types of gigs?

Gig Work: What you bring

So, here’s a little graph just showing some examples of things and kind of what you bring. So, straight up the middle, you’ll notice at the 45 degree angle, you’re starting to see the pay. The pay is going up as you bring more. And what do you bring? Sometimes you’re bringing your skills and sometimes you’re bringing your resources and sometimes both. I think that it’s just important to notice this. Basically, things get more and more complicated, the skill sets can range from basically having very little to no skills, just owning a – basically being able to walk a dog or deliver groceries up to being able to perform brain surgery or write computer code. And of course, as a nerd I’m going to say something about writing code being cooler than brain surgery. But then the resources when you look at what type of resources you bring it might just be your phone, but then again up to scale of you have to bring a drone that has high-definition cameras on it, or a car with insurance. Those types of things. So, basically the pay goes up and you need to be able to bring and do more.

Why is it Different Now?

So, why is this happening now? Why is this changing. We’ve had gig and contract work for a long time but why is it getting different now? Basically, a lot of wat’s happening, we’re more connected. There is cell phone and internet coverage everywhere for people, except for that one woman in the photo who I just love that shot with here. The one without the phone. But you can see that we’re going everywhere and the devices and the people are able to connect wherever they may be. The other thing is we’ve gotten really good as a culture of humans at turning big jobs into small discrete tasks. A great example of that is a hospital. And so, you have everybody from an intake person to a billing person to a doctor to a physician’s assistant and nurses and nurse’s assistant. And so, your visit through a hospital takes on many many tasks performed by many different people with varying levels of skill and specialization.

Other things that are influencing now is the globalization allowing for work and communications and payment to flow across borders. And that allows a much larger pool of workers. Also, things like all the sensors and improvements and making sensors cheaper whether it be a GPS or temperature and pressure sensors that we have on our phones. Those things are kind of much more prevalent.

New Tech to Enable Piece Work

And that kind of leads into this idea of what is the bigger tech that’s allowing us to do this kind of piecework. And I’m saying piecework and I’m actually trying to get this idea that these things can be done by humans and they might be done by machines. We’re going to be sharing some work pools. And so, I think one of the things that is important to note here is that we’ve gotten much better at turning large jobs into discrete tasks. A lot of this work has been going on for decades with the work in supercomputers and parallel computing. We’ve created new algorithms and AI is helping us actually divvy up work. And then on the other side of that is once we divvy up the work is actually making sure that it gets done correctly. I think you look at little things that are happening in there now like with Captcha where your validating somebody’s work potentially. So, we’ve gotten better at making distributing or breaking up work into little pieces, then checking on it, and finally putting it back together. And so, another idea is that we are able to create training simulations and create the situations and such that you can train on an individual task instead of spending years and years learning something you might just do a simple task and learn that in 4 or 5 hours, for instance you might play a game and learn how to diagnose a break in an x-ray. And you might just know how to look at a leg fracture but you’ve been trained on that one thing. And that could be you or that could be a computer.

Piece Work – Not Just for People

Another interesting to note is that it’s not just people that are part of this kind of new work in gig economy, but companies also. When we look at something like Amazon, Amazon is aggregating lots of companies, lots of sellers into their world of work. And so basically Amazon acts as that company that provides the trust and the organization and the guarantee. And so, the sellers are basically acting as a gig companies. Convoy is another example of that where they’re trying to become this shebang aggregator and they’re using AI to maximize trucking loads and routes and basically divvying up that work into different places.

Hybrid Intelligence

And so, finally this all comes together as sort of a hybrid intelligence where there’s people and robots and AI working together. I think a great example of this is looking at the five levels of self-driving cars. We seem to be as a society pretty fascinated by self-driving cars. And so, I think it’s a good way to look at how machines and humans are working together. So, if we just look at the very extreme end with level 0 is basically what we’re used to in large part where we drive the cars and there are no robots and there is no AI involved. And then you start getting up as you go you move along, most of us are familiar with things like Tesla’s autopilot. That’s a level 2 where it’s able to control things. It’s kind of an enhanced cruise control. Humans get to take a break but humans are in control most of the time. And then start moving up and you get into places where like level 3, robots can do some very simple things, basically like you’re stuck in traffic it might be able to keep driving for you but like the speed picks up or the traffic clears or you need to exit, that’s the human in control. And then finally you get into those situations where basically the robots can control the car almost always unless it can’t in which case there might be snow or construction. But again, humans are going to have to be able to step in.

So, in all those situations, humans are working and at the very end is where robots are completely able to do the task completely after training and working with the humans. And so, that is the level 5. That’s something we’re trying to get to. But it’s still a ways out. But I think that this is just a good way of showing how humans and robots work together and Denise is going to talk even more about the future of work.

The Rise of Automation.

Denise: Great. Thanks Erik. So, just taking a step back and I figured I would start by looking at the big global picture here. Erik touched on this a little bit earlier. I love that photo of everyone with their phone’s out and the one woman in the middle without the phone. There’s definitely a level of interconnectivity today. Everyone has at some point of their lives interacting with a bit of AI and so I think that’s definitely a really fascinating point to touch on. And there’s a level of demand and supply here. As consumers, we constantly want more. We want things faster. And better quality in half the time. And that desire drives demand for the sharing economy and businesses in answer to that demand have responded by trying to of course leverage automation to deliver these experiences efficiently and quickly at scale.

And just to touch again of how that automation is really touching on every part of our lives these days, you just think about Alexa or Google Home, self-driving cars which are touched on in his earlier slide, Netflix recommendations, all of these really have a component of AI in them. And this one graph that I found here on the left-hand side that I thought was a really great representation of just visual representation of how AI touches on each aspect of the org. This really goes to show how big of a market AI has become for business productivity and just some interesting stats that I have come across is that employees spend around 20-40% of their time searching for documents manually and lost productivity leads to around a trillion dollars of lost cost for companies every year. So, there’s a huge amount of room for optimization there and I think that this really just goes to show how there are a lot of more manual pieces of work that AI can take on. And I think that one of the common fears or misconceptions is that AI will really take over your work completely but I think that what this goes to show and what hopefully the rest of my slides will go to show is that AI really is going to be taking a lot of the pieces of your work that you maybe don’t enjoy as much or are a little bit more routine that they can take that on so that you can focus more on the parts of your work that you do enjoy the most.

Training Comes First

And so, let’s hop on over to the next slide and I’ll talk a little bit more about the first points why AI is still very much in it’s infancy. So, one of the first really, really crucial things to keep in mind and to be aware of is that training comes first. So, machine learning algorithms need to first be trained on how to think and how to process things. And that is really why crowd-sourcing companies like Observa are so important. And we learned this first-hand definitely the hard way at Ricoh Innovations. This is why we love, value and realize the importance of partnering with crowd-sourcing companies. So, our retail execution and solutions and Ricoh Innovations is called retail intelligence and it’s a computer vision technology that helps consumer packaged goods, CPG brands, and brick and mortar retailers efficiently track corrective actions such as incorrect placements, out of stocks, and things like that and understand trends like share of shelf. And so, for each of these products on a shelf as you can imagine there are a lot that a user wants to track, our machine learning algorithm needs to be trained and needs to learn to recognize these products. So, on average that means that the algorithm needs at least five to ten photos of the products in various angles and distances. You’ll see here this photo here would all be Dove deodorant here. This is just an example of the different variations depending on lighting conditions or the way that the product is stocked on a shelf it will look very differently. And to the camera and to the software, unlike the human eye, it needs to be trained to recognize that this product with so much glare on it is the same as the one that is very cleanly captured as you can all the way on the left-hand side at the top and the bottom photos would look very different to a camera and to a computer.

And so, if you think about the average number of products a CPG produces on a yearly basis and on a seasonal basis the task is pretty daunting. And again, that’s why it’s very important to work with gig economy workers that can be assigned to stores closest to them and they can cover multiple branches of a store to capture variations like that.

AI Needs Us

The second big point that I want to talk about is that AI is actually not that smart. So, we have seen lots of really great stories, interesting stories about how AI has beat humans in board games and how AI is now taking over and able to build other AI and help build things in factories. But really, a lot of times AI doesn’t have the ability to self-troubleshoot. And some of you may be familiar with Elon Musk’s new – and talking about automated cars – with the new Alien Dreadnought factory about his idea was to make it fully automated. And he very quickly – he was really trying to build this to help keep up production of the model 3 cars and very quickly realized that the factory kept falling short of about 3,000 vehicles a week. And what he realized was that a lot of these intelligent robots were actually creating bottlenecks rather than helping optimize and create efficiencies. You’ll see in his quote how excessive automation was a mistake and that humans are still very much needed in the loop.


And so, the next slide here, I’ll talk a little about humans in the loop. So, very quickly just to provide some context, deep learning in machine learning really refers to the number of layers in a neuron network and is really modelled after the human brain. And I love nerding out about this a little too, but essentially, a lot of the inspiration comes from modeling the human brain and a lot of the thinking is in the ultimate goal is to get it to think independently like that but today’s deep learning systems don’t really resemble our brains just yet. There are still a lot of work and research to be done in order to get it to the point of general intelligence which allows AI to reason and to learn on its own and to build mental models on its own. That’s really why it’s important for humans to still be in the loop.

Humans can really help train as I mentioned on the very first slide to tune – so really coming and stepping in whenever there are certain parts that need to be troubleshooted and then testing the data to help build out the algorithms. And this is really why once again trained – it’ll be very important for crowd-sourcing companies and taskers and people who are working on gigs to be trained to learn how to help – to be trained to train this technology and then to help optimize and correct and really work together with the AI technology. And I think the reason why I really see this relationship continuing to work is not really through humans helping AI completely take over at some point but really AI will help once again a lot of the automation stuff that is a little bit more mundane and a little bit more repetitive while humans will really help come in and either personalize a lot of that or come in and do more of the creative work on top of that and a lot of the reasoning that layers on top of that. And I’ll talk a little bit on one of my later slides – my last slide actually – about how you can step in and really start to future-proof your work and some of the skills that will be important to layer on top of that. But let’s hop on over to my second to last slide here.

The Human Touch

The human touch. So, this is one of the really important pieces that will continue to – this is where humans and where the individuals working in the gig economy will continue to play a very huge part. I mentioned personalization earlier. So, research shows that a majority of customers – about 77% – choose or recommend brands that offer more personalization. And no matter how personalized chat bots or personalized virtual assistants become – you know we now have AI that’s making phone calls for us – people still aren’t quite – it won’t be quite the same as talking to another human, right? At the end of the day, there’s still a lot of limitations on how a chat bot can pick up on emotions and a lot of the nuances of conversation. So, a lot of that kind of more human-to-human interaction will still be very important. And it will still be very important for humans to really pick up on a lot of the pieces that require brands to predict and to understand behavior.

Future-Proof Your Work

And then the last slide here, talking about future-proofing your work. So, I touched on a lot of these points earlier. I just wanted to really reiterate and kind of put them all together on one slide. A lot of the pieces that a lot of the things that you can do to future-proof your work, first of all being here today and being on this webinar is a really great first step. I think staying curious and technology is always changing, software is constantly evolving even on a month-to-month basis so it’s just great to always be keeping an open mind and learning and soaking in all this information. Empathy is very important as well and that really is that human touch and understanding nuances and understanding behavior and helping predict is a very big part of that AI is still not completely able to function on its own. Industry knowledge, I mentioned that as well. Just keeping up to date with your industry and just really building yourself as an industry expert will go a very long way. And then, creativity and problem solving, that will really plugs into that troubleshooting aspect of AI how there’s still a lot of gaps or holes in that overall workflow there. It’s very important for a human to come in there and decide where AI kind of flows next or help guide and once again troubleshoot and help some of the things out.

Thank you! Observa x Ricoh

And I think that’s it.

Erik: Yeah. So, thank you for attending and coming and we wanted to be able to answer any questions if people wanted to ask anything. Feel free to type those in. We are here for questions if you would like to ask any. It doesn’t look like anyone is chipping in with questions. So, I would like to thank you again very much and please free to reach out at any point. You’ll get a follow-up email from the webinar system and feel free to just reply and ask us anything that you might have to ask. But otherwise, thank you very much and hope you have a great rest of the week. Thanks Denise.

Denise: Thank you.

Observa End of Year Wrap-Up - Webinar Transcript

Observa Title Slide

Good morning and thanks for joining me and Observa today for our webinar. We’re going to be covering our end of the year wrap up for 2018 and within this webinar we’re going to cover a couple of topics: our Observa index report which this was our second year doing this, measuring the cost of goods different baskets of goods across cities in the United States as well as our findings from Black Friday.


So, first a little bit about Observa. So, we help consumer brands, third parties – such as merchandisers, distributors, brokers, as well as retailers – measure store execution. And we do this to help them have a better means of understanding what’s happening in front of consumers to fix that store execution and perform better with consumers.

Index Report Overview

So, first of all we’re going to cover our index report. So, this is the second time that we’ve done this study as I mentioned before. And we’re doing it across the United States and our results revealed a significant trend on how brands of retailers’ price products in different markets. And we showed a correlation on how that pricing relates to median incomes across the United States as we move from city to city. So, we found that the price of most items increase with the size of the average paycheck across the cities that we measured. But it does it as slower rate than income actually increases. So, the baskets of goods don’t necessarily cost the same as a percentage of income.

So, they cost more in the cities where the income is lower. And so, that leads to an income gap. And of course, the income gap is going to affect a lifestyle towards choices for some of these consumers.

Index Report – How

So, how did we do this report. So, we measured and did observations across 22 U.S. cities and the campaigns in the cities were conducted during February and April and the retailers that we looked at were consistent across these cities. And, we were looking at stores like Target, Kroger, Wal-Mart, Whole Foods, Natural Grocers. And we created four different personas. So, these are lifestyles that loosely correlate to common stages in a person’s life. And so, these four different baskets of goods were fun foods. And so, this Price Index includes items like Cheetos and Oreos and other fun food items like that. The next is parenting. So, the parenting price Index includes things that have like baby items, like diapers in event bottles. And then we moved into healthy foods so, the Healthy Foods shopper price index includes items like organic bananas and Justin’s almond butter. After that then we looked at the golden years. So, the golden years price index included items like insure and playing cards and so, on.

Index Report Findings

So, within our findings we found that the yearly cost of a basket – so, we first were measuring the basket of goods each week. And then the prices increased at different rates and you’re going to see this as the slopes of the lines in the cities. So, you’re going to have the highest cost being for the golden years, so, these are the people that are the most advanced in life in general and then the lowest cost is down at the bottom with the fun foods that’s the blue line on this graph. So, the results immediately showed that the basket costs increase as a city’s median income increases. So, there’s a correlation there. But upon further analysis these increases did not create an equitable difference in the relationship between the variables. So, they’re not they’re not rising at the rate that the incomes are rising. For instance, you can see that the lines are relatively flat here as the income increases but they are increasing so, there is a correlation it’s just not as strong as is as you might expect. So, this means that the costs of basket goods did move as a direct relation to income in these cities as we go up in the median income.

Findings Continued

So, what we’re seeing here as we then look at this as a percentage of income is that the percentage of income required to purchase that basket of goods decreases dramatically as income increases. And so, this is a pretty strong slope here. One thing that this didn’t take into account is the cost of other things that you might need to purchase whether you’re living in the low-income city or the high-end income city such as housing or transportation. We’re just looking at the basket of goods which includes food.

Example Basket – Healthy Foods

So, let’s look at a couple real examples. A couple of the baskets themselves. So, now as we dig down into the data and now we’re looking at what does it look like in individual cities you’re seeing on the graph on the left here. So, this is healthy foods going from Memphis going all way down to Palo Alto there and then on the right we’re looking at the actual costs. So, what do bananas cost in Memphis versus Palo Alto right? How much does a LARA bar cost between those same locations and so on? And so, you’re seeing variation here in the actual prices that we’re collecting as we did these observations with our people out in the field.

So, what you see here is that healthy foods have the second greatest positive correlation for price in median income. So, the cities with higher median income are paying more by the dollar amount for these items. However, when we dig further into the data and look at specific cities we find that, for example, by percentage of income is 3.5 times more expensive to eat healthy in Memphis than it is in Palo Alto. No. What’s funny here, what’s interesting, is this doesn’t take into account or that measure doesn’t take into account the dollar spent. So, Palo Alto still does spend more money but they’re spending only $300 more in expense per year than Memphis. But as a percentage income back it’s to the figure I said earlier, it’s 3.5 times more expensive for that person because the income level in Memphis to buy that basket of goods. Palo Alto does spend more dollar-wise but it’s more expensive as a percentage of income, much more expensive in Memphis to eat healthy.

So, what you’re seeing is we look into the individual items here is the chicken, almond butter, and milk or the items that very most in price from city to city. And you know knowing that chicken and milk are perishable items, right, you may be able to contribute that fact to higher distribution costs there locally. So, I’m not completely sure but that’s something that we might be able to determine from the data. So, LARA bars, bananas, and GT’s kombucha demonstrate and stayed at the most consistent pricing across the cities. So, LARA bars, that’s to be expected. Shelf stable item. And so, that’s a brand that’s basically probably had standard pricing for the most part across the US.

Example Basket – Golden Years

If we look at another basket, so, this is the golden year, so, these are the personas of the aging individual. This has both the strongest correlation where the city’s baskets were close to the line from that previous graph we showed in the steepest correlation or with the greatest slope in the line meaning that this basket pricing changes the least with income changes between the cities.

And so, we saw you know with the last one with the healthy eating that Palo Alto was $300 more than Memphis for instance. Well in this one we find that it’s only a $100 more for this basket of goods in Palo Alto than New Orleans but it is still 3.5 times more expensive to purchase the goods as a percentage of income in New Orleans than it is Palo Alto. So, in New Orleans is on the top here in this one it was Memphis on top in the last basket of goods is the greatest difference in price. So, the interesting stuff around the correlation here. So, another fact is that Raisin Bran was the most consistent in pricing and So, that probably has to do with Kellogg’s having more standard pricing again across the country.


So, what are the takeaways from this. So, the report confirms that there is a disparity in the cost of the basket of goods as a percentage of income between U.S. cities. So, this is going to result in a challenge in the availability of products to lower income folks. And So, that’s due to a lack of access to some of these products or that’s what the result is including what foods that they’re going to choose to purchase that they’re able to purchase because of the high relative cost based on their income.

So, we found that as median income increases the cost of each basket tends to increase as well. However, these price increases do not level the playing field for all consumers to have the same access to products. So, once again it’s that availability issue. So, when we calculate the percentage of the yearly income that each city spent on the baskets studied we revealed this economic imbalance.

Consumers in low income areas need to spend more to live on in any of the lifestyles. It is clear that consumers have limited purchasing power in many areas of the country. And So, this leads that opportunity for brands and retailers to change this. That’s the way I look at it. You know you’re what you’re seeing in the data is that there’s a choice problem. People in the lower income cities are not going to feel as comfortable spending the dollars on some of these products. So, how do the brands respond to that? What can they do to increase efficiencies and lower pricing in those areas to attract more buyers? And that would result in a win-win for the brands for consumers for the stores if they’re able to make some price concessions in these lower income cities.

Observers on Black Friday

So, let’s move on to the next study here. So, this was on Black Friday – oh and I just want to point some out here, for all of this there’s more information on our blog. You can see that URL there it’s observanow.com/blog. It’s also, you can find it on our website. It’s in “Resources’ and then ‘Blog.’


Attendance and Lines

So, what do we find here at Black Friday? So, we looked at the attendance at various stores across the country. You can see them listed here. We have Best Buy, Macy’s, Target, J.C. Penney’s, Kohl’s, Kmart, Big Lots, and True Value. So, we sent out our observers to these stores across the country. And what we found and what we’re measuring are things like foot traffic which you can see on the left and then the length of the line and what’s interesting we look at this is it’s pretty well correlated right. You’re seeing that Best Buy, Macy’s, right they’re the first two with the most foot traffic in the graph on the left. And the longest lines are the one on the right. And so, there’s a strong correlation there.

There’s one obvious exception and that is Target and Kohl’s here. What we’re seeing on the on the graph on the left is that Target had a greater foot traffic. So, people walking through the front of their store on the left and yet they have some means of greater efficiency in their stores to end up with more preferable lines to the consumer in the graph on the right. So, whether they are staffing accordingly to get more people to process at the register, whether they have better technology to be more efficient at the rest or better training for the employees, there’s some reason why they’re able to deliver better service to the consumer and end up with shorter lines.

So, in previous years department store successes declined, but this year it appears that some of those stores saw more foot traffic and longer lines. That was a good thing that we noticed out of attendance here this year for Black Friday.

Mood of Customers

So, another thing that we measured was the mood of customers. So, beyond foot traffic and length of line we asked our observers to look at the emotional response of the Black Friday crowds at each store. So, here we’re measuring people that are generally happy with the yellow and then people that are less happy with the blue. And so, this was by counting those people that are that are there at the stores waiting in line and so, on. Right. So, what are our research shows is a strong likelihood of shopper or what other research it shows a strong likelihood of shopper misbehavior where people are unhappy. So, you know there are crazy news stories about people getting in altercations and other things happening at stores on Black Friday. And obviously those are things that we all want to avoid right. And So, what’s going to contribute to that? So, it’s things like long lines, a lack of goods available in the store, you know things like that.

So, in stores with low foot traffic and short lines such as True Value and Kmart, back from the previous slide, consumers were reported to be happier. So, those shorter lines. So, we can see that here in this representation. Meanwhile stores like Macy’s and Kohl’s which have longer lines reported more upset customers or people that were less happy.

Best Buy

So, let’s look at some specific example let’s start with Best Buy here. So, Best Buy is well known. It may be infamous for their Black Friday deals. So, they draw these massive crowds. People start lining up in the middle of the night at their stores and they basically flock to these deals and it happens year after year, there are stories about it every Black Friday. So, that’s pretty much expected this year and that’s what we found. Right.

So, our observers report it back to us that the stores were busy, they had long lines, as for the customers they seemed primarily positive with this store unlike some of the other stores that we observed that day. So, this is all likely due to Best Buy’s experience attracting these shoppers’ year after year and then keeping them happy and being able to deal with the influx pretty well of shoppers, the crowds that they attract.

True Value

Yet when we look at other stores – another example kind of on the opposite end of the spectrum – is True Value. So, it was not the case there. They offered lower deals than any previous year, so, maybe they’re finding that their deals on Black Friday don’t produce the financial results you’re looking for, I’m not quite sure. But they’re not making the same attempt to attract those people to the store. So, fewer deals. And that’s likely resulted in this lower foot traffic, you’re seeing empty stores here in the photos that were recorded, short lines, small crowds did ensure happy shoppers. So, for the people that did make it to True Value they had a good experience.


Macy’s so, they offered slightly higher than average deals than in the past and they saw great foot traffic and long lines. You know there’s been some negative news about Macy’s or unfortunate news. They had recent closure of about 100 stores. Yet we’re seeing now growing sales for them, so, maybe those – I don’t know if it was direct from those store closures, maybe it’s because we’re now only looking at their better stores and better performance stores in the past. But what we’re seeing for Black Friday was positive for them. And obviously the analysts out there arguing about this what is the future Macy’s and there’s conflicting beliefs around that. But from a Black Friday perspective if this traffic was any indication, shoppers clearly still choose to shop here over some of their competitors.

Wrapping Up our Wrap-Up

So, that was basically what we found out from Black Friday. And there’s more of course on our blog as I mentioned before. So, I want to thank you for joining us today and I’m going to kind of wrap up and summarize here. So, today we showed two examples of how the flexibility of the Observa platform can be used to gather market information quickly and effectively. You know we normally show things such as the measurement of product merchandising on the shelves, so, store execution around general merchandising which is obviously a problem out there in the market. Our unique platform enables us to measure it like no other company out there. And we also, measure promotion execution. So, whether we’re talking about secondary displays in the store, end caps, demos, we’re able to do things in tight time windows. What we did today was something different than that. Our platform was highly flexible and we’re able to measure things that most other companies can’t do especially not at the speed that we can.

So, I want to thank you for joining us today for our end of year wrap up and to see what we learned in our annual index report and during Black Friday this year. Observa is here to help you and your company measure how your products are doing in the market in front of consumers. Please consider Observa in your planning for 2019 or reach out if we can be of any assistance. Even if you just want to talk about what your plans are we’re happy to be here to help with that discussion. Please send an email to sales@ I can be reached directly at hholman@ Thank you very much for joining us today.

Observa AI Intro & Demo - Webinar Transcript

Observa title slide

Adam: Welcome everybody to the Observa AI & Observa for Field Reps webinar. Today we’re going to be focusing mostly on Observa AI. I want to thank everyone for coming and we’re going to try to make this as interactive as possible, so if you do have questions that pop up during the webinar feel free to send those in and we’ll either answer them in real-time or at the end of the webinar.


So, for those that are familiar with us, and for those who are not, we are Observa. We provide real-time information for what is happening with your products on the shelf at retail. And so, kind of phase one of our company has been allowing you to gather metrics that you previously didn’t have the ability to go out and get. So, using the crowd, going to stores, collecting that real-time information so you can manage the point of sale and you can increase sales, eliminate voids and out of stocks. And phase two is building on top of that with new services for field reps and AI technology that really takes our offering to the next level and allows for analytics and calculations that previously would have taken a lot of time to be done manually.

Your Speakers

So, today’s speakers: I’m Adam Kirk, director of operations here at Observa. Previously kind of did account management and sales, so I might be familiar with some of you who have joined us today and some I have not met before, great to meet you. And Erik Chelstad, our CTO and Co-founder is here as well, and he’ll be speaking here in a moment more on how this service works and I’ll touch more on the features and then on a couple of use cases for both CPG brands and for retailers.

New Observa Features

So, as we said, exciting new features. We’re really excited to share these with you today for our current customers and our future customers to take advantage of these offerings. As I said, one of those is Observa for field reps, and we’re not going to focused on that too much today, but I will touch on it briefly. Really what it is is it’s bringing all the services that we have – the qualitative and quantitative data, the images at store level – and putting that into the hands of your field reps. And we really think when you combine our crowd of users, both auditors and merchandisers, people with some sales experience, and combine that with your field reps, it really allows you to have nation-wide – actually, across all North America – coverage of any retailer in real-time so that you can go in, find out what’s happening on the shelf, and make changes quickly to give consumers great experience with your products on the shelf. But, we’ll touch more on that probably in a future webinar, but today we’re going to focus more on Observa AI and what that means for our customers and what that means going forward.

So, Observa AI is an artificial intelligence solution that combines image recognition and sales data to give you even more information on what’s happening with your products on the shelf. What’s exciting about this is as I mentioned before, it opens up a whole new range of possibilities for metrics and insights and analytics that can be taken off the shelf, because previously there was a time that if you had a 100 SKU’s or you want to do an entire category to gain insights of the shelf, that takes manual data entry and then afterwards, the calculation time on the metrics that we’ll get into here in a second, we’re saving a bunch of time by letting the computer tell us which products are on the shelf, where are they located, where are the out of stocks, and then doing the calculations in 30-60 seconds it’s going to save our reps, our observers, and managers a ton of time on being able to gain valuable insights.

Observa AI Features

So, what I’m going to do here is actually open up our preliminary report we made using some of this image recognition and AI technology and kind of walk through some of the metrics. And as I mentioned, these metrics are just the start. With artificial intelligence, we’ve opened up now the capabilities to have metrics that previously you just couldn’t get in the amount of time I mentioned. So, as I’m going through these – you know, metrics that you have in your business, metrics that you would love to have for your CPG brand or retailers and know that what we’re going through here today is just the start. And there’s so many different metrics we can add on top of this.

Website page with images

And so, what we’re looking at first here is share of shelf. And I’ll show you the interactive nature of this report. And so, what it does is it blocks the brands together. So, this is a Walmart location and as you see in the gray here, we have the Walmart private labeled brands. Old El Paso is kind of the example brand we’re using for this report. And so, what you can do is you can click in here and see the data behind the products whether that’s how many facings the product has, which products are there, you could tie in sales data here on this interactive feature. And again, this generates in about 30-60 seconds from when somebody takes a picture in front of the shelf or takes a few pictures if they are going to be doing a section or a category.

So, the first one that actually comes up here is called Ring of Buyer. And this is a cool metric that we’ve kind of created here at Observa. It allows you to dial in where your product is versus key products in your category. So, a lot of people talk about whether your product is at eye level, where are you on the shelf – and that is hugely important, but what really matters is where you’re located in your category. Are you in the center of your category? When somebody’s looking at maybe the leader in your category, where you’re located compared to them and this brings in some visual metrics behind the scenes that say you’re in rank one, rank two, or you’re right next to those key products and that allows you to measure over time how you’re doing compared to some of those key products in your category.

So, for example, if you look down here, these products, the other taco kit products, really far away from most of the set, they’re maybe down in rank seven or eight which is going to be problematic for somebody who is going and looking maybe at just the largest collection of products in this category is. Sales data by product – we think this is really exciting because it allows you to finally figure out the “why” behind what is causing sales to either go up, go down, stay flat, and so by being able to go in, identify a product, figure out how it’s been selling, and do this whether it’s at a chain or across a category allows you to finally know which levers you’re pulling are actually having an effect. Because you can sit at the headquarters, put in a great marketing plan, a great new packaging, and look at the sales data and maybe you think you understand what is causing that drive in sales or that drop in sales. But without actually being able to go to the shelf, look at which products are there – which competitive products are there – and really identify those things allows you to finally say, “This is the ‘why’ behind what’s driving sales.” And then you know which levers are working, which levers are not, and which you can pull or push going forward.

Sales data by manufacturer here. Previously it was broken down by brand. Target sales, another sales metric brought in here. And then down to share of assortment and share of shelf. So, this is tracking over time your share of the assortment and your share of the shelf compared to your competitors. So, very valuable information, very competitive trick to see who may be taking shelf space from you or who you may be taking it from. Also, you can look across different retailers and say, “At this retailer, this brand has a ton of share of shelf. Maybe they’re targeting this retailer.” Or maybe they’re not. Maybe it’s an opportunity for me to really find a retailer or demographic that I can take a nice foothold.

Section facings. How are your facings growing compared to the category? Again, if the category is growing but your facings aren’t, you’re actually losing space within that category or within that section of the store. Manufacturer Facings and Sales. Another sales metric here brining in the facings compared to sales. So, if your product has a lot of facings but is low on sales, you might want to take a look at that product mix or if you only have one product and it’s doing great, maybe it’s time to add some more products to that category or work with the retailer to get more of your SKU’s onto the shelf.

Here’s some – and again these tools are available for field reps on Observa. So, this is stuff that can be operational. So, there are some operational stuff down here for a field rep that goes and scans the shelf and says, “How many products are here? Do I have all of my facings?” This is facings count. Which products are not here at all? So, which SKU’s are missing. And so, we view this data both operationally and from a management standpoint as a very valuable tool whether you’re going to use it in real-time at the store – take that data and back in 60 seconds, cut down on a bunch of time you’re spending in front of the shelf doing manual counts of products and inventory levels. Or if you’re going to use it more of a management tool over time as I said, tracking trends in your category and such.

Artificial Intelligence as a Service (AlaaS)

So, we’re going to go back here and going to loop in Erik here in just a second. And so, what Erik’s going to talk about is how this Artificial Intelligence as a Service and so, you all know of SaaS, or Software as a Service, it’s been a great step for businesses to be more efficient and we really think Artificial intelligence as a service, giving this as a service to you, us doing a lot of the behind the scenes work so that you can just take advantage of the information is going to be super valuable. So, I’m going to hand it off to Erik and he’s going to talk here for a few minutes kind of on how we’re doing this technology.

How it Works

Erik: Thank you Adam and hello everybody. Just, I will be brief, I’m an engineer, so I will try not to go into any of the really nerdy stuff here, but again, always available to answer those questions at a later time if you’d like. There’s a lot of buzzwords that you might have heard out there around AI. Deep learning, and neural networks, object detection – and I think that’s it’s important just to note that these are the tools that we are using. Again, if you want to dive into that – and just the object detection I think this is important to note, this is if you look at the picture, you’ll see that there’s Tabasco and that is an object and it’s detecting where the object is versus just sort of classification of pictures. So, those are the buzzwords on what’s happening.

Training Data

And what does Observa do for you in this case? So, one of the big things that you have to do for the AI is you have to basically train it. And so, you need to feed the AI as much data as possible. This is something that we’re very good at Observa. So, we’ve spent a lot of time figuring out ways to collect data and initially to create a training dataset, we’re able to send our observers out – again, as Adam mentioned these could be your folks in the field or ours – but we can collect a lot of different images. This is necessary because the AI needs to be able to see different things – so different lighting, different angles, potential different products based on where you are in the world and then you take that data and you actually need to label that data. So, it needs to be – again, it has to do with creating the boxes which need to be drawn manually at first so basically the AI can determine what is in the box – but in the beginning you are telling it what’s in the box.

Questionable Input

So, this is what we do to say what’s in the box. Something that is unique about us as far as Observa is that we can – the AI is not perfect. There’s a very high probability that it’s going to get every potential product on the shelf, but occasionally it misses something. This could be a new product out on the shelf, it could be just a blurry photo – there are many reasons it might know what the product is, but since we have the observers in the store, we are able to actually have them clarify what’s happening and collect more data so that not only will the data be correct when it comes back to you, but also, the AI can get a little bit smarter as it now has a new piece of data to train on.

How it works / Putting it Together

So, basically, all those little components go together in a way – and there’s a timeline on here: 48 hours – and this would be basically how this works as a timeline. So, if we work with you, we’re going to get your initial products and some information. This might be some things like the marketing info and the products including some of those beautiful marketing photos that you have, but then we’re going to go ahead and create a campaign and in the first 24 hours, have people go out and get lots and lots of images. And then the next 12 hours, we’ll be able to get that data, clean it up, label it, and feed it to the machine so that 12 hours after that we can have a trained model that is now able to recognize your products or your competitors’ products on the shelf. So, we also, again our system will continually monitor in that anything that comes in that’s questionable we’ll be able to determine what that is and make the AI smarter as well as continually dealing with updates. We see that as new products are released whether they’re yours or your competitors, whether there’s packaging changes, whether it’s holiday assortments or those types of things, we can deal with the updates and continually get that going.

You Own All the Data

So, that is really kind of that in a nutshell. And I think it’s really important to remind you and everyone that you own the data. Anything that’s collected with our system is yours. This is all of the data and training, it is yours. And so, we’re not taking your data and sharing it with your competitors. This is very specific to you.

Adam: Yeah, and just to touch on that, that’s important because for a lot of our customers they’re going to have competitor data and their own data in here and so we wouldn’t share that across different competitors.

Use Case: CPG

Great. So, coming back here, one of the use cases for this I touched on in the beginning that we’re really thinking this is going to be a valuable tool for both CPG brands and for retailers. And so, on the CPG side, there’s so many different aspects of the business that can take advantage of this information whether it’s on the sales side, whether it’s in the marketing department, whether that’s trade marketing or shopper marketing, what’s on the inside, consumer insights, shopper insights, management, executive level, any sort of analytics individual. And it’s both operational as I spoke about and a management tool. It’s digitizing the shelf. It’s giving you those metrics that maybe you’ve come to learn on an Ecommerce site or on Amazon for example that have real-time information on who’s visiting sites, where things are coming from, what’s causing things to change and getting that now on the shelf in real-time. And so, on the operational end for field reps, we really think this could save a field rep 75% of their time when they’re in front of the shelf. Especially across a large category in the store. So, instead of spending all that time counting those products, figuring out where voids and out of stocks may be, instantly take a photo or a couple of photos and get that information back in real-time.

And then again for the managers, being able to spot trends in those categories. At the end of the day, we’re looking to drive sales and what I talked about was which levers are working, which levers are not working and actually being able to understand that across any store in the country – any store in North America for that matter – in real-time. It’s going to be super powerful, and at the end of the day the goal is to give the consumer – the person that is going into the store a great experience and are able to understand are they getting that experience, are we getting what we paid for maybe on share of shelf, is our promotion happening? All of those things now powered by AI is going to save brands and Observa a ton of time out there in the field.

Use Case: Retailers

And then for retailers, same thing, operational and management tool. And again, just like field reps are out there in front of the shelf, making changes in real-time and counting products on the shelf and looking at shelf tags and figuring out where things might be in or out of stock. This now available for the retailer employees as well. It’s something to think about as we advance this product. We do have different APIs and things like that that can tie into systems like an ordering system to save, again, more time. Less time manually entering in data, more time letting the computer do the work for us.

As we all know, we are always smarter when we have our smart phone with us. For example, we have the ability to Google things, find out stuff in real-time, access spreadsheets. The computers are making us more efficient and this is just the next step in that. Employees are going to be in stores for a long time. We’ve seen robots, drones, different things like that. The human element of having somebody that is able to make decisions and act quickly with help from the computer is going to make that person more efficient and more valuable over time and again, a great management tool to track what’s happening over time. Track things like planogram compliance which is something that I didn’t touch on earlier, but that is something that a ton of money goes into creating these planograms, predictive analytics, all of those things, but if it’s not actually executed at the store, if the products aren’t where they’re supposed to be, all that hard work, all that money spent behind the scenes isn’t actually happening out on the floor and we’re going to be able to track that over time and provide those valuable metrics.


So, again the summary of what we’re doing here is we’re digitizing the shelf, we’re taking those metrics and hopefully turning them into experiences people or consumers are used to having online where all the products are right there. There’s great details on the products and everything is being executed on and taking that to store level both for CPG brands and for the retailers themselves and at the end of the day, what Observa AI does is it provides the ability to do calculations efficiently and quickly than were previously possible either manually or with any of the systems that are out there right now.

Try it yourself

And a great thing, something that we’re doing that I don’t think is available anywhere else is for you to go to the store yourself and experience the AI in real-time. Of course, everybody here on our team has done it. Some of our customers have done it. It’s been a great experience for them. And so, here’s some steps on what you do to sign up. If you look at the email as well that you clicked on to sign up for this webinar, there are links for you to go create an account on our website, signup, login to our mobile app which is going to be on the iOS or the Android store, just search for Observa and you’ll find it. And you can go into the store, go in front of the shelf, take a photo, see how quickly those metrics we’ve discussed here today come back as well as there’s a possibility for the brands to have joined us here today. If you want to see this with your products – you know, we showed an example of one brand here today – but if you want to see your product, just send a message, send us some photos of your products and we’ll set up a campaign for you to be able to go into the store and see your products with all of these awesome metrics behind them as well.

Get ahead with Observa!

So, if you do have – yeah get ahead with Observa. Yeah, so what we do here is make sure your products are executed at the point of sale for both retailers and CPG brands. Different things you can do whether you’re looking at competitive analysis, looking at your brand specifically, but at the end of the day our goal is to give you both operational data to save your reps and employees time out in the field, as well as for management level how do you really understand what’s happening at the point of sale to help drive those sales.

Thank you! Observa

So, again I want to thank everybody for joining us here today. You can reach us as sales@ or I know a lot of you have joined us or spoken to us personally whether it’s me or some people on our team, feel free to shoot them a message or give them a call as well if you’d like to talk a little bit more about what Observa AI or even Observa for field reps may mean for your business or how you can start to test that out and see if it is something that makes sense for you. So, again, thank you to everybody for joining us here today. Doesn’t look like we have any questions. I’ll wait a couple seconds if anybody does have any that they want to send in here at the end. But, yeah, thanks again everybody for coming, please reach out to us with questions afterwards if you do have them whether it’s sales at Observa or with somebody on our team who you are familiar with. We look forward to talking to you all again soon. Thank you.

Observa - Small Brands Taking Bites out of Big Brands - Webinar Transcript

Observa title slide

Good morning. I want to thank you for joining us today for this Observa webinar. Our topic today is small brands taking bites from big brands. This is a really interesting topic to me. I love the whole market-space around companies and acquisitions and entrepreneurialism in building up brands and companies and I’m excited about being able to share a little bit about that with you.

Your Speakers

So, my name is Hugh Holman. I’m the CEO and one of the Co-Founders here at Observa. And my background is in technology and retail, sales, and marketing strategy. And we started Observa back in 2015 and are delivering analytics and insights directing actions for brands of all sizes and retailers. So, something interesting here we have a new dog in the office that happens to be my dog. But his name is Disco and there is a picture of him there. He’s actually a silver lab which is not a very common breed and a lot of people haven’t seen it. So, something to share.


So, Observa – we help brands gain greater efficiency and effectiveness based on giving insights and directing actions. And we have a platform that’s pretty expansive at this point that helps them do that. And so I want to thank you for joining us today on this topic and also start off by touching on some changes that we’ve seen over time. They’ve been pretty drastic and if you go back to the early 2000’s and even before that with the rise of natural and seen what was happening in the natural food space, we saw a lot of smaller food brands starting to show some success in the market that was disruptive to big brands and has now really changed the grocery landscape. It’s really interesting. So, this is a study here by Boston Consulting Group where you see this graph and one of the things that they point out in their study is that the really large brands are still seeing some growth. So, you see 2.5% there for the top 10 companies with the large brand portfolios. But it’s nothing like the growth in small brands.

So, the overall market of course is growing but what we’re seeing is that the new share of the market is really being captured by these small brands. We’re talking small in this case means less than a billion dollars in sales. A lot of people may be watching this saying “Well, that’s actually a pretty big company” and I agree with you, but in the case of consumer brands, anything under a billion dollars is considered small in this case.

Small Brands Successes

So, what was feeding that? Well, there’s a huge shift in consumer wants and expectations and it had a lot to do with education route. Health, feelings about changes in the environment, and going back to basics. And what feeds this is the need for authenticity, for transparency, for understanding what’s in their food and being more conscious of what people are putting into their bodies is at the forefront of that and another aspect around natural and organic you start to come into context with locality and what is produced in your general community and so people started looking for things that were produced locally, and shopped locally, and buy local brands and products. And as people researched more, it’s more likely that they are going to develop more emotional connection with things. So, uniqueness. As I mentioned before health-conscious and kind of a push towards being more healthy and natural. And what we see with the small brands of course is that they are able to innovate. And this is so challenging in large organizations – making decisions and getting all the parties involved to move things along whether deciding on what to work on or actually getting a project done. Doing R&D and innovating and so small brands are able to do that.

“Small” Brand Success

As an example here, we look at Kind. And they did something really interesting because they had a product and they got some feedback from the market and because they were small and they were able to adapt quickly, they took that feedback and they rebranded themselves with a clear transparent package. And you can see what’s in the products. So, it’s got a clean ingredient deck. All the ingredients are listed. You can see what you’re going to be eating and I think that that really resonated – or we know that that really resonated with the public. And it’s shown by the growth in their sales and their ability to capture market share and take it away from some of the big brands like Kelloggs and some of the people that were really dominating in that granola bar market. And the other thing is in creating a kind of attachment with the community, this spread kindness campaign that they did, it did resonate and they saw great success because of this. And so, this is an example of a smaller company innovating, capturing market share very rapidly and using transparency – what people are looking for – understanding what the consumer’s wants are and responding to it. Transparency, this spread kindness campaign, and creating an emotional connection.

Big Brand “Failure”

So, why is this challenging for big brands? That’s on the opposite side of the spectrum, right? So, why are they growing slower? Because they are growing still. So, the top 10 brands, 2.5% like we talked about before, but it’s just not at the same rate. They do have challenges in innovating, right? And so, it leaves them challenges in either doing that or figuring out other ways to deal with the fact that they are not keeping up in a market share perspective. So, it’s not that they can’t maintain their companies based on the growth, but that’s not what a big brand’s owners expect and especially if you’re publicly traded and the pressure of Wall Street, you need to deliver results and results are based on what the market potential is and if you’re not maintaining your market share you’re losing to the street and you’re going to suffer because of it.

So, you’re pushed then to do things like maybe acquire. And so, we’ve seen this happening again and again here. Where companies like those here on the board – Coca-Cola buying Honest Tea – that was an interesting acquisition because they bought in small. They did a small investment and I think they bought like 13.5 or 13.5 million they spent initially. It was a small amount of money for a percentage ownership. And then it came back later as they worked with the company and better understood the potentials and how Coca-Cola and their distribution system would be synergistic and help Honest growth, right, then they understood the value more.

General Mills went out and bought Annie’s and that was a big acquisition. So, much higher dollar amount just jumping in and these are just two different ways to go about expanding your market share and buying sales. But, Annie’s I think they were doing $200 million in sales when they were bought by General Mills for $840 million so, it’s a 4 times multiple of the revenue that they’re buying. It’s going to have to pay off over time. So, it’s kind of challenging. In the chart over here on the left, it shows more or less some of the statistics around some of these big acquisitions but what you’re seeing is the challenge that the big brands have which is to maintain their market share and how are they going to do it.

Company Examples

So, some other examples here. I think that this Campbell’s soup one is interesting. Campbell’s, once again a big brand portfolio, they own lots of different brands and different products and different spaces in a store and they went out and they watched Snyder’s Lance which is basically somebody up and coming and natural and positioned themselves against Frito Lay and they bought them. And this was a massive acquisition. And what they saw was a slide in their ability to maintain market share and they went out and bought somebody that represented effectively 20% of their sales, so post-acquisition they become 25% of the sales of Campbell’s and that’s just a huge buy-in to another company and it’s very interesting.

Here’s another example with yogurt. Yogurt has seen a great increase in the size of the market. And that’s exciting to watch in and of itself. But what you have is some dominant players, and I’m using Yoplait as an example here, where they failed to innovated and they kept doing the same thing and milking their cash cow if you will, but you had up and comers coming like Chobani’s a great example and Fage is another one, but Chobani bypassed them. Essentially, moved past Yoplait. What we see recently is Chobani has come back not with yet another Greek yogurt option but as a regular yogurt option to go head-to-head back against Yoplait. And so, they’re fighting head-to-head in the category that Yoplait dominated and Chobani came up in alternative for Greek yogurt. So, that’s very interesting to watch.

In the ice cream space here on the right, Ben & Jerry’s was the alternative. And they were going against Dreyer’s and the other major brands and of course they got acquired. And then we see Halo Top. Their growth in 2017 was 2500%. So, this is a new product differentiated in the market, so they’re going after health-conscious, you can eat ice cream but it doesn’t have to have all the calories and it resonated with the consumer. They’re taking a lot of shelf-space in stores. And it’s driven a 25% increase in sales in 1 year last year. And I just think that that is remarkable. And so, here you got Campbell’s who’s staying in the game by buying and other companies – I guess Ben & Jerry’s is owned by Yun Labor – but they are the largest ice cream seller on a whole. But Halo Top is somebody that they are definitely watching.

Retailer’s Role

So, let’s pull this back to what does a retailer do? So, we’re watching this powerplay by small brands and big brands and some acquisitions and so on – just general disruption in the market. And what does a retailer do? Well, they need to make sure that they are also responding to the changes, consumer wants and needs, their expectations, and of course we’ve seen that in the rise of availability of natural organic products in grocery stores which you can see in this chart on the right side of the screen there. It’s one of the things that consumers are looking for to draw them to come back for store visits, but selection of store brands is something that we wanted to point out here. You know, if you don’t have a selection so they don’t have to go to your store and then another store and then maybe another store to get all their products, if you can actually meet all their needs, they are more likely to come back to your store. And so, having that broad selection and being able to listen to your customer – so I think that loyalty programs are helpful because they actually give your relationship with your customer. Hopefully, you’re asking them somehow. What do you like? Or looking at the data and asking them what is it that’s missing? What do I need to add next and so on.

But the next thing I want to highlight here which is so near and dear to us at Observa is the point to make sure that your products are always in stock. We measure that for our clients and it’s just disastrous at times to see how often brands are running out, consumers are coming in, they’re disappointed because they can’t find that product on the shelf in the store.

Retailer’s Role (2)

So, let’s give a couple more example that are more specific to chains. HEB we watched them created essentially what I would call an innovation pipeline. So, they’ve developed this community event program which includes their own representatives and they invite local brands, local producers to their stores and talk to them about their products. And this is essentially giving them a way an audience to not only tell HEB about their products, but it’s also given to be a pipeline to look at products that are up and coming, that are unique, that might be good options for their consumers and help them. Then they can work together. Are there challenges with their packaging? Is their package size wrong? Are they hitting the wrong price point in the store for the consumer? And sometimes as a brand you don’t have as much information and maybe you’re not creating the right product for the market and I think that this is a really cool program that HEB has.

Kroger created some programs for smaller brands. But they’re mostly around how do those brands do things like introduce their product to the consumer through sampling and demos in the stores. Or through a POS, point of sale materials, to do in-store education and that kind of stuff and I think that that’s great. Especially from such a large retailer.

Whole Foods – I’m a little confused with where Whole Foods is going right now but I think I understand from a data perspective, but they’re actually pulling away from local and so we’re seeing less brand representation there from local producers, less opportunity for uniqueness which we knew from Whole Foods in the past and we’ll see how that develops over time and I’m not quite sure where they’re going with that. But we’re going to see it continue to develop I’m sure.

What does this mean for you?: Retailers

So, once again for retailers what do they do? So, variety is the spice of life, give new options to your consumers, make them excited when they come to the store, give them something to try, make sure your products are always on the shelf, always stocked. A way to give them something to try is through the demo, sampling, I think it’s so important and obviously it’s the most effective form of marketing for a brand, so making that easy for brands, inviting them in and helping the up and comers understand that that is an option for you. Make sure that you have a selection of natural and organic, I think everybody knows this at this point. And then make the shopping as convenient as possible for a consumer so that they come back to your store. They are finding everything they need and it’s always in stock and it’s always convenient for them.

What does this mean for you?: Big Brands

What does it mean for Big brands? So, I think the main thing here – and this goes for everybody – is always listen to the customer. What is the consumer thinking? Where are their changing expectations? And make sure that you’re thinking not only what are they saying today but what are the trends of that messaging. What are the trends coming across as and how are you going to address those wants and needs with that portfolio so you go down with the line and so are you going to consider acquisitions or are you going to buy market share? If you are losing market share how are you going to deal with that? And if not, are you going to do it in house, you’re going to be developing your own products, you got to be listening and adapting fast. You got to move quickly and you’ve got to figure out how to create that R&D pipeline and develop products quickly, get them to market quickly, get into the hands of consumers fast for that feedback. Do you have the right packaging, do you have the right messaging? And so on, so you know whether or not you have a winner or you have to keep innovating on the product until it’s the right fit for the market.

What does this mean for you?: Indie Brands

For smaller brands, marketing innovation – again listen to the consumer, do a lot of listening. Once again, this is the most important thing, understanding what the market is asking for and making sure that you are producing products that meet their needs and innovate. You’ve got to get your brand out there. And the most effective way to get your brand out there is to get it on the shelf. And so, figuring out what these programs are with the retailers, making sure that they have awareness of your products. Going to the shows I know that we spend a lot of time at shows like Expo East and West and the fancy food shows and wherever you can get your brand name out there do it. But you’ve got to get your brand in front of the consumer too. Take advantage of those sampling programs. Don’t take the retailer’s word that you’re on the shelf. If they may buy your product and they may make it in the warehouse, may even make it back to the store, but is it making it to the shelf – you need to manage the product at the shelf and use a service – maybe ours – get your own people out in the field to find out. But make sure that you don’t leave that to chance.

Observa can help!

So, how do we help all of you? So, for the retailers we’re all about increasing efficiency. Observa’s mobile apps and services for driving actions at the store using our Artificial Intelligence and image recognition to measure planogram compliance and direct actions to fix where products are out of place, or maybe you have out of stocks, other issues on the shelf there we can help you with that and our products are really fast and efficient to make your people in the field more efficient.

For big brands, it’s about accuracy of the insights. We’ve been relying on kind of challenging data, slow data for so long. I love Nielsen and IRI and the information that we can get from them – the aggregated data – gives you a relative idea of how you’re doing compared to your competitors in the market. But it really doesn’t help you fix problems at stores. And that’s where everything matters. It’s right there in front of the consumer. The consumer is everything. You have to make sure that your products are being merchandised as well as possible on store shelves. Expectations are only increasing at retail stores where 90% of the sales occur based on people’s online experiences. So, make sure you’re measuring at the store. We can help you with that. You can use our mobile apps. You can use our crowd-to-go where you can’t expand the reach of your people and we’re happy to do that. So, let us know if you want to talk about it.

Small brands – don’t leave it to chance. Make sure that you’re on the shelf. We can help you with that. Make sure that you’re trade marketing dollars – the scarce dollars you have – to buy shelf space and that you’re getting an ROI on it. That you’re actually getting that shelf space that you paid for that when your applying for a temporary price reduction that it actually occurs when you ship out a display that it actually makes it to the floor. We can help you with all those needs.

Thank you! Observa

So, I want to thank everybody for joining us today to talk about this changing landscape and what’s happening between big brands and small brands and the rise of natural organic and retail. I want to thank you for spending a little bit of time. If you want to learn more there’s our email address, just shoot us as note at sales@ Also, we’re going to be at Groceryshop. I’m actually on stage on Sunday. If you’re going to be there, I’d love to have you in the crowd and we’ll be telling more about Observa and showing more about our Artificial intelligence and image recognition, our ability to scan shelves and measure planned RIN compliance instantly. And it’s really a pleasure to have the opportunity to speak there. We have a booth. It’s booth S20. And so, come join us at the show. We look forward to seeing you there. I want to thank you for spending some time with us today. Thanks.

Stimulating Impulse Purposes Webinar - Transcript


Good morning. I want to thank you for joining us today to share a little bit about marketing and simulating impulse purchases at retail.


So, to start off with, I’d like to tell you a little bit about Observa and what we do. So what Observa does is we help our clients with real-time information about how their products are being merchandised at retail locations, this is unique data, that we collect across North America and parts of Europe. And our goal there is to provide brands with the means to better gain marketplace visibility and ensure that their products are being marketed to their brand specifications and according to the agreements with their partners whether that’s the brokers, their distributors, or the retailers themselves.

Hugh Holman

So, a little bit about me. My name’s Hugh Holman. I’m the CEO and co-founder here at Observa. And I’ve been working in technology, strategy, and retail for my entire career. A couple fun facts about myself, I don’t know if you guys have kids, but if you do you’re probably very aware that tomorrow is Star Wars day or it’s often referred to as a May the 4th be with you, because it is May 4th and in my household this is a very important day my kids are very excited about it. In my past, I’ve managed some large sales team including at AquaStar where we were doing about $400M or even approaching half million dollars in sales there. I take that experience along with some experience at another large retailer, Savers, the largest global thrift retail brand and combined my learnings there to create Observa with my co-founders Erik Chelstad in 2015.


So, what are we talking about today? Well we’re talking about a problem space that most brands are aware of, but I think needs a lot more attention and not only attention in the sense that we need to be looking at it and understanding it, but really focus the effort on ensuring that your company is spending the right amount of your resources and time and effort to resolve the issue.

The problem space is that brands just don’t get a return on their investment from their promotion dollars. And this is the shocker statement here: 90% of trade promotion dollars are wasted. And this is from a Bain study promotion. But you can look at studies from other large firms like Nielsen, IRI. I think Nielsen statistics, depending on the study you look at, shows 72% of trade promotions result in a negative ROI. Another one said 55% don’t break even.

The point here is that this is a massive area of large spend and a lot of companies really just ignore it and it’s not to be ignored.

Marketing Spend

So, there’s been a shift in marketing budgets over time and this chart shows it a pretty well that it used to be that brands would talk directly to consumers broadly with regular advertisements.  Whether it was TV or print, lot of print advertising. And that shifted over time. And it’s taking the dollars out of the hands of the brands, giving it to the retailer’s predominantly to buy shelf space. A lot of this transition occurred with the introduction of barcode scanning and how that changed the balance of power in who controlled information and understood the value of different shelf space. And it’s continued over time and what we found is you need to influence that consumer at the store.

The Consumer

So, the reason being that their decision-making is happening at the store. So, they are – instead of creating lists – there’s two different charts here and they kind of show this change over time and where that purchase decision is made and what’s happened is consumers are going to the store and they’re making the decision there. Ad that’s about three-quarters of the time. So, if you’re not influencing them at the store, then you’re missing out on the point in time where they’re actually choosing what to buy.

How to take advantage of consumer indecision

So, how do you take advantage of that? What are your options as a manufacturer – as a brand – to influence that consumer to purchase your product, to get them to choose yours over the alternatives. Well there are three primary mechanisms that were offered by the retailer and ones is to do a display promotion of some sort, right?

Second is to buy better positioning, we know that purchasing eye level slots or somehow convincing that buyer to put you in that position in the store is going to increase your ability to sell products, it’s going your consumer’s attention better than at the bottom of the shelf or the top. And what’s interesting when you start to dig into this is positioning isn’t necessarily the same for all brands. Here, eye level for adults is different than eye level for kids and so on. So, you need to be focused on what’s right for your product.

And then the third is to do promotional deals, right? So, and we’ll go into the different types of those here a little bit later, but promoting your product is one of the keys that I’m sure a lot of you spend money on.


So, let’s look at displays to start off with. End caps, the retailers love to sell end caps and they make a lot of money off of them. If you’re looking at the breakdown on spend, this is where the retailer’s get a huge percentage of their trade promotions fund and we know why. This is the beachfront in the store and put your product at the end of the aisle, so for the consumers that aren’t walking down the center aisle, you’re going to grab their attention being on the end, so whether you’re buying a normal end cap, it’s a sidekick, you can see in this Duplo example here on the right there’s a good sidekick and that has grown in popularity. Or you can be doing other things, like a demo, which is a little more challenging. You have a human involved, and so it’s a little harder to get organized and executed effectively. Or displays elsewhere in the store, free-standing display for instance.


So, the next is positioning. So, I mentioned beachfront property, getting to the front of the store is key, you know, and it’s very expensive. But what you’re doing is buying new customers, new customer acquisition and introduction to your product and it can be very effective in increasing awareness and by positioning your brand relative to your competition and just offering it in a different part of the store than it’s normally found where consumers are more likely to walk because frankly there’s just more traffic there. And other than that, working on the buyers to get better positioning on the shelf is key, right? And so, this is challenging I know. Often it’s the category captain’s, the key brands there, with the most market share that kind of command the best positioning on the shelf and you’re in a battle with them either having better unique data to share with that buyer on why you should be in that position or you’re opening up your pocket book and spending more money and buying somebody else out of that premium spot on the shelf which is of course very costly.


And then there’s deals. And this is the easiest place to play, right? So, whether you’re doing some sort of TPR (Temporary Price Reduction) or offering a percentage off, or dollars off, or doing a coupon of some sort that has a start and end date and maybe you’re doing this in a kind of a repeat fashion over time during the year, you maybe have a program around your TPRs. It’s a good way to get people to fill up their cupboards or freezer or refrigerator at home with your products and buy more than they otherwise would. It also helps people make a decision to try your product over at competitor.

Other ways to grab attention, maybe an X for Y, whether you’re doing a 10 for $10 up here, which means you’re probably an inexpensive product obviously to get something like that. But a 2 for $1, which is also similar to a BOGO, they’re doing some sort of a deal to grab their attention. You’re trying to differentiate from your competitors and influence that consumer while they’re standing there looking at their options at the shelf.

What products are best?

So where does this work best? Well, and this is kind of a sticky area because the worst and best for cheap products or commodities and the challenge here is that all brands need to be promoting their products. All brands need to be trying different things and figuring out what works for consumers, right? What’s going to engage them, what are they going to respond to? And so, we all have to do it. It may work better for a cheap product, something at a low price point, but that doesn’t negate the need for all brands to figure out how to engage the consumer in the store, how to attract their attention, how to get them to decide on your product over your competition. And so, this is something just to be aware of.

I think that it puts the pressure on higher price products actually to spend more of their time and pay more attention to ensure that they’re getting a return on their investment. So, cheap products do well for promotion and commodities. So, things we fill our shelves with anyways. So, things we have to have. You know, the toilet paper, water, and things like that that you’re buying every day. Something very interesting is that frozen bread has popped up and is an outlier right now as something that is doing very well in promotion. And maybe it’s because it’s a less common item, meaning that most people don’t buy it and so it’s a new item for them to find and so they’re trying it more often.

So, think about that if you’re introducing a new product. How do you increase awareness, maybe you promote a lot more out of the gate and we see the big brands do this of course as they roll out new products, lots of promotion? So, a couple that are interesting here where promotion is exceptionally challenging are beer and ice cream. So, what are we saying here? They have strong brand affinity, consumers do, and they choose products more consistently there, they have the brands that they like and they purchase them over time and it’s really hard to make them switch and so, once again, it’s going to be more challenging, you’re going to have to try different things if you have beer or ice cream as one of your categories and products in that category. You’re going to be testing on what types of promotions and how do you switch things up, whether it’s display promotion or deals, etcetera.

What might fail?

So, what might fail? So why don’t these TPRs work and this kind of goes back to that Bane shocker of the 90% failure. The statistics from Nielsen and why you don’t always get a positive ROI on your trade promotion spend. And you know, it comes down to execution. Did what you pay for happen? Did you get the position on the shelf that you paid for? Are your products actually on the shelf in front of the consumer? How do you know that, right?

And this is a problem that we find here at Observa with a lot of our customers is they pay for space on the shelf whether they’re paying the slotting fee or doing free-fill, whatever it is, and it just doesn’t show up there. And in about 40% locations we find that the products don’t make it to the shelf. So, retail execution. It’s the same with promotion. If you give dollars to the retailer to do a dollar-off temporary price reduction, is that discount being passed on to the consumer or did the retailer pocket those funds and buy down on their average cost for inventory? If you paid for displayed promotion, did you get the end caps? Are your displays setup in the store? And so, it’s really important that you’re measuring what you’re doing and that you’re paying close attention to these activities, because if with your marketing plan this is how you’re expecting to grow your sales, if there’s no execution, you’re not growing, so it’s challenging.

The other thing is your strategy. What is your strategy? So, you’re trying different things. So, you’re doing some display promotion, you should be measuring: did the displays make it to the floor? You’re doing TPRs, you should be measuring. Did the discount get passed on to the consumer? What are the results from that? Did I get the sales lift that I was expecting? And you need to try things over and over. It’s kind of like this Mario Brothers example here. Did the consumer connect with this? Does this make sense for them? Did it cause them to buy more or were they curious about it at all? So, you need to try things and finds what works and often what works changes over time, so you can’t necessarily stick with things over time assuming that they will continue to work.

Course correct

So, you’ve got to course correct. And so, sometimes it’s challenging, you kind of get in a habit and correcting the direction of the battleship is not a fast process. And I think that this chart in here which is from Nielsen shows you know you can continue to spend on your trade promotion, but it can become less and less effective over time. You need to adjust your strategy based on the results that you’re measuring. So how do you do that? So, what you’re measuring needs to be actionable. And so, we’ve been trained by the information industry around retail to use scanned data. And these are in a big four-week chunks, aggregated data, not-store specific. And we can run the strategy of our company based on that, but times have changed. And it’s not fast enough.

Things move too fast today. You need to have store specific, store level data to do anything about it, right? And you need to orient your resources at your company to be able to react. And so if you’re not doing that today, if you don’t have your internal employees, your brokers, your distributors set up so that you communicating down your distribution chain all the way to the buyer at the store about the challenges that you’re finding in your marketing and what’s effective and what’s not effective, what’s happening and what’s not happening in regard to execution then you’re probably not moving fast enough and you need to reorient your resources to be able to do that. So, you need to measure in time to be able to fix things. Which means you need to fix things quickly. So, measure early.

If you’re doing a promotional effort with a new retailer, measure to see if they did it. If you’re doing a promotion for a specific event let’s say it’s a holiday period of some sort, measure the first couple days of that event. If you can correct early on, you will fix that spend. If you can’t, at least you have data to share with the buyer on the fact that the execution didn’t occur. And you can ask for those funds to be spent later. You have an argument to make there and you have a partner if you’re sharing unique data with that buyer. So, measure early, take action early, and fix that store execution.


So, to summarize on this, a lot more purchase decisions are being made by consumers in the store. You have to affect that. You have to talk to them at the shelf when they’re looking at the products and making that purchase decision. There are many different ways to influence that consumer. So, you need to try different strategies and measure. So, you try, you measure, you improve over time.

Remember that if you have inexpensive products, commodities, it’s easier to influence that consumer. As your price goes up, it’s harder. So, you need to measure it, whether you’re a brand with cheap products or more expensive products, but it puts a greater burden on the brands that have a higher cost product or where there’s a higher brand loyalty to measure their trade promotion spend for results. Measure against their strategies to make sure that they’re working. And you need to ensure success by correcting early and often over time.

Observa information

So that’s it for our presentation today. I want to thank you for joining Observa and spending some time here with me today. Let’s see, let’s look and see if we have any questions out there. Look real quick, go ahead and submit any questions that you might have. And I don’t see any right now. Okay, well I guess that wraps it up for today. Thank you again for joining Observa for this webinar. Thanks.

Fast Data: How to use your data to influence buyers in real time Webinar - Transcript

Title Slide: Overview of Fast Data in CPG & Retail

Hello and welcome to Observa’s Overview of Fast Data in CPG & Retail. A brief introduction about Observa, we are a company that does audits and we do retail audits for folks that are interested in finding out what’s happening in real-time with their retail sales and marketing efforts.


You can learn all about us on our website or by calling us if you’d like. So, enough about Observa. Now we’re going to talk about me.

Erik Chelstad

I’m Erik. I’m the CTO and Co-founder of Observa. Before this, just to give you a brief background so you know that I am somebody potentially worth listening to, I was at Honeywell where I redesigned a lot of the aircraft safety equipment that is used every time you fly. I worked at a company called Isilon which was a local company here in Seattle that we took from 3 people to 350 in a relatively a short time, in a couple years. And I think more interestingly, I created a bunch of applications for outdoor guides and mountaineers to relay safety information back to forecasters and people that are making safety predictions down in the lowlands for other people where they are traveling in the back country or on the roads.

Fast Data: What is it?

So, today we’re going to talk about fast data. And the question is often: “What is fast data?” I think a good question to start with is, “Why is it fast?” And what we’re talking about is something that is being measured often. A good example of that is going to be oil pressure on a machine in factory. The other thing that might be happening fast is something that happens often, so this is sort of in aggregate. Something that happens often would be purchases in a chain of stores. So, as a single small store, you might only get a purchase every 20 minutes, but when you have a chain of stores, thousands potentially across the country or the globe, you have a lot of purchases happening. So, what’s making all of this fast data?

There are sensors out there on lots of things now. You may have heard of things like the internet of things and this is lots of little things reporting data on a pretty consistent basis. Sensors might be temperatures of things like a meat cooler or oil in a machine. It could be how much vibration is happening at any given moment. That might be on machine or bridge or roadway. Cameras are making this. So, cameras are producing lots and lots of data whether it’s a security camera or a camera to check on stock levels, potentially a camera attached to a drone or a robot. So, there’s lot of data coming in from those sources as well and that’s bound to grow. Then we’ve got human actions. And when you think of human actions, we have things offline like purchases or how they’re driving or where they’re walking, but also online things like where they’re clicking or viewing or reading. How much time someone’s spending on a webpage. And then finally there are environmental things that we might be measuring. So, this could be weather and air quality which is obviously something that we’re thinking of fairly often here in Seattle right now. But, what is happening in the environment that’s being measured?

Big vs Fast

So, I think one of the buzzwords that’s been going around for a few years is big data. And it feels a little bit like some of the big data is being replaced by fast data at least in the buzzword sphere. So, these images on the screen are chosen specifically to sort of draw an analogy and think about this. Big data is mountains of data. This is lots and lots of data. You’ll hear me throw different analogies today whether data is a lake or an ocean or a haystack but think of it here as a mountain. There’s lots and lots of it. And it’s kind of just sitting there. From a faraway perspective, it’s just there. When you start zooming in on it, or when you are on it, kind of going back to potentially some of the information that I was helping people collect in the backcountry, when we’re skiing down a mountain, you’re making decisions really fast. It’s the same mountain, but you’re looking at one part of it. I think it’s also important to note that the decisions your making are not just based on what’s happening right then, but they are happening based on – the decisions are part of what your training and what you’ve learned over an entire lifetime. So, we’ll come back to that idea as far as how big data and fast data interact.

Big vs fast examples

But, I think a really good example that I like, again it’s another analogy, but it’s the proverbial needle in a haystack. So, big data is I do liken that to finding a needle in a haystack. You have a lot of things and you have to scan through it and we’re coming up with different computers and techniques and algorithms to find the needle faster. But we’re still looking for that needle in a haystack. Fast data is more about you’re watching the haystack and waiting for the needle to fall into it. So, you’re there the entire. So, you know when it happens, and you notice it. I think if we just go into some of the things that are relevant to us with regard to big versus fast data, we think about financial. I like this idea because – I’m going to come back to credit cards fairly often – basically, because we all use them and are familiar with them – big data is who gets a credit card. And that’s based on an entire lifetime of credit history and maybe some other activity, but it’s based on everything you’ve done. But then, is a purchase going to go through? Is it a fraud or not? That’s an instant decision based on all the data that’s available right then. So, that’s the fast data versus the slow.

On the retail side, I think about big data being used for resets, quarterly resets, where somebody is looking at all the data – somebody or an algorithm – is looking at all sorts of data from the year and the quarter before and deciding on what’s going to happen. Versus fast data might be just in time for inventory and purchasing things because of the activity that we see on the purchase history from that day. In CPG, I like the idea of big data being we’re going to come up with a new product container design and what we’re going to do is we’re going to do studies of thousands or millions of handshapes and arm positions when they are using our product and we’re going to base our new product design on that. Versus changing a label on a can based on maybe the Twitter feed for the last 12 hours, what hashtag is popular right now.

How fast is fast?

So, fast data. How fast is fast? And I like to kind of put this into perspective at different levels of fast. And maybe you’re not as nerdy as I am, and you don’t think that a nanosecond is one billionth of a second, but it is. It’s really fast. That’s the type of thing where the nanosecond level of fast data it might be something like self-driving cars where decisions need to be made as fast as possible to apply the brakes. Maybe closer to home would be the anti-lock braking system in your car because most people have that at this point. Then if we go a little bit slower at one millionth of a second, we’re looking at microseconds. These are things like ticket purchases where we don’t want people to accidentally buy the same thing when there’s only one available. Milliseconds. This is one thousandth of a second and now we’re talking about e-commerce purchases. So, when you’re buying something online, it’s ok, it buys it, and you don’t really notice. All of these things you don’t notice they are happening so fast to a human. They are happening in real time. The next level up is the second. And back to our friend the credit card, this is sort of a credit card authorization when you’re buying in the store, or if you own a store, you know that the credit card authorizations can take a couple seconds and it’s fine. It’s just part of the process. So, there’s a scale on here if you want to look at it. But things like the PC chip, 10-9, that’s the nanosecond. The neuron synapses, this is your brain.

Real world solutions

So, now if we kind of move on to real world situations, when we think about this being that it’s a new buzzword, you’re going to potentially get a lot of people telling you you need to buy new hardware or software to work with fast data. I would be wary of that a little bit. I don’t think that most people need this when we talk about the speed of things. Again, if you’re developing self-driving cars or something else that needs that real-time processing, yeah, you’ll probably need new things. But for the most part, you can do fast data with what you have. And you may need to involve some other techniques, things like batching and buffering. That’s what you think of when you think of batching and buffering. Again, if we look at batching of transactions at the end of the day and you process them all, similar types of techniques.

Buffering, the example in the picture there is a dam where that is being buffered. That water is being kept behind and used as necessary or controlled that way. It’s very similar to that. I think another example of batching might be if you’re bringing firewood into your house or your cabin, you are not going to carry one piece at a time. You are going to batch up a load in your arms and then carry that completely, that whole load in. And it’s much more efficient.

So, edge computing. This is another buzzword you may hear every so often. It’s computing is basically the idea that you are, to be somewhat pedantic, you are doing computing at the edge. So, you’re doing it on-site. I think an example of this if we go to an extreme, let’s say we have a self-driving forklift inside of our warehouse, and what we wouldn’t want to do is wait for the forklift to communicate back to some servers in the cloud to determine whether or not it’s going to need to apply it’s breaks. We want that processing and that power, we want that to happen right there on-site. Somebody uploaded a new viral video of their cutest cats and the internet slows down and your forklift starts running into things. That is not what we want.

Pipelining. This is just kind of the idea of making sure that things are happening – the right things are happening at the right time and in the right order. I think, expanding on that, it’s a good thing for the next slide here

Save it all?

where you kind of wonder about fast data is creating a lot of data because it’s coming in fast, and do you need to save all of it? And this is a question for you and kind of your business model and ideas I would say the reasons that you would save all of it is 1) there’s some sort of legal requirement, security cameras, or hospital records, these types of things you want to make sure you always have it. But you also can use that data later for things like machine learning and finding the needles in the haystack that we talked about as far as going in and finding some other patterns. I would say that storage is pretty cheap right now and you cannot recreate the data, so do think about saving it. Think about efficient ways to save it and use it later. If you want to save money on saving, basically one of the recommendations I have is looking for significant things in your data. You don’t have to always save everything. If the same thing is happening over and over, that’s not interesting, you don’t have to save it necessarily. I think when I put on the slide purchase versus browse, the idea is if somebody is just walking through a store, potentially they haven’t done anything significant that you need to save and look for behavior patterns later. However, if they go and buy something or they stop in an area and linger and look at promotions, you probably want to record that and see what that does. That became now a significant visit.

And then of course, how valuable is your data? If you are saving lots of data for somebody that is actually making a purchase, then that guy is probably pretty valuable later because you can try to figure out what influenced the purchase. And likewise, especially if you can link that to what customer and make that data very valuable so you can aggregate their data and look back on it later, that could be very good. The picture, by the way, is a hard drive from the 1970’s that probably can store about as much data as our phones.

“Real world safety”: Techniques to Avoid Catastrophe

So, what about real world safety? And I say “real world” in quotes because the topic that I’m going to bring up is Jurassic Park. So, yeah. On the screen I can do air quotes “Real world, Jurassic Park”. So, if you remember Jurassic Park, if you’ve seen this movie or the read the book, it’s basically there was a catastrophic event where dinosaurs ran around and started eating people because they really only had like three people in a computer room drinking Mountain Dew and those people managed millions of dinosaurs on an island with thousands of tasty tourists. So, all they really had to do to avoid that was to have enough people to actually be watching what was going on. They would’ve noticed that the dinosaurs were leaping over fences or eating each other or behaving in ways that they shouldn’t have. So, the idea of keeping a human in there to watch out for what’s going on.

The stock flash crash, there was one of these – a couple of them actually – in 2010, and this type of thing happened because there were a bunch of algorithms that were basically processing, which you might call fast data, the stock market movements and they all reacted at the same time to cause a pretty major crash. Now, the stock market fortunately shut itself off. That kind of trading was halted. And I just think that’s an important thing to look for. Another thing that they could have done is those people that wrote those initial algorithms could’ve been watching for this. If I’m trading stock A and based on what happens with stock B and suddenly stock B starts crashing, so I’m going to start dumping stock A. Maybe I should look for something else. Maybe there are other inputs to be considered so I can correct things as they are happening.

I think another example in the real world is lending money to friends. We all know we want to keep our friends. And so, we also know that in general when you loan money to friends a lot of times they just won’t pay it back or they might not consider that a priority. So, in order to do that safely, you limit your exposure. You loan people money you don’t mind losing generally. And it won’t affect your friendship negatively. So, in the same way with computers and algorithms and fast data don’t necessarily let the computer control everything. An example of this might be if you’re buying inventory, let it decide plus or minus 20%, but don’t quadruple the inventory purchase or buy nothing. That would be a problem.

Another thing is the portfolio idea. I take the example of Warren Buffett and the Ponzi scheme. Warren Buffet spreads his money around and has a portfolio of things and that again limits his exposure to any single event that would cause him a problem versus a Ponzi scheme where somebody might be putting all of their money into one thing, one hope and dream. And of course, as we know, Ponzi schemes never do pay out.

Humans in the middle: Robots Need People Too

So, I think another point here is that robots – I would say that robots are only human, but the point here is that you’ve been working, you’ve managed people before, and one of the things that you’ve learned is that you can’t just let something run completely unsupervised forever. So, it’s the same for robots. They need this help. So, they might be processing fast data, making fast decisions, but you need to have humans there, humans in the middle to throw the emergency break and stop things. It’s different than a parking break, this is the thing that stops the train. So, they need to be able to halt things that are going on that are bad. But humans are also providing feedback and training data into the algorithms that are making those fast decisions. As the world changes, the algorithms only know about their tiny, microscopic view and so the humans need to be there to help train and feed them, like a pet.

The Future of Fast Data

So, what’s going to happen for CPG and Retail in the future?

I think we can look at some things and as I talk about them, remember this is sort of future and look for your own inspiration and your own ideas.

Retail Ideas Soon

But if we go to retail and we talk about things that might happen very soon, I think a couple of good examples when you look at stocking levels of things, changing the sort of assortment and high traffic areas or just in time ordering, those are things that can happen fast decisions based on fast data. Cameras, like I said before, are producing lots and lots of fast data, and examples might be: is someone about to steal from me? Is somebody about to buy from me? And if so, how can I get in there and do something? And then sort of analyzing foot traffic flow because you may want to adjust your displays on a daily basis, or on an hourly basis even based on what’s happening.

Retail Ideas Future

The farther out future for retail, I think there are some great ideas out there with things like real time pricing or matching instant coupons based on the current shopping basket. This might be somebody who picks up limes and tequila and next thing you know, they are being confronted with a coupon for a specific brand of Triple Sec because obviously they are pretty likely to go and purchase some ingredients to make a margarita. So, but that decision to present them that coupon might be based on what’s in their basket, what their buying patterns were before, what the weather’s like outside. So, those types of things that are going on. For inspiration in this, I would look to E-commerce right now. Look at the things that have been going on online for the last few years and see what you can glean from that. And also, loyalty programs. They’ve been changing, and they are becoming much more real-time as time has been going on.

CPG Ideas Soon

So, what about CPG? What is going to happen soon: production. So, when you’re looking at when should a production run happen based on real-time data coming in about the price of electricity, how much of the raw product or raw material should I order based on what’s happening out there in the world. Not just what I’m selling or what I’m wasting but think about what is actually available and what the prices are doing there. So, sort of making those fast decisions and making yourself more productive.

Distribution, there’s the idea of the just in time release of inventory says that if you have a 100,000 units of – ah, let’s go with air purifiers, it’s wild fire smoke all around the northwest here right now – air purifiers, if you’re trying to go get an air purifier, who would have known in a factory 10 days ago that Seattle was going to be a hot market for air purifiers? The computers could’ve noticed that in the upward trend in purchases and started routing more from the warehouse or the factory into the Seattle area based on people increasing their buying.

CPG Ideas Future

What’s going to happen in the future for CPG, I think some of this idea that noticing plant problems way in advance, so your equipment is vibrating 0.001% more than it has been in the last few days during a certain time when the humidity is at a certain thing, so you can actually address those sorts of things. So, you are recognizing all that data and doing something about any potential future failures in advance and getting them when they are a lot cheaper to fix. Sort of adjusting ingredients based on trends. The idea that you have a cracker with a flavor and you start adding flavors that who knew that people started wanting blueberries along with their wheat crackers. But the fast data would start noticing that and be able to start adjusting ingredients and shifting things around. Again, this is sort of future and not current.

For inspiration, I would say look to places like electric companies and utilities, airlines, they are definitely taking advantage of a lot of real-time data in order to make fast decisions with fast data.


So, in summary, fast data is not the same as big data, but they can be closely linked, and fast data can create big data. Fast is a relative term. One billionth of a second or one second are both pretty fast. Most fast data are not going to require you to buy anything new. Maybe you need to make some new types of decisions and strategies, but you’re not going to have to buy a lot of new stuff. Fast data is here now, and you could start using it to increase your sales, decrease your costs, or making incremental improvements. And finally, I would say it is a new thing and look to other industries, verticals, or channels such as other industries utilities or online for inspiration about how you might use fast data.

Final slide

As always, if you have any questions or thoughts, you are welcome to get in touch with us at Observa anytime. We’re happy to talk about what Observa does, what the smoke is like in Seattle, or just anything tech, retail, CPG… we’re happy to talk. So, thank you very much for your time and have a great week.

Blockchain & Retail: An Overview Webinar - Transcript

Title slide

Hello, and thank you for coming to the overview of blockchain in retail. And basically, we’d like to emphasize that this is an overview and it’s an overview of blockchain and an overview of blockchain in retail. So, I think the first thing that we want to talk about is we’re Observa.

Erik Chelstad

I’m the co-founder and CTO of Observa. I am an engineer and I think that this is important to note just so that we can see that I’m not just your average person on the street giving you advice on blockchain. But, I am a technology expert. I’ve written software that keeps planes from colliding and also helps keep backcountry travelers safe out in the world. Just to humanize, we always do this, but I like cooked carrots but not raw ones. I don’t know why that is. Maybe you know someone who is also like that, but it is a thing. Maybe not like not liking cilantro. But I do like cilantro.

Blockchain in Retail: Goals

So, what are we going to talk about today? Other than carrots. We’re going to talk about blockchain and really there’s three major goals and of those, I think that number 1 and 2 are sort of the foundational overview of blockchain. And really, I want this to be a good basic understanding for you and also for you to be able to talk about blockchain at cocktail parties. Number 3 is for the overview specifically for anyone in retail. That’s meant to provoke your thoughts so that you can spur conversations with your colleagues and peers or go in and do more research if this is something that you are into.

Blockchain in NOT cryptocurrency

So, to go into these foundations, I think it’s important to say that blockchain is not bitcoin. The paper that created bitcoin, that thesis and that idea, also created blockchain. However, I guess you might say that blockchain is to bitcoin as fruit is to an orange. In that, I mean orange is a fruit, right? So, a cryptocurrency is a fruit and, if we go with this analogy, bitcoin is an orange and cryptocurrency is a fruit and blockchain is food. So, I think that might be a little slight of say, but I think it’s important to understand that this is not bitcoin, and this is not cryptocurrency.

A distributed, decentralized, and public digital ledger.

So, if you look up blockchain, you are going to get a definition like this. A distributed, decentralized, and public digital ledger. That’s kind of difficult to consume, so what I want to do is break down each of these rather highly technical words that’s underlined on the slide. So, if we go and let’s talk about a ledger.

Blockchain Basics: A ledger

So, a ledger is something that you know, and it is exactly what you think of as a ledger. It records every single transaction that happens to something, the thing that you’re ledgering. Examples of this are bank records and doctor’s charts. If anything is to go wrong with a ledger, you know that immediately. When you reconcile your bank accounts, anything that you see you know when it’s gone wrong and you can immediately take action to correct that. I think we’ve all experienced that with our accounting, either personal or in business. So that’s a ledger.

Blockchain Basics: Decentralized

The point of something being decentralized. In the world of computers and the concept of something like blockchain, I think an example is – think of a centralized system – we’re going to go to a bank. If a bank computer is down, basically you can’t interact with the bank at all. You can’t touch your funds that are in the bank. You can’t give people money and you can’t receive money. So, these diagrams sort of show, the one of the left, you’ll see that all the little computers on the outside are talking to the central computers on the inside.

Another example of this is something that is decentralized. That’s represented in the picture on the right. And that is all the little computers are talking to each other. I’m going to call these computers “nodes”. And so, just think of a node as just something inside of a network. In the case of a bank computer system a node is a bank’s computer or your computer or anyone else – your phone that’s talking to the bank. Everything in there is a node. So, there’s centralized. I think that’s a pretty standard concept that we all get.

And then there’s decentralized. And while decentralized is all around examples are fairly technical, but a really good example is the English language. Basically, anyone of use that speak English can talk to each other. So, we’re able to talk one-on-one to each other and the language I use with you, my peer, is going to consist of things about retail and planogram compliance. However, the version of English I might speak to my mother is going to #1 be devoid of any profanity, but also probably not going to use words like planogram compliance. But we can still communicate, we have our own ways of talking, and there’s nobody in control per se of the English language. The rules evolve, nobody set them. And basically, the thing about decentralizing, you get a lot of benefits of this.

One is that there is no single point of failure. This is basically, if you imagine we go back to that bank situation, the bank main computer could go down because of a mechanical problem like the banks on fire, or it could be a software error, it could be a power outage, there could be attacks on a centralized system much easier because there is a single point to attack. That could be a physical attack – that’s somebody breaking in and destroying that computer, or a cyber-attack.

A good example of what we are talking about here is the internet was actually designed to not have any single point of failure. It was designed to survive nuclear war and enable communication even if that was happening. So, there’s also the idea of no fake data inside of a decentralized system with blockchain, because every single node, again a computer in the system, has a copy of the entire ledger, the thing that we were talking about. So, if there is a bank count that we’re doing a ledger of, every single node has a copy of that entire ledger. That is the one piece that is centralized in that the data is the same across all of them. So, how that data gets updated is that when a change proposition comes in every node in the system has to agree, they basically take a vote amongst themselves and if they decide that it is an acceptable transaction, then ever node updates at the same time. Otherwise, the change is rejected.

Blockchain Basics: Distributed

Another concept in there was distributed blockchain. So, distributed in this case means that those nodes that we’re talking about can be anywhere geographically. It could be on any type of system or equipment, but it will not have the same software because the technical specifications of the software has to be the same so that they can talk to each other. You can see on the graph on the left, sort of the hierarchy of a distributed network there. And a really good real-life example is on the right with a basic order chart. The CEO might be in control, but the CEO speaks to their managers, and their managers have local control over all their people that work to them. So, work gets distributed through the leaders from the CEO to the managers to the people that work below those managers.

Blockchain Basics: Hash

So, kind of a little bit of an aside here. This was not a word that was in there but it’s important concept is a hash. And of course, every time you say the word hash you have to make a joke about a greasy diner and then look at that picture and then think about breakfast and move on to what it really is, and I like to think of it as a digital fingerprint. So, anything that you have that can be represented digitally whether that’s a health record, a bank account, a picture of your kids or a song, you can make a unique fingerprint for that exact thing. And when I have a fingerprint like that, I can know when anything has been changed, because the fingerprint will no longer match what was originally recorded. As an example, if you create a word document and then pass it around to all of your colleagues for review, and they’re just supposed to read it, and it comes back to you and you compare the fingerprint to what you sent out, if anybody has changed anything along the way, that’s immediately known. So, how does that come into play with the blockchain?

Blockchain Basics: Blockchain lifecycle

And so, this is – it’s going to come in right here. This is also where the words “block” and “chain” come into the naming of this concept. So, if you look at the idea of block 0. So, this is the one in green on the left. This is the genesis block. This is the start of your blockchain. So, if we go with the bank account analogy for this, let’s assume that block 0 is your initial deposit. So, you deposit, let’s say, $1,000 into a bank account. That is block 0. Now, you make a transaction in that bank account. Let’s say you’re going to withdraw $100, because that’s all your mortgage is is a $100. And so, you go in and you make your $100 withdrawal and that is block 1. That is the transaction. And so, it gets recorded in there. And what it does is it takes that hash, or that fingerprint, from block 0, and puts that inside of block 1. And then you record your data and the time that that happened. And so on, and so on, and you do this with every single transaction that comes through in this way each of these transactions is an activity and it’s a block. And then you chain the blocks together, each one having a fingerprint of the previous block so that way you can make sure that nothing gets changed underneath the block. So, nobody can fake a transaction, nobody can insert one that wasn’t there before because the fingerprint will be different. And all of the other records will notice that across the blockchain network.

Blockchain Basics: Public

So, of the other words that was in there was “public”. And I think what’s important to know about that is that public is the record is public but not necessarily the contents. So, if I go back to this previous slide, you can see the blockchain along there is public, but the data that’s in there could be anything. It could be clear text, so “Harry loves Sally”, or it could be encrypted text. So, if this is something that you want to keep private, then you encrypt it on the blockchain like that. So, that would be your health records or any of your financial transactions or maybe even your messages to your spouse. So, when we talk about public, we are going to reuse that concept, or that word. But we’re going to reuse in the instance of, sort of, the public versus private blockchains.

And I think what is important here is that you would have a public blockchain for something like bitcoin where you are trying to get the entire world to use it freely. But a private blockchain might be an internal company system where you’re tracking parts within your company and you don’t need anyone else to see that. Or even your bank account which is going to be private to you and the bank. So, they are going to have an entire blockchain that is private to their members. And again, you cannot manipulate the information. Basically, the blockchain history is there. You can put changes in, but you can’t remove history. I think that’s a very important concept when we look at some of the things about what we could use blockchain for.

Where is it now?

Ok, so that was kind of your technical briefing and overview of blockchain. So hopefully some of that makes sense and again I’m always open for questions. My email, you have it or it’s going to be at the end of this presentation. So, where is blockchain now?

Gartner Hype Cycle(s)

Basically, what is going on with blockchain is, I think the representation is looking at Gartner Hype Cycles. If you’re not familiar with this, it’s a really nice way of looking at what’s happening with different types of new technology. So, you can see the blockchain idea in the red box and it is on its way down from the peak of inflated expectations into the trough of disillusionment. The naming on this is great and kind of fun, but it really brings up the point that we’ve been hearing about blockchain constantly in the news for the last year or so, and so it’s kind of on its way down and going towards actually reality and getting back into the mainstream.

To give you an idea of what was in that same position 5 years ago: virtual assistants. And I think that the virtual assistant market, millions of people have their own from their phone with Alexa, Siri, Google Home, so the big players have gone after that and people adopted it. Augmented reality was also in that same spot 5 years ago. You can see it again on this chart, it’s down at the bottom of the trough and disillusionment. Disillusionment. And really what we’re seeing is it probably didn’t go as far as they thought. It is not in everyone’s home, and the big players, as much as they have been trying it, it hasn’t really taken off as much as a lot of people would have hoped. Especially the people that invested billions of dollars in it.

Ten years ago, in that same position on the Gartner Hype Cycle was RFID. And RFID may not be something that is in everyone’s home, it’s not as pervasive as maybe it was thought, but it’s definitely changed a lot of what we do in retail and supply chain management. So, I also want to bring up that earlier this year, Gartner did a study on blockchain and kind of a state of where is it now talking to CIO’s of companies. Only 1% of the CIO’s said that they had adopted any kind of blockchain solution currently and 8% had had it in their short-term plans. So as much as blockchain has been pushed in the news it may not be that prevalent in the short-term.

Why care?

So, I’m just going to go over a couple – there’s a number of thoughts on here – but I’m just going to go over a couple really quick. And I wanted to point out, with any of these things, if you’re interested or if this is something that you do personally or something you do at work, just type in any idea that you like and you can then add the word blockchain to it and will probably find somebody talking about that online and get some more ideas. For instance, this next one, we look at banks – faster cheaper banking. The financial sector has been by far the biggest investor in blockchain technology up to this point. It is going to allow faster, cheaper banking and processing of payments and that sort of thing. So that is definitely on their radar and they are making it work. And I think anything else that you want to look at whether it’s going to be on the ability to control the flow of art or maybe making ticket scalping go away, those types of things. It’s out there. So, these things are important to normal life.

Retail Possible to Improbable

And what about our life in retail? So, I’m going to go over a few ideas. Again, I’m going to present a lot on the slides, for instance here:

Short-term Primary

There’s a lot of things on the slides. but we’ll just talk about a couple. So, the picture of the coach bag, the point here is that you can authenticate items. And this is in the short-term. This is probably going to happen within the next few years is giving high-end luxury items the ability to determine their origin. I think it’s probably going to start with high-end items and right now, this is actually being done inside of the marijuana industry here in Washington state where people are required to track things from what they call seed-to-sale. And so, I think it’s an interesting prospect. It’s not being done with blockchain currently, but it can be easily moved into blockchain being used for that. And then that would be a decentralized system that everyone could have a view into, and so that’s probably going to wind up saving a lot of money. So, watch for that with states integrating that in as they roll out new regulations for things like marijuana. And also, this is definitely going to take – it’s going to help out with supply chain in the short-term and you’re going to see that with regards to tracking shipments, tracking components and parts either within your organization or larger ones or across organizations that are shipping things and moving things around the globe.

Short-term secondary

Short-term secondary. I wanted to bring up the idea that there are secondary impacts of blockchain across the ecosphere of retail. And these are things that may not directly impact, but they will impact retail. So, this is the idea of money being loaned on actual sales. Because you can make your sales ledger viewable to somebody that might want to loan you money. I think an example of this right now is that Amazon does this with some of their suppliers. They definitely do it around the holidays and the reason they feel comfortable loaning money to their suppliers is because they can look and see the historical sales records for those people on Amazon – or those companies. So, again you might see new entrance into this, because you would give them a real-time view of your sales. As we talked about, banks are investing heavily. So, the banks are going to be reaching out to retailers and brands trying to find partners and to help them recognize the return on those heavy investments into this technology. So, as these people are reaching out I would recommend if it’s something that you are interested in, go for it. Become an early adopter with your financial partner but make sure that they are the ones covering all those costs. They should cover the costs because really what they are looking for are partners and usage.

Long-term primary

Now we go long-term. So, let’s go years out from now. Many years out. Ten years type of thing. Five to ten years. So, as a primary effect, one of the items that might happen is that vendors might be getting paid when they sell and not in advance. So instead of payment upon delivery or net-30, or net-60 you might be paid, if you’re a brand, you might get paid as your items sell. So, this would almost be a consignment shop. This would shift the risk from, say, a retail store to the vendor. But on the flipside of that, vendors should have more visibility into their data. That is, you’re going to get real-time updates, so instead of having to work through lots and lots of different data systems to see how you’re selling or waiting for months for an aggregated report, you should be able to get that in real-time, because you’re going to want to get paid in real-time.

Long-term secondary

So, what about long-term secondary? So, one of the effects of blockchain…again, crypto is part of blockchain… so even though blockchain is not crypto more consumers may use payments that go directly – so it wouldn’t be going through a centralized authority, that would be, in this case, lots of credit cards and as you may know it’s about 3% of the cost of operating a retail establishment. And so, if they are not going through that in the long term, the credit card companies are going to have to find another way to make money. So, the financial institutions are going to have to catch up to this idea that they can’t just pluck that many off there because they trust of the ownership of cash is going to be decentralized. Since that chain of ownership can be tracked when we’re talking about products being moved around, taxes are going to keep up with that. So, for instance, even though a vendor might not get paid on delivery, they will get taxed some of that amount. Think about the delivery as a taxable event, same way that if you exercise options from a company that is taxable even though you may not have sold your stock yet. So, that’s going to definitely change. So, the governments are going to get into the chain of ownership and they are going to promote that because it’s good to know. It’s going to get a little bit more complicated but imagine that the burden and complexity is going to be pushed out to each member of that. So, you are going to be, or your store is responsible for filing your taxes and government will even be able to actually see more into what is happening. So, with that, let’s talk about what’s very improbable.

Improbable primary

I’m just going to mention a couple of things that are improbable due to blockchain and some of these things are fun to think about. You might read about these, but it’s probably not a good idea to plan on these or invest in them. So, I think something to keep in mind is that blockchains, they are about keeping accurate records, not dodging authorities or any other responsibility that you might have. So, it’s not going to eliminate taxes, and you’re not going to be able to do fully anonymous purchases. That’s something that you can’t do now very easily and so these things like crypto are still leaving a good trail.

Impossible secondary

The secondary improbable effects is banks won’t go under, we’re still going to need financial entities, accountants won’t be going out of work because imagine you had a perfect record of all your sales, you’d still need an accountant. And taxes, you probably aren’t going to get any simpler. They may become more and more localized, but the point is there is going to be more information, so the government will be able to, again, tax in different ways but theoretically you’ll also be able to provide them with information easier. So, less of a burden to you.


All right, so just to summarize what we learned today. And here’s our summary. Blockchain is not bitcoin. Blockchain solves the problem of decentralizing trust and authority and this creates efficiencies. These efficiencies are going to be take advantage of right now while blockchain still has a long way to go but it’s definitely gaining traction in the financial sector first. So finally, the last summary point is that there are some short-term possibilities for blockchain in retail, but the disruptions are likely to show up in the long term.

Observa information

So, again, I just want to say thank you very much for attending Observa’s webinar on an overview of blockchain and blockchain in retail. And you can see the email there, sales@ Feel free to shoot me an email or give us a call at that phone number and we’ll be happy to answer any questions about blockchain, about Observa, anything you’d like to talk about. So, thank you very much.

Money & Pain: Supreme Court Sales Tax Decision Webinar - Transcript

Title slide: Money & Pain

Hugh: Hello and thank you for joining us today. This is Observa and our presentation today is on Money & Pain: A Look at the Recent Supreme Court Sales Tax Decision.


Hugh: And first of all, a little bit about Observa which you can see here on the screen. We help consumer brands and retailers measure their effectiveness of merchandising and marketing to consumers at brick and mortar stores.

Your speakers

Hugh: So, a little bit about who you are listening to today, I’m Hugh Holman. I’m the founder here at Observa and I’m CEO and we’re delighted to have Rachel Le Mieux join us from Peterson and Sullivan. Because of the topic and the nature of taxes, we asked Rachel to join us today. She is a tax fraud expert and particularly versed in this state taxation issue that’s come up with the Supreme Court. And it’s been going on now for 26 years starting with a Quill and now through the most recent case with Wayfair. And that’s really the topic of conversation today. So, thanks for joining us Rachel.

Rachel: Absolutely. Happy to be here.

A Note

Hugh: So, we’re going to start off with a disclaimer. So, here’s are note here we’re not providing tax advice. But we are going to be talking about this topic today.

Timeline of Wayfair v. South Dakota

Hugh: And so, a little bit of the timeline. As I mentioned before, this goes back to 1992, a ruling of middle court versus the state of North Dakota and in that case, they stated the retailers without a physical presence in a state did not need to pay sales taxes. So, about collection of sales tax. And of course, this has become so much more relevant with the rise of e-commerce and online sales. So, I think it’s a really timely case. Obviously coming up this year and about 2 months ago with Wayfair, but the timeline really not much happened after 1992 and it wasn’t until May of 2016 before there started to be kind of a little bit of noise around this. And finally, on June 21st, the Supreme Court finally took this case and declared that states now have the right to tax online sales. And so, they basically reversed the original case of Quill and are leaving it up to the states. So, we’ll talk a little bit more about that later.

South Dakota’s claims

Hugh: So, what was South Dakota bringing up in this case? Well, they were feeling cheated a bit. They were saying that they were not able to collect state revenues. They were watching their state tax revenues go down as there was an increase in online sales. So, people were buying more online and not buying some of those products down at the physical stores that were located in the state of South Dakota. So, they were no longer collecting sales tax on those sales and they say that they were missing out on it. They did some analysis and they came to the conclusion and put out a number that if the federal government were to look at this overall, the taxes would increase by $34 billion. So, they were talking about $50 million in their state and they were saying it was a $34 billion windfall for taxation overall.

So, and why were they saying this was an issue? Online shopping, as I was mentioning, were replacing the purchases out of their physical brick and mortar stores and they were also talking about the competitive nature of business which is often what the Supreme Court is listening to when we’re talking about business and do the physical stores have a disadvantage to the online stores because the online stores may not have been collecting sales tax and therefore were able to sell things for a lower price and consumers were being drawn to that lower price online and choosing not to buy out of the stores. They might go shop at those stores, but they weren’t buying anything.

And I just want to make a note here, and you can see it on the screen that when the government accountability office actual did their own estimate looking at that $34 billion, they came up with a number that was closer to $8-13 billion. So, what is the number? I guess time will tell.

Wayfair’s claims

Hugh: So, what was Wayfair’s claim? So, this came up again just starting in 2016 and obviously went to the Supreme Court. But what was the Wayfair claim? They said that larger companies were already paying tax in the states where they have physical location, so this has to do with something they call “Nexus”. You’re doing business in a state, therefore you have to collect and pay sales taxes to those states, so we have some large chains listed there, Macy’s, Target, Best Buy, and Amazon’s doing this too. So, for the sales that are Amazon sales, meaning it’s their products being sold by them and it’s not another company necessarily using their platform, they are paying tax on that everywhere anyways because they have warehouses in across 45 or 46 states. They are already paying taxes. So, these big companies are already doing it, but they’re claiming it’s already basically happening for a majority of the online sales.

So, the other argument is varying tax laws. So, now you start talking about the smaller business. So maybe a mom and pop shop is doing online sales, or even mid-sized businesses. Taxes are complex. And it’s complex because of how tax laws go into place. You have obviously federal laws and state laws, and that’s what we’re talking about here. But there are also local laws. So, there can be county tax, city tax, so it can make the landscape very very challenging. So, that was the other claim by Wayfair was that it would be too much for the smaller businesses to deal with.

Final “Ruling”: 5-4 for South Dakota

Hugh: So, what was the ruling by the Supreme Court? So, basically, their final ruling was that they threw out the Quill case from 26 years ago, they overturned it, and they said that states have the right to tax. So, they have the right to tax online sales where the goods are being delivered within their state. And so, in doing that, basically what they said is now this goes back to the states. Each state can decide individually what their local rules are, what their rules are for taxation. So, some states right now don’t charge a sales tax. Many states do. But when are they going to charge sales tax for small to mid-size businesses? What are the rules around that? That goes back to the states. And so, for instance, they sent this back to, regarding Wayfair, back to South Dakota saying “What do you want to do South Dakota? It’s up to you.”

Other States’ Reactions

Hugh: So, with that I’ll hand it off to Rachel and you can take it from here.

Rachel: Great. Thank you very much Hugh. That was a great background on how we got to where we got to today. You know, we’re really only 55 or 56 days into the ruling that was issued by the Supreme Court and so there’s a lot to be learned that has been learned in that time frame and a lot that still needs to be looked at and decided upon over the course of the next one month to twelve months out there. But let’s take a step back and look at what really was South Dakota’s law? South Dakota’s law which said that any out of state sellers outside of South Dakota that had at least $100,000 in sales, so that’s a value-based threshold, or had 200 separate transactions, which is a volume-based threshold – and those two things are very important to keep in mind, value-based versus volume-based – had those transactions in South Dakota had to collect sales tax from customers in South Dakota and remit that to the Department of Revenue.

So, in contrast for example, you have the state of Washington that said if you had $10,000 in sales into the state of Washington, you needed to do one of two things: you either needed to register and collect sales tax from Washington customers, or you needed to provide notice to the customers that they had a use tax reporting requirement to the state and you had to submit to the state a list of customers and what it was that they bought.

So, we have these state thresholds. So again, a value-based threshold or a volume-based threshold and it differs between the various states. So, now that the Supreme Court ruling has come down, more and more states are adopting a South Dakota style of Nexus. And again, it’s called economic nexus. But again, wouldn’t it be great if we had uniformity amongst the 50 states? And of course, we don’t. We have now over half the states that have decided to adopt an economic nexus threshold which includes both a volume-based or a value-based and sometimes it’s an “and”. Both value-based and volume.

Other States’ Reactions: cont’d.

And the effective dates of these things are all over the map and we have some slides later that will show you basically what those things are. So, again, economic nexus is now the law of the land. Some people say is physical presence gone? And the answer to that is absolutely not. Physical presence can still create a nexus, but absence of physical presence now your value or volume of your transactions can do that as well. So, the Supreme court said, “South Dakota, we like your law. And therefore, we are going to say that economic nexus is now present and ability to be the law of the land.” But here’s the thing that I think a lot of people forget, South Dakota has not actual yet ruled and created an effective date for their own economic nexus law. All the Supreme Court did was say that “South Dakota, we like your law, but we’re remanding back to your own Supreme Court to issue a ruling not inconsistent with what the Supreme Court did.” So, what that really means is that South Dakota hasn’t even implemented their own economic nexus yet. We still have a date that’s to be determined on that.

Nonetheless, other states have gone ahead and changed their laws. And this is going to be something that, again, at the very beginning of the slide we said this is not tax advice, and we’re giving you information as our best guesses, but we’re here to tell you that 55 or 56 days in and we’ve had more than half the states in the United States implement economic nexus. So, the good news is that the Supreme Court tried to say that retroactive application of the law is discouraged. Now, that does not mean that states could not do retroactive application. And there are some states who may in fact want to go back and say, “We did not have to wait on the Wayfair decision in order to believe that we had the ability to impose economic nexus.” So, there are a handful of states considering retroactive application. We out here in my side of the world, which is the tax consulting world, are hoping that they will rethink that and be happy with the fact that this is now there, and businesses are going to come forward and register and start collecting taxes.

16 states have enacted economic nexus provisions regarding sales & use tax

So, as we move through now through the next 12-18 months, as legislative sessions begin to start, we believe that more and more states are going to implement this. In this particular slide, you will see this is 16 states, but this is actually as of July 11. Almost one month ago. We actually now that it is something more along the lines of 27 or 28 states that have actually implemented. So, this changes every day, and these are here. But here you can see the states, you can see when they are going to have their lobby effective. And you can see what their various thresholds is. Remember, South Dakota was $100,000 in sales or 200 transactions.

Here, just in this list, you can see that we have states like Georgia that $250,000, you have Ohio which is $500,000. And you can see the various effective dates: Ohio, January 1, 2018. Ok? That’s prior to the actual Supreme Court case. You can see that you’ve got Georgia with January 1, 2019. Still a number of months left to go to have people be able to get set up and registered. So, again, these things update constantly. Every day, my staff comes in to the office and immediately looks at the news wires to see what other states we have to add to this list. And like I say, it’s more than 16 at this particular point in time.

Many Services May Be Taxable

One of the things that we need to make sure that people keep in the forefront is that services are also subject to sales tax in many states. That is one of the reasons why the Supreme Court agreed to accept the case between Wayfair and South Dakota is because South Dakota happens to have one of the broadest based sales taxes out there. They tax everything. Okay? So, if you are providing accounting services, if you are providing legal services, those sales are actually subject to sales taxes in South Dakota. And that was one of the reasons, just one of the reasons, that South Dakota went forward to the Supreme Court is because of how broad-based it is. And I think a lot of times people forget that certain services are subject to sales tax.

And so, what this map shows you are the states that actually have said sales taxes will apply to services. I think people are familiar with say construction related services or repair services, something that touches tangible personal property. You go out, you have your car repaired, well there’s a sales tax associated with it. But I think they also forget that things like engineering services, architectural services, sometimes consulting services related to software implementation, those things can be subject to sales tax. This is just a depiction of that, so you really have to pay attention to, and keep track of, where you’re doing business, what you are doing, and will the state say that what you’re doing is subject to sales tax.

Notice & Reporting Requirements

We still have states that have what are called notice and reporting requirements. What that means is, and Washington is one of them, Washington has implemented their sleeper statute from around 1940 that says that they can implement nexus to the full extent of the law which is now, of course, economic nexus. But they still also have a notice and reporting requirement which means if you are selling – right? We talked about this before – more than $10,000 of tangible personal property or taxable services into our state, and you have not registered to collect sales tax, you have to provide notice to your customers that they have a use tax reporting requirement. And this map depicts the states that have those requirements. And it’s important to pay attention to that because if you as a vendor fail to notify your customers that they have a use tax reporting requirement in some states, the penalties can be quite onerous. When these laws were designed, they were designed specifically to be so onerous that the states thought that the vendors would just opt to register and collect sales taxes rather than go through this onerous notice and reporting requirement. But don’t forget that those requirements are still out there.

How does this impact you? What you should be doing

But how does this impact you, okay? Well, it’s really really important that you evaluate what your business does. Where do you do business? And what do you do? If you say, “Oh, I just sell tangible personal property. I don’t need to worry about it.” Well, again, remember some states impose sales taxes on services. You need to look at your revenues by delivery state. Where am I actually doing business, okay? And what am I selling? And then you need to be able to, “Okay, what do I – where am I?” If you don’t know those types of things, then we highly consider that you have someone conduct a nexus study for you. And what a nexus study comes in and does is it says, “What are you doing? Who are you doing it for? And does a state impose a sales tax on that?” Because there are three things to think about. Is my product subject to tax? Is my customer subject to tax? Or, is the way that my customer uses my product subject to tax? And those are three things that absolutely need to be looked at in terms of how you do business to determine whether you have created an economic or a physical presence nexus in particular [inaudible 18:24 audio garbled].

So, it’s important to check those not just once every five years, not just once because I had somebody do a nexus study when I started 20 years ago, but you really need to be looking at it almost on an annual basis and maybe more frequently if your business changes frequently. And then you basically need to evaluate what you have and what you need. “Am I doing 10 sales tax returns right now and I think I can handle that with my one clerk that I have doing that?” Or, “Have I suddenly become subject to collection and reporting requirements in 25, 26, 27 states not to mention the states that also have local tax reporting requirements like the Colorado or Alabama where they have home rule where you have to file a sales tax return every single local jurisdiction you do it.

So, if you’ve gone from filing 10 sales tax returns to having to file 100, or a 150, or 200 and if that is in fact the case then potentially it’s time to think about automating. So, you need to evaluate where you currently are, where you’ve been in the past, and where you are now under this recent ruling.

What New Requirements do Sellers Have?

So, again, what do you have? You have to track. You have to track what you’re doing in the states. Not only the value that you’re doing but the volume that you are doing. Okay? And you have to track the laws. The laws are constantly changing. Okay? Like we’ve shown that one slide said 16 states when we put this together, but we know that it’s 25, 26, 27 states and it changes rapidly. You have to set up your systems to be able to collect and remit. And you have to constantly be re-evaluating. This is incredibly important. If you already have agreements set up with facilitators, like Amazon, eBay, Etsy, like any types of marketplace facilitators, you have to sit back and say, “Is it time to re-evaluate?” Because you’re are going to ask them potentially to be doing things for you. Or they are going to be forced to do this for you in terms of collecting and remitting and there are always costs associated with that. So, this ruling more likely than not is impacting more margins. And if they aren’t impacting it right now, they will. So, it’s time to re-evaluate those contracts and talk about how you can make sure that your margins are protected.

Evaluation Tools

We have some evaluation tools. This here is live on – actually not this particular tool – this particular tool there will be an email address you can email to me – it comes to me – and if you want to have this nexus evaluator tool, we will email it to you for free. Basically, what it does is we make it as current as we possibly can. This one was actually as of August 10. But I have one already that is effective as of August 14. This is a tool where you can go into the states and you can put what your total sales value is and what your transaction volume is in a state to determine whether you have triggered economic nexus or not and then it will also tell you what is the effective date for when you should be registered or when you should have registered. And after going through this tool you need additional assistance, you can obviously contact us, and we would be happy to answer any questions or help you get registered. And again, there will be an email address later on that you can use to get a hold of this tool. Back to you Hugh.

Who could the ruling hurt?

Hugh: Ok, so then the next question is who is hurt by this, right? So, small e-commerce stores might see issues with these new laws. Because they are going to have to adjust their prices. They have to account for the charge and the change in laws and their need to actually do all of the things that Rachel talked about they have to track. They have to report, they have to submit the taxes, they have to collect and submit. There’s costs associated with that. And so, how are they going to adjust their prices to accommodate for their additional administrative burden, right? And so, that’s the thing. There are about 16,000 different tax authorities out there so that’s something to be aware of.

So, the next is 3rd party sellers. So, Amazon does like a mentioned before about half their business through their own products selling themselves and the rest are the people that are using Amazon as a platform, kind of like what Rachel was mentioning with SD and so on, all those people are going to be impacted as well. And so, what does that mean, right? We’re going to find out over time, but so far, they’ve largely… these are international companies selling products into states and they’ve avoided paying it, so they’re going to have to figure out how to do it.

Consumers. So, how do consumers make purchase decisions? What’s going to happen to them? What happens when the small business raises their prices a little bit? Do they switch to the larger e-commerce business? Do they start going to a store? You know, we’re not sure completely but consumers will definitely be impacted. And then government tax agencies. This seems like a windfall, right? It seems great for the taxing authorities, they’re going to get more money, that’s an obvious thing. But it also creates a burden on them. That means that they have to actually start tracking what’s happening and be able to enforce the rules and so on. So, this does create change for them as well.

Who could the ruling help?

So, smaller brick-and-mortar stores. So, who could this ruling help? Small brick-and-mortar stores and in this case does it push people from online back into the stores to create more local purchases? It could. The big box stores. Because it levels the playing field from a tax collection perspective online with the other online retailers, you know, if these big box stores online are already collecting tax it doesn’t change anything for them. So, will they become more competitive against people that have been beating them out on price? You know, that’s the one of the things.

The large online platforms are the people that are going to be able to get the tools in place to help their vast communities of sellers the quickest – right? It’s much more efficient for somebody big to put in tax software to help thousands of sellers on their platform than it is for an individual business to put those tracking systems in place. As I mentioned before, the government wins because of more revenue. And then the obvious thing that this all filters down to the software companies that are able to track these things. So, companies like Avalara, Taxify – those types of companies that have been in this space, they’re helping people that are already dealing with nexus. The companies are already tracking this for paid in certain states. They are in a good position right now.

Will prices increase when the ruling goes into full effect?

So, will prices increase? It’s difficult to say, but what we know is when costs increase over time, prices do increase. So, there’s been a lot of math on this and the number that’s been thrown out that I’ve seen is that on average a rough estimate is that prices should increase about 1.6% for items that are sold online. So, we’ll see over time. And then what about areas, right? So, I talked a little bit at the beginning about taxes – or states that don’t actually collect sales tax. What’s the effect there? Does it drive people over borders to do purchases from their states? Maybe. I think that that happens today. Right? If you have sales tax and you’re on a border city and you’re going to a brick-and-mortar store, do you jump across the border to buy those same goods? Some people do, right? So, how much? Not sure. And so, I guess we’ll see over time.


So, let’s summarize. So, basically, the Supreme Court ruled to overturn Quill and pushed it back to the states. And so, each state gets to determine how they are going to impose their sales tax collection on these online sellers and hopefully that there is some sort of standardization but as Rachel showed us, that’s not what we’re seeing so far. Right? Whether the limits are set at $500,000 in sales or a $100,000 in sales, it’s up to the state. You know, South Dakota Supreme Court rules. Are they going to go retroactive? Will many of the other states go retroactive? We’re not sure completely but that’s all up in the air so it puts risk out there for anybody doing e-commerce today.

So, will we see a shift towards brick-and-mortar sales as a cheaper option? We might. And we’ve seen a lot of changes in brick-and-mortar stores. There’s been a huge disruption in the space. There’s a lot more people looking for buying local and if the local options become less expensive too, it might push more sales that way.

What do you need to do though? What do you need to do as a seller? You need to create a strategy for tracking what’s happening, for reporting what’s happening. And let me bring back up the point that Rachel made about services. This is a huge issue. If you are selling goods and services, the rules are per state. And so, some states might impose taxes where the services accompany the goods. Other states may have taxes just on services in general. And so, you need to have a strategy. You need a way to track this.

For More Information

So, as Rachel mentioned before, you can reach out to her and ask for that sheet to help you understand what your situation is under these new laws. Go ahead and email her at SALTgroup@pscpa.com for their Nexus Evaluator tool. We also have Rachel’s email there at the bottom and her phone number so feel free to reach out to her and Peterson and Sullivan for assistance with your tax questions and needs.

Get ahead with Observa!

And then a little bit about Observa again.

Rachel: Sorry to interrupt, if you could go back to that slide, I just noticed that the phone number is incorrect. So, I just want to get that corrected out there. The phone number is actually 206-382-7711.

Hugh: Thank you Rachel.

Rachel: You bet.

Hugh: So, once again, a little bit more about Observa. We help consumer brands and retailers improve what’s happening at the shelves. So, we measure what they’re challenged at measuring which is the last mile, how products show up on the shelf, whether companies actually get what they pay for with shelf space, with their promotions and we do this across all channels whether we’re talking about big box stores, supermarkets, drug, convenience, wherever you’re selling your products. The other thing is if you’re looking to adjust your business model, maybe you’re moving into new states like we’ve been talking about here in this session, we can help you with competitive analysis. We can get out there and find out who is on the shelves today, how would your products fit into the mix and help you with those arguments when you’re taking them to the buyer for those chains.


And with that, we’re at a wrap. And I want to thank you for joining Rachel and me today to hear about these state sales tax changes and learn more about what your challenges and risks might be associated with the changes both today and going forward and how to plan for managing for these changes. Once again, thanks for joining me and Observa today and thank you Rachel for joining us. We really appreciate your helping us with your expertise.

Rachel: You bet. Thanks everybody.

11 Top Retail Marketing Strategies Webinar - Transcript


Title slide – 11 Top Retail Marketing Strategies
Hugh: Good morning. This is Hugh Holman from Observa. I want to thank you for joining us today for the 11 Top Retail Marketing Strategies. This is a webinar in our series and we really appreciate you tuning in today.

Hugh: So, a little bit about Observa: we help consumer product goods brands and retailers measure store execution and improve the consumer experience there to drive increased sales. And we have a crowd of observers across North America and parts of Europe that help with this and we collect real-time information and retail for them.

Your Speakers
Hugh: I’m joined today with Emma Rhoads who works here at Observa. She focuses on marketing and research. And her research plays into the presentation today.

CPG Manufacturers & Marketing
Hugh: So, I’m going to start out by talking about marketing spend in general before we go into the 11 top retail marketing strategies. So, for consumer goods companies…you know, a lot of the spend, as you can see from this pie chart, is in the store. So, trade promotion and shopper marketing comprise spend of things like paying to get on the shelves – so slotting fees; paying for promotions – such as temporary price reductions; endcaps; secondary displays; you know, other things that happen there at the store and often this is because they’re parting with their dollars, giving them up to the retailer to engage in the promotion for them, whether it’s a display advertisement, or maybe getting an ad in the search bar.
So, I just want to point out that it’s as huge percentage of the spend. The fastest growing spend right now is actually digital marketing. So, what happens online, and it’s been really consuming traditionally advertising has been – it has been taking away from the traditional advertising as it has grown. It represents, you know along with traditional advertising, a pretty big chunk there. You know, 33% of the overall spend. So, when you’re looking at the shopper marketing and the trade promotion it’s not quite double, but it’s pretty darn close. And so, for consumer goods companies this is just a huge amount of spending and most of it is going directly to the retailer. So, you know, what can you expect out of this spend?

Marketing Smart and Effectively
Hugh: So, with the digital marketing this chart here really shows effectiveness. So, you know, how is it driving that purchase decision? You know, what is it doing for the shopper? How is it influencing them in making a choice and choosing which products to buy, which brand to buy? With digital, it’s really easy to measure. And a lot of the spend is transition there or you’re seeing the fastest growth at least. And it makes sense, you know. You take an action and you’re able to measure reaction, especially if you’re doing e-commerce and it’s driving an immediate purchase.

So, for online, if you’re directing the consumer or you have the ability to do that, it’s a fantastic way to not only raise brand awareness, as all brands can do online is raise awareness of their company and their products, but to drive that sale for e-commerce. But besides that, you’re going to be way more effective with every dollar if you actually influence the consumer at the retail store which is where, you know, 90% of products are still sold.

So, which of your dollars are being effectively used at trade promotion and shopper marketing in front of that consumer, in the store, where 70% of the purchase decisions are made and is your best place to spend the dollars. You just need to make sure that you are able to measure there and actually our company helps with that which I mentioned earlier. And so, it’s just something to think about as you’re deciding where to put those dollars.

Research, Evaluate, Innovate
Hugh: So now, I’ll hand it over to Emma and she can run through the 11 top retail marketing strategies.

Emma: Alright, hello everyone. I’m going to jump right in to the first one here. So, research, evaluate and innovate. It’s arguably the most important and really the jumping point for a lot of the other tips that we’re going to go over. Really, you need to be researching. You need to know what’s going on whether it’s in your industry, whether it’s the competition, whether it’s the trends going on, you need to know what is happening and you need to evaluate where your product stands within the whole context of that. And then from there, you innovate your product whether it’s a new product, an old product, whatever it might be, you need to keep innovating as the markets are changing and consumers are changing.
So, we have a prime example right here: Old Spice, a popular brand many people know. I’m sure many people might be wearing it right now. Back in the 40’s and 50’s they were very popular. They had these advertisements as you can see here right in the middle of the screen. Fast forward 50 years later and they still had the same kind of look going on. They had the same consumers, but 50 years older. And it was known as the smell of grandpas. One of those things that I saw online. So, they researched into it and figured out that they probably should change their branding.

So, they completely rebranded. Researched who they wanted to go after. Decided on a younger audience. They figured out what those people liked. Figured out his humor, the advertisements and now – also the product design you can see that all along the side there – they changed the product design. Went for some funny commercials. If you’ve ever watched the Super Bowl, you probably noticed those before. And saw truly positive results. Their sales went up and people enjoyed it. They got a whole new audience of consumers.

Know Your Consumer
Emma: So, talking about research, one of the main ways to research is really just to listen to your consumers. Anyone that has ever worked in consumer goods or retail know that consumers like to be vocal sometimes. And it’s actually really important to listen to that feedback. They will tell you what they don’t like about the product and what they do like. And listening to that can get you really far when it comes to innovating your product. So, they are the ultimate end buyers. They’re dictating whether your sales are good. They are the ones buying the product. You need to know: do they like it?

An example of this is High Brew coffee. They are a cold brew coffee company that’s been around for about 4 years now and last year they were receiving numerous people giving them feedback about how they didn’t like the sugar and how it tasted funny. So, they researched into it and found out that 1 in 3 coffee consumers don’t like Stevia. It tastes different to their taste buds. So, High Brew took this into consideration, changed their recipe, and overall it was a really positive result. They found an increase in sales and really positive responses from consumers across all media channels.

Keep up with trends, not fads
Emma: Another great thing to look at with research is trends. And that’s not to be confused with fads that you want to distinguish here. An example of a fad might be the current fad you’ve been seeing maybe in the news: raw water, where people are drinking water from creeks and lakes and not filtering it. That’s a fad, probably only be around for a little bit of time. Meanwhile a trend that’s similar would be health-conscious food choices. As you can see on the graph at the bottom there, over the last 9 years organic food sales have doubled from 20 billion to 40 billion dollars. It’s clearly going to be around for a while. People are sticking to it. So, when you’re innovating your brand and evaluating where your product stands it’s really important to be looking at something like this. Like, how does it stand within the idea of health-conscious food choices? Can you make any changes to your ingredients that would draw more people in? Or are there ingredients that people wouldn’t want in that product?

Look local
Emma: So, moving more towards actually marketing. First of all, when getting a brand out there it’s important to look local. This is one of my favorite tips to look at and talk about. It’s really taking the idea that when you start out your small brand, it’s really hard to get into a big store, but if you just do it step-by-step it can be really helpful to look local. So, there’s three different ways to do this.
You can position your product with a local story. Something that appeals to the people in the area. If you’re products made three blocks down, that’s going to sell to people in that area. People are excited about that. So, you can talk to people, tell its story, vocalize that to the area. Another way is to embrace local channels. We’ll talk about this is in a little bit and how this relates to Bartell Drugs. But yeah, just local stores that want to support you. And then also, partnerships with other local businesses that are probably more established, have a little bit more of a following, they already have consumers that come back. Maybe those people are willing to help you. They started out small too and they want to help.
So, an example that we have on here is a Seattle Brand: Joe Chocolates. And they started out making it in a little kitchen nearby in Seattle in a small house and they moved their product into Bartell Drugs which is a Seattle drugstore. They have about 65 locations in the Punchestown area. Sales went well, they helped each other out, next thing you know, they’re able to go to REI. A little bit bigger of a company, it’s nation-wide, but they’re based in Seattle and that’s where the headquarters are. Their positioning with their story had to do with a lot of outdoors…making it had to do with having an awesome outdoor story. So that really lined up with where REI was at. So, REI was willing to help them. They got their product into REI. Next thing you know, they have established great sales patterns and they’re selling their products. So, they were able to go to Nordstrom, another Seattle company and get their products there by telling the same story which improved sales this time. And after all that, they managed to go to a national grocery store, Whole Foods, and get their product in the store. And a large part of this is because Whole Foods is one of the many grocery stores that will actually have regional, local programs that look for products in the area to incorporate into the store at a regional level, and if it goes well, the national level. So, the last opportunity is to start local and build it.

Look Good Online
Emma: As Hugh was mentioning, digital is growing, so it’s important to talk about. If executed well, it is really effective, and it can impact the consumers. So, really when we’re looking at online and marketing online, you look at three things. That’s going to be who your consumer is – really important to get that down, who is it that you’re going after? Because then you decide where are they online? Online is a large place and you need to know where you can find these people that you’re targeting.

And next is how to integrate ads into wherever they are online. And there’s many ways to do this, but you need to figure that out and it shouldn’t be disruptive to the point where people are annoyed by looking at the ads. It should spark interest and generate brand awareness. Quip did this really really well with their electric toothbrush. You can see how they were targeting a younger audience, millennials, and used big words like “your iPhone” and “Warby Parker glasses”, “trends” and showed the competition next to it and it was slick, clean, showed up all over Facebook and Instagram where their audience was. And it worked out really well for them, they were very successful.

Emma: Influencers is a new topic on media. And really, it’s the fastest growing online customer acquisition method right now with most companies wanting to increase their budget. You have 59% of companies wanting to increase or influence their budget. And influencers are people who have a following whether it’s on the media or even in person. This could be anywhere from a very targeted blog about mom bloggers that have 5,000 people that read their articles every week or watch their cooking channels every single week. And using someone like that to get a very specific targeted group, or a larger one since there’s the Kardashians or someone that a lot of people follow. It’s not very specific but a lot of people follow them. And just simply having them use your product whether you send it to them for free and hope that they will or pay them. There are a lot of ways to use influencers. And right now, companies are making $6.50 for every $1 spent on influencer marketing, so this is a big thing people are looking at right now.

Smart Product Placement
Emma: Smart Product Placement. This is moving into in-store execution which is really really important as Hugh’s already mentioned. 70% of people are making those decisions at the shelf of what they want to buy. They are impulse purchases being made, it’s really important to make sure your shelves look good and your products in the store. Four things that we’re going to talk about now right here is eye-level is buy-level. It’s the whole idea you can see down in the picture with the kid with the cereal, you’ve got kids cereal at kid’s eye level while adult cereal typically would be at a higher shelf level. And that is really important.
Displays and features, we’ll get into trade promotion a little bit more later, but it’s really important to have put your product in places in the store that it typically wouldn’t be in.

Secondary placement, we see two examples on here, Mentos and Coke, for those of you who don’t know this little science experiment, throw a Mento in, it explodes. For those things that you might not go to the store thinking that you want to buy that, but once you see them next to each other in a non-traditional placement, you might be like “oh, that’s a good idea. Let’s do that.” Strawberry jam is a great example of that, it can be next to the peanut butter, next to the waffles, and it could also be next to the cheese. And it’s frequently bought with all of those items, but they are in different parts of the store. So, really placing it anywhere can stimulate those impulse purchases.
And the final one is the check-out area. We’ve all stood there. You’ve looked at the candy and you’ve bought it, because it works. You’ve got nothing better to do than look at it and it triggers that impulse buy.

Hugh: Yeah, this is something that we love to see for our customers where they’re able to put their product in multiple places in the store whether it’s getting their product into more than one section of the store on the shelf or if they’re just creating points of disruption with additional displays for instance. It’s fantastic if you’re able to get the product out from the central isle in front of different sets of consumers. And make them aware of your product. And I think as Emma pointed out, the association with other purchases that they’re making is a beautiful way for you to get them to try your product again, right?

Look Good, Stand Out
Emma: Alright. Big one, simply: Look Good, Stand Out. As we’re talking about here, people are making these decisions at the shelf. If you have the option of a hundred items, your eyes are going to go somewhere, and your packaging better look good. A great example of this is RxBar. On the left there, you can see their original packaging when they weren’t as well-known of a company, but they did a complete package redesign to the one on the right and jumped to the third bestselling wellness bar after changing it to follow those five tips you can see on the side there. To help you remember it’s SHAPE acronym. It’s just really important to tune into all of those and make sure that you are looking good and that you are standing out on the shelf and that people will try your product.
Hugh: I want to point out here, I think they did an excellent job of really understanding their consumer. You look at the package on the right, the redesign. It’s just like what is in the product. It’s a clean ingredient deck, they are showing it’s a simple recipe, there’s nothing bad in there for you, it says “No B.S.”, kind of funny. And I just think this was just so successful and it obviously resulted in a $600 million exit for RxBar as they were acquired by Kellogg. So.

Lead by a Sample
Emma: Lead by a Sample. I’ll say it again if you guys didn’t catch it, Lead by a Sample. It’s a little joke in there for everyone. Samples are very impactful. 81% of consumers are more likely to buy if a free sample is offered. We’ve all been there, you try a sample, you want to buy it. It triggers something in your head, you like it you want to buy it. And what’s really interesting is how that can be applied across beyond consumer goods to tech companies, any company. if you provide a sample and people can see it and try it themselves, they are way more likely to buy that product.

Hugh: Absolutely, and this is something that we’ve been working on here at Observa as well. Sampling is the most effective form of marketing and if you make it easy for someone to try your product they’re more likely to – I know in my household if there’s a company sampling at the store we’re way more likely to buy it. We try lots of new things because of sampling.

Trade Promotions
Emma: Trade Promotions. I promised you guys we’d come back to this and we did. If you haven’t caught on this is a pretty important thing to be working with. People spend a lot of money on it. A lot of companies are really dropping a lot of money on it and it’s really effective as we saw at the beginning when Hugh mentioned that. Using displays and features or putting that draw-back in places that maybe people wouldn’t have gone originally, and also it just draws in that attention and is triggering that impulse buy. We can see in the graph from Nielsen over here that sales for food specifically are increasing 26% when it’s featured and 38% when a product is displayed. And those are major sales lifts. You know, that’s having a huge impact in the store when executed correctly. But it needs to be executed correctly and you need to make sure it’s actually happening because a lot of times it isn’t.

Know Your Shelves
Emma: So right on the topic of that, while you’re trade promotions look good, so do your shelves. You need to make sure that your product is there and that it has many facings and that it’s SKUs are there. It’s so frequent that we’re seeing product voids where somehow that shelf tag is knocked off, there’s no product there, it’s gone. And it’s supposed to be there. You don’t want that. No missing SKUs. Or even a lack of facing.

Let’s say a company agrees to have three facings in a store, if they only have one or they only have two that’s not getting restocked as frequently as some items, it’s so easy to run out of stock that way. So, you need to make sure that all the facings are there. And then out of stocks, as we mentioned, if you’re out of stock you’re not selling your product. It’s not being put on the shelf. You need to make sure you know that’s happening. It’s a must-win grab battle. Everyone is right there picking their impulse buys, making their decisions on the shelf. Your product better be there, and it better look good.

Hugh: Yeah, you really don’t want to disappoint your consumer. So, if a consumer is going in to buy your specific product and you’re not there, they really have two choices, their choice is to leave the store and try to go somewhere else to buy it, or purchase one of your competitors, and neither is good for you. So, you’re the only one that’s going to manage the shelf, and so it’s something that you need to take on as your responsibility if you’re in CPG.

Emma: I’m going to hand over the summary to Hugh here to cover what we just went over.

Hugh: Thanks Emma. So, the first point here is manage your marketing spend. You know, we looked at the pie chart at the beginning. You’re probably spending some money on advertising. Maybe more of it’s going toward digital these days as you’re increasing your brand awareness which is awesome. Maybe you’re doing e-commerce and you’re able to drive immediate purchase and that’s fantastic if you can. If you’re mostly in brick-and-mortar and most brands with their products are, managing your trade promotion and shopper’s market is very very important. You can’t expect the store with their distributed execution and so many locations to get it right. They just can’t. We know that all the big brands manage it at the store level. I mean if you have your own field workforce doing direct store delivery, you’re touching the shelf every day, maybe twice a day, some stores maybe just a couple times a week, but you’re out there. The rest of the brands leave it up to somebody else and it just doesn’t get done. You have to manage that marketing spend and making sure that you’re actually getting results out of it, that it’s effective and that it’s getting executed. Know that you’re getting your self space, know that your promotions are happening.

And second, really know the consumer. You know, don’t let your product get stale because you aren’t listening. Make sure that you have that feedback look going and you’re listening to the consumer, you understand how they are making purchase decisions, and you’re finding out are there other segments that you’re maybe not tapping into. So, it’s not only the people that are purchasing your product today but the other people that might purchase your product. You need to make sure that you understand the consumer community and that you are able to target them effectively with the right messaging, packaging, etc. And innovate, right?

Once you have that information, understand it, try things with your products, try things with your marketing. And you’re not going to get different results, you’re not going to improve results without trying new things, right? You can’t expect things to change if you keep repeating the same actions. And really drive that purchase decision. It’s about influencing the consumer where they make the decision, whether you’re doing it online with e-commerce or whether it’s in the store, right? You need to not only make sure that your shelf placement is effective and that your packaging is popping through to the consumers, you’re catching their attention and that your promotions are drawing their attention and having them choose your product over a competitor’s and put it into their cart. Also, finding placement in other parts of the store where maybe a consumer that doesn’t walk center isle down the path where your product is on the shelf, you know, catching their attention. So, it’s really important to influence them where they make their purchase decision.

Get ahead with Observa!
Hugh: So, a little more about Observa again. We help our clients ensure that their displays are done correctly. That their shelf placement is as expected, that all their SKUs are present, they have all the facings that they should, that their priced competitively by the work retail. And that their promotions are happening. And we do this real-time across North America and parts of Europe. And let us know how we can help you, but you can’t manage what you don’t measure. And then the other thing is competitive analysis. We help lots of our clients, you know, understand markets that they’re not in and markets that they are in, right?
So, if you’re going after a new chain and you want to understand how your product would fit into their set, how are their stores different from rural areas to urban areas or between markets. We can help you understand the product mixes and what the competition is there, or in the markets that you are in: do you really get out to the stores? Do the other people in your distribution chain get out there? Do you have a feedback loop from those people? We often hear that our brokers take care of that or our distributors will. We find out that most of the time that’s not actually true and that the feedback loop isn’t working and we’re able to help with that. So, let us know if we can help you, your broker and your distributor.

Observa information
Hugh: So, thanks again for joining us today to cover the 11 top retail marketing strategies. I want to thank Emma for joining me today with this presentation and we really appreciate your reaching out if you feel we can help or even if you just want to engage in conversation about what you’re doing with your marketing and how it’s going with your placement in stores. Thanks again.

AI Overview for CPG & Retail Webinar - Transcript

Title slide

Good morning. Thank you for attending Observa’s overview of AI in CPG and Retail. And I wanted to encourage you, we’re going to go through some things today and if you have questions or comments feel free to put those into the chat here but also feel free to follow up with me afterwards, I’ll send out my email after this to everybody, so you can do that. So, our purpose today is to have an overview to AI that is hopefully useful to everyone in retail and CPG.


So, basically, what is Observa? You probably already know, we’re not here to talk about Observa today, so I’ll just briefly say that we provide real-time insights for retail sales and marketing effort. And so, you know what we do.

Erik Chelstad

Who am I? Again, this is about AI, but I figured I’d introduce myself, so you know that I’m not just your average Joe off the street to talk about this. I started Observa with my business partner Hugh Holman a couple years ago back in 2015. And, just looking at my career, I was at Honeywell for a while basically writing the code that is in the aircraft that keeps them from colliding. I think that is fairly important just in what we do and that background in safety and technology. I was on the cover of Mountaineer magazine. That’s important, I think, just because a lot of that, I spent a lot of my time creating technology that’s actually out in the mountains. I’ve worked with different avalanche centers around the country to basically collect information that will prevent avalanche exposure for back-country travelers as well as providing aggregate forecasts. I guess Observa is interesting because I took that on and I think that potentially the grocery stores are an even more dangerous place to be than back-country avalanche terrain. It’s great not being to have a joke.

Artificial Intelligence HERE and NOW

So, what is AI? What’s going on with that right now? Basically, AI is one of those things where if it seems like it’s a smart computer, it’s AI. And that’s fine, we can pretty much go with that definition. It changes all the time as our computers get smarter and we get used to things. Machine learning, and you hear a lot about this now, it’s a subset of AI. So, basically what we’re doing as computer scientists or engineers we’re feeding in as much data as possible into an artificial intelligence system and the computer’s learning from that data. We don’t tell it exactly how to learn, and so it picks things up and it makes connections that may sometimes surprise us. What’s really happening right now, we’re seeing if you look in the lower left of the slide we’re looking at things, reading things, guessing what people are going to do, and trying to have very simple conversations with people. That’s sort of the state of AI right now.

What can AI do right now? Image classification

One of the important things to go over is computer vision or image recognition. And there’s kind of two parts to this that I’m going to go over, the image recognition I think are really important for retail and CPG. One of the first concepts is image classification. In this entire talk, I’m not going to dive too much into how things happen. I’m just going to go more about what is happening and hopefully what the ramifications and implications are for us. So, image classification if you look at the images being displayed up there you can see that the images at the top level of each it’s saying what’s in the picture. For instance, container ship. It’s showing that it recognizes that there is a container ship in that picture. And what happens is, and if you look to the left of each word, you’re going to see a blue or red bar graph. And that’s giving the confidence of how high the confidence is that the computer thinks that thing is in that image.

So, this is one of the first components to what we often call deep learning. Which is this idea of feeding it lots of data, the system, and letting it learn. This is really good for recognizing things or broad concepts. You’ll notice that it could pick up something like container ship or motor scooter, but you might also have a concept like library or, given a person’s face, happy or sad. Those types of things. You need lots and lots of images for this. Thousands generally. And why do you need so many photos?

Why so many photos?

Basically, when you look at the photos – like these are examples of pictures of glasses. And we as humans can really figure this out quickly that these are all glasses; however, to a computer if for instance you trained it and you had pictures and every picture you gave it was of glasses with round frames, it’s going to really learn that it’s going to assume that all glasses have round frames so when you give it something else like the image in the lower left with sort of the square rectangular frames it may have difficulty figuring that out. And then if you think about it kind of going like we’re going reverse “Where’s Waldo” mode, and you look at these glasses and you try to see, “Okay, so I see that they’re different and why are they different?” And at first you might say the shape of the lens you might then go and say “Well, the colors are different” and then you can start looking at other things that are going to happen on here. The background is important, the shape, the angle of the actual picture. So, are the glasses is sitting sideways or are they straight on? Is there a glare coming off of a flash or a light in the picture? So, the computer really needs lots and lots of images with different contexts like this so that it can start determining when glasses are in the photo.

What can AI do right now? Object detection

The next kind of concept that is big in the image recognition is object detection. And this is the one that you’re starting to see more and more of this being shown whether it’s on a Syfy movie or TV show or just in marketing material. It’s kind of pretty cool and pretty interesting and it seems really neat because you can do things like if you look at any of these pictures you’re actually seeing which objects are in the photo itself.

So, this is a newer piece of technology this is a little bit more advanced than the classification. Classification you can buy off the shelf right now at any of the major vendors like Amazon or some of these others. The object detection is newer, and few people are starting to offer this but it’s rare right now. It’ll be state-of-the-art and available over the next 6 to 12 months. And primarily what the problem isn’t getting a data set that that works. I think the image in the middle here where you see a notebook, glasses, and coffee, you can probably find one of those off the shelf that will recognize those types of things. But when you want it to recognize your own products or your competitor’s products, you’ll need to train that yourself. And that’s the part that’s really rare right now. But the nice thing about this is that it actually locates objects. So, you can see them on the screen, you can draw boxes around them; but more importantly the computer knows what is actually in the image.

So, you can start doing things when you take things like this you can start doing – if you’re in retail you can start looking at the share of shelf, planogram compliance, and these can be measured because the computer knows each object that’s in there. You can count things. Obviously, there are limitations: it can’t see behind, it depends on the angle of the photo, what’s going on. Because whereas a human might be 6 feet high and can see into the box of candy bars and all the way towards the back and can either make an estimate as far as how many candy bars are in there based on the height or the thickness and the computer may not be able to do that, especially right now.

I think it’s also, just to go back one slide, I wanted to point out that for manufacturing processes the object detection can be very good at detecting things like labels being off-center where you can recognize a bottle and then also the image of the label and if it’s not in proportionally where it should be, it can flag an error in there. So, there are definitely ramifications to everything we’re doing

What can AI do right now? Predictive Analysis

So, if we look at it other things that AI’s doing right now, predictive analytics. So, this is what we talked about. This is a pretty common thing that’s been going on for quite a while. We almost don’t think of this as AI anymore because think of an example like our computers might know that it’s coming up on the rainy season and so the stores going to start pre-ordering umbrellas and hats. That wouldn’t happen without the data going forward in this. I mean we may go kind of gut feel but also the computers going to predict this. Again, it doesn’t even seem like AI anymore but maybe 20 years ago, this was a big deal. So, what is it doing now? We’re constantly evolving this and it’s getting better and better where it’s picking up on smaller trends. You might even see things where it’s predicting when events are happening.

So, I think of even a convenience store chain with they are going to start ordering things in advance because they know that the rodeo is coming into town. And so, they’re stocking up on certain things that people might need like sunscreen or beer. So that’s going to be happening and it’s going to go with that with regards to what’s happened in the past. I think a lot of this is future-proofing yourself for constant improvements in this system is to just get and keep your data that you have so when you do capture things go head and keep it. Data storage is getting cheap and so you should hang on to it because you never know exactly what you can extract more from that day in the future.

What can AI do right now? Predicting Single Consumer Interests

So, another thing that you’ve seen quite a bit of is sort of predicting a single consumer’s interest in something. This is – I think a classic example of this is Amazon’s been doing this for a long time. I know originally, we saw this as if you read this book you might be interested in this one. And then of course Netflix came along and started suggesting things that you might like to watch and so this is really a classic example in retail you might see this looking at the baskets of goods and looking at what a consumer’s purchasing and then suggesting coupons or discounts for them based on that. One of the problems there is that you’re only getting a small snapshot of what they’re purchasing right now and so you’ve seen a shift and I think we’ve all participated in this is in loyalty programs. And that’s an attempt to capture the entire shopping journey continuing forward for years or the lifetime of a customer. And so, with that you are able to capture an extensive dataset which is one of the things that is very necessary for predicting what a single consumer might buy.

You can infer things the more data you get about individual customers. You can start to infer that, you know, I’m a consumer and I behave a lot like my brother does or like my friend, we buy pretty much the same things. So, we can start making inferences based on our shopping habits. So therefore, if my friend likes these types of movies and we’re all watching 80% the same the 20% deviation is where recommendations might come in. Online retailers do have a bit of advantage here because it’s so easy for them to collect all of this data. So again, what can we do to try and capture as much of this data as we can and attribute it to an individual consumer. Even if anonymizing is good it’s still okay to capture the data and anonymize it if there are any surprising issues.

What can AI do right now? Sentiment Analysis

The sentiment analysis is something you may have heard of. It’s a little bit nerdier. But this is trying to capture what’s going on with people’s text. We spend a lot of time in text communicating obviously texting. But also emailing, Facebook Messenger, slack, these types of things that we do. This has been a challenge. People have been trying to do this since the 60’s or 70’s to do sentiment analysis. What they kind of created came up with a list of things and certain words are considered negative and some are considered positive. That was kind of an initial way but that’s a really difficult situation. I mean if you remember, probably in the 80s, Michael Jackson made bad good, so this is one of those things that the computers are trying to pick up on.

I like these examples we have on here. You know the text of things like, “Way to go Delta. My flight was only delayed by 7 hours” and the computer might see certain phrases and think those are good, but if you take the entire context, it’s actually a bad thing. And so, this is where AI is helping and getting much better. We can use this type of thing to look at really positive or negative comments and try to reach out to people proactively if that’s happening in social media. Obviously, we’re not going to capture those things in what’s known as “dark social” which sounds really ominous but it’s just basically private communications.

So, when you are texting your friends, nobody’s really analyzing that. There’s no computer system that is picking up and is able to send that off to retailers and let them know that something bad is happening. But you can look at things like media and social posts and it also helps you compare your brand with competitors if that’s something you are looking to challenge yourself and be better.

What can AI do right now? Natural Language Processing

Natural language processing, or NLP sometimes you’ll hear this, this is similar to the sentiment analysis in that you’re trying to figure out what humans are actually saying. The difference here is you’re not just measuring but you want to be able to interact and respond. So, this is – you’re seeing this in the voice assistance that we have going in our homes into our phones. I just wanted to say that it’s still pretty basic. It’s useful, but like sentiment analysis it’s really hard to detect things like jokes or sarcasm. And so, we humans do a lot of nonverbal communication when we speak and so it’s getting better but this is something is being worked on. I think just as an example of this, think of how many times you’ve done a voice to text and it’s if you were to just hit send you could be embarrassed or just have a good joke with your friends that you said something really odd.

The Future of AI

The future. So those are the things that are really here right now. And this part I just wanted to talk a little about what is coming. So, things to keep an eye on, but not necessarily to put all of your resources into it. I think if we look at this as something Bill Gates talked about where he said that we as people often overestimate the change that’s happening in 2 years, but we underestimate the change in 10 years. So, these things are on the horizon. Again, I would watch for them but maybe don’t – just be careful of throwing too many resources at them, unless of course you have lots of resources.

What could be coming?

So, I think we’ve been hearing a lot about Amazon. And Amazon is definitely putting resources on lots of different pieces of technology and retail and consumer packaged goods. The company is all over. But one of the things they have is this Amazon Go. It’s a store. It’s basically all AI run, so if you walk in and grab something off the shelf and walk out and you get charged for it and so the idea is this is cashier-less stores. And the thing about that it’s great it’s a really neat idea but it is going to be very expensive to set up and get going correctly and given the way Amazon reacts or behaves with their technology it probably won’t be selling this anytime soon.

So, it’s going to happen, others are going to come along and copy this and that’s going to be kind of way that goes. But watch for it. It’s really interesting it’s fun to see, but probably not anytime really soon. self-driving cars, something we hear about a lot and this is again it’s on the way, it’s getting better. It’s been around for quite a few years, but it’s definitely gotten a boost with cheaper computers and a lot of this deep learning, the computer vision tests that we talked about at the beginning is getting better so it is making it faster and easier to do this. There’s going to be a lot of legal concerns. Some of those are showing up right now as people are getting injured or killed. And so, we’re going to see more and more of that coming in. I think, again, when I was creating safety avionics it would take us at least five to six years to get a product fully tested and out the door. So, we’ll probably see something similar start happening with these self-driving or autonomous vehicles and that it’s going to take much longer to develop them get them out.

Probably also and thing to note, there’s this idea of level 4 and level 5 autonomous vehicles. Level 4 being something that can move around in a place that it has been and has basically been completely designed to work in. Level 5 is the idea that you could just take a car and drop it off in the middle of the snowy mountains and it can drive around. So, we may see level 4 soon in the next 5 years or so. But level 5 is probably quite a ways off still. So, with that in mind we may see things like obviously warehouse forklifts self-organizing and Warehouse pods or maybe the warehouse moves around and shifts itself. So, a lot of those types of things that are very well-known delivery systems but definitely we not be taking the humans completely out of the loop on this yet.

What could be coming? Self robots and delivery drones

Couple other things that I think are in the news and cool: Shelf robots going around scanning shelves and measuring this Walmart is definitely putting a lot of money into this. It’s a nice thing, it’s really cool. Again for a tech person like I think it’s really cool but it’s going to be expensive to get going, it’s going to be a little bit difficult, kind of like a self-driving cars it’s going to have to really be good at knowing the area so it’s going to be difficult to react to the things like messes, or customers in the way, layout changes. These types of things are going to cause problems and it’s going to be pretty expensive maintenance lift. So, you’ll need teams of people to do that, at least in the short-run. Probably really only be useful in some of the big box stores where there’s ample room to move around and when products have a consistent layout. Delivery drones, again neat but like the self-driving cars. We’re going to see a lot of regulations showing up. Amazon here is putting a lot of money into it, they’re starting to get moving on it, but it’ll be an expensive thing. It’ll probably be more like military is autonomous or what we think of as drones, which are not fully autonomous. They tend to have pilots working there. So that is something we’ll see especially with regards to safety concerns and regulations.


So, kind of in summary, some of the AI is here right now and you can use it to improve your business if you’re not already. And I guess my plug here is that we at Observa are using that now. We are using image recognition, we’re using Predictive Analytics, but you can use – there are plenty of other companies like us that are starting to use this, so I think it’s important to find somebody you like and work with them. The future has some really interesting things. But, for practical purposes they are still a long ways out.

So, I think what’s important to do is really think about what makes you different, what makes your brand special, and focus on that. Just like you always do but look at what pieces of technology can help you get even better and differentiate you even more. And along those lines a lot of techs will be a commodity. I think of back in the day when you would host a website, if you wanted a website, you had to pretty much host it yourself, at least that was what we nerds did. And also, if you were building a factory at one point in the past, you had to make your own machinery. That wasn’t something you could just buy off the shelf. But now you can get these things off the shelf and it’s going to be the same with AI and a lot of the robotics that are coming. So, unless you have a lot of resources watch what’s coming and get on board with things as they make sense for your business and help you differentiate your already well-known brand.

So, I’d like to thank you very much and again this is Observa and you can reach out to me at any time if you’d like. And I can talk more about technology. And also, I will again follow up with this with my email and you can find me there. Thank you very much and have a great rest of the week.

From Brand Loyalty to Retail Loyalty - Transcript

Title slide

Good morning. Thank you for joining us today at Observa for this webinar focused on brand loyalty and the current switch that’s happening, you know, with retail given the disruption in the area towards the retailers switching the customers towards loyalty to their stores. So, retail loyalty.


So, you probably know a little bit about Observa if you’re here watching the show, but we help brands measure store execution and drive sales by, you know, fixing what’s happening there in the last mile in front of the consumer.

Hugh Holman

So, a little about me, I started this company in 2015 with my co-founder Erik Chelstad. You’ll see in the photo there I just got some transition lenses in my glasses. It’s a new thing for me and it kind of makes it easy when you go from dark to light. My past, I worked for a national retailer, Savers, which was across the US, Canada, and Australia selling a thrift or used products. And also driving sales strategy and marketing for AquaStar which is a global seafood brand.

Over 20%

So, in our topic today, here’s the key, right? Over 20% of products purchased in retail are the store’s brand. And, you know, why is this happening? Why is it becoming such a large percentage of the products sold at stores? Well, what’s happened over time and we’ll get into this in more detail is that the quality has gone up and the brands are better able to match the consumer’s demand for product features and quality and their perception has changed over time as they built these brands into something more than just an alternative but something that they’re actually accepting and the retailer’s obviously are able to manage price and so they can offer a lower price with consistency and quality and they’re getting greater loyalty. So, where did it all start though?

What are private labels?

It started back in the 70s, and some of us lived through this and remember it with generic products that we would find at the grocery stores. These basically were predominantly just white packaging with some black labeling on them and they gave a low-cost alternative. And during the recession there, is when these showed up as a lower-cost alternative to the national brands at a time when people had less dollars, fewer dollars in their pocket. You know, it made sense and some consumers switched over to this and so let’s look at that look at that a bit. Or we’ll look at that a bit in a minute.

How are Private Label Products Made?

So, let’s talk about what they actually are. So, there’s some confusion by some as to how private labels are actually created. They think that maybe the stores are actually developing these products. That that’s not the case all. The stores are not in the business of producing goods and so therefore what they’re doing is they are relying on, often, their partners that are delivering their own branded products to them. So, in this example here we’re showing that Cheerios are made by General Mills and then, you know, they go to Trader Joe’s and say, “Hey, can you put Cheerio’s on the shelf?” Trader Joe’s says, “well, you know, that’s not really our store format. We need you to pack it in our label. How about we’ll call it Joe’s O’s?” And if they come to an agreement to do that, then General Mills packs Joe’s O’s off of their production line. Maybe it’s even the same recipe, you know, sometimes it is; sometimes Trader Joe’s or the retailer will ask for a change in the product specification you know maybe they alter the amount of sugar or salt, you know, some of the ingredients in the ingredient deck. Or maybe it’s exactly the same product just packed in the store’s brand and then Trader Joe’s puts it on the shelf to sell to consumers. What it does for General Mills is it allows them to keep their factory operating. It means that a competitor is not producing that product for Trader Joe’s, but what it doesn’t do is help them build their brand, build the Cheerios brand at Trader Joe’s. And Trader joe’s is an interesting example because they have a high percentage of private label and we’ll talk more about that later, but this is true at all stores were it’s a challenge for brands to decide what to do.

When are private label products more popular?

So, here’s the history, right? So, when we go back in time to the 70’s like we were talking about before, you know, these generic brands were developed, and you see this in the graph in the lower left. And there was a huge upswing there you can see in the recessionary periods where consumers would switch to those generic brands because they offered an alternative product at a lower cost. But then what happened over time is these private labels became true brands really competing, you know, where they had your brand names, they were just generic where they developed call outs on their packaging to speak to features of the products and you can see in the upper right graph there was another major shift towards private label in this most recent economic downturn after the financial fall out where you can see that increase and move to private label and it really is kind of plateaued after that so people didn’t switch back like they did in the 70s, and it appears at least over this most recent period that people were staying with private label and that is a very interesting trend meaning the consumer brands are having more to compete against.

Recent Shift to Quality

So, it all comes down at quality and the ability of the stores to engage consumers with these private label brands and have them consider them as a true competitor to the national brands. And another thing that’s very interesting that we’ve seen is that this isn’t unique to the United States. Sometimes we think of our economy, it is the largest in the world, being very unique in a lot of ways, but what we’re seeing with private label is this same message, this same pattern, is occurring everywhere around the world.

Private Labels in the Digital Age

So, there’s one other aspect that’s very interesting that we’re seeing in the digital space, online. And, you know, we’ll talk about Amazon here because they’re the dominant player in online retail and what they’re doing in transitioning this consumer shopping patterns and how people are searching for and buying things is away from the brand name completely. And it’s not that they don’t have private label brands, because Amazon does, they have Amazon Basics, they obviously bought Whole Foods now, 365, you know, online as well and many other brands and departments, you know clothing and so on.

But what we’re seeing is it how they are training the consumer to shop for products is not necessarily around searching for the brand name but instead searching for description or features and then looking at rankings on Amazon to choose the item that they’re eventually going to buy. So, they’re training shoppers away from shopping on brand. And it’s very interesting. Something to be aware of and if you are a consumer brand, national brand, or you’re selling your product around the United States under your own brand name, this is something you need to be aware of because you need to be educating consumers about the features that are unique to your brand. And so, will get into that more in a bit.

Who’s Doing it? – Target

So, who’s doing private labeling? It’s kind of across the board. I like this Target example because there producing private label in so many different departments in their store across so many different categories of product and if you look at the matrix in the lower left you see all these brands. I mean, and these are real brands. This is so far away from the generics of the 70s and, you know, this is their current offering but if you go to their website you’ll see that they have three new brands that they’re introducing this year. And so, this isn’t sitting in a static state they’re advancing their use of private label and the number of products offered in their store under their own brand. This helps them differentiate and drive consumers towards buying only from them. Because if you start shopping for their store brand you can’t get anywhere else.

Who’s Doing it? – Walgreens

So, another example is Walgreens. And Walgreens has other private labeled brands. They have their own Walgreens brand, in fact. But what I like about what they are doing here, is that they’re doing segmentation. So, this product, this nice brand, they’re creating three tiers of product. They are segmenting the consumer into a value-oriented shopper, a premium shopper, and then somebody that is more health focused, you know, looking for organic. And so, they’re taking one brand and tiering it and segmenting the consumers with it. And so, this is just another way that private labels have been taken to a more advanced state and this is what national brands do and they are doing this with their own store brands.

Who’s Doing it? – Costco

Costco is a great example of converting consumers to store brand with, you know, Kirkland Signature and you see it throughout their store. And you know, we have an example up here in regard to a bar and its obviously Kind is the brand that they’ve been taking market share from here. And I’m not sure if Kind is packing the product for them. If Kind is manufacturing the Kirkland brand fruit and nut bar, and if so at least their manufacturing plant is being efficient. If not, then they’re just losing shelf space to somebody else’s packing it for Costco, but these are the types of things you have to watch out for. And how do you create that differentiated product? How do you innovate to compete against, you know, these brands from the store?

Who’s Doing it? – Trader Joe’s

Trader Joe’s is interesting because they’re basically a store of just private label. And it’s not to say that they don’t have other brands of product there, but that’s really their focus and when you go sell with them I’ve been sitting with them, it’s just a conversation about, you know, can you pack this under our label? And they’ve done some interesting things with their internal branding where they kind of a mix-up the Trader Joe’s brand or kind of changed it slightly for products from different countries, you know, you can see the olive oil here from Italy is Trader Giotto’s and you know the Chipotle is Trader Jose’s. I think that’s cute. And it’s just a high percentage of the products they sell and they’re able to you know differentiate their store offering considerably from anybody else because of this private label.

How can Brands Compete?

So really the question is how do you compete against this, right? And it’s a very big challenge, you know. It’s hard enough especially for up-and-coming brands to compete with the national bands anyways, you know. If you don’t have DSD with your own people in the stores managing what’s happening there, making sure that your execution is what supposed to occur it’s hard right, so you know when you’re working with the buyers how do you get the right location for your products? How do you get on the shelf the way that you want to be presented for your segment, right? Are you getting at eye level? Are you able to get additional points of disruption in the store whether it’s, you know, end caps occurring, display promotion? Are you working with other brands where your products are complementary so that you can be positioned next to their products in the store and doing co-promotion? You know, those types of things that allow you to be unique to the consumer and call out to their attention, you know.

And I think it shows right here that 18% of shelf space is private label now, you’re competing with the store to get your product on the shelf against their products. And you got to make sure that that’s happening and that you once you get an agreement that what you expect to happen is actually occurring. And so, then the other thing is in speaking to the consumer, how do you attract the retention, right? How do you innovate in your marketing and your packaging and understanding the consumer and speaking to them with your products, with call outs that are meaningful to the and stay current in the trends in the market and making sure that your packaging stays current for your target consumer segment?

You got to innovate. It’s innovation in your marketing. It’s innovation in your product. What are the trends around flavors and, you know, what’s coming out of, if you’re in the food space, what’s coming out of food service in the restaurant space so that you can invent your product there; if you’re in fashion and clothing what’s happening with the premium haute couture producers and how do you include those changes in your products; and consumer electronics, the same type of thing, right? You have to differentiate. You have to be different than these store brands.

How Can Brands Compete?

So, an example here is Philadelphia cream cheese. You know, they are interesting in the market because they are really the only brand. It’s a billion-dollar market. They have 70% of the sales and they really only compete against private labels. And so that’s why I think this is a unique example. So, they have to compete. So, every store you go into, there’s a private label to compete with Philadelphia. And yet, how did they do it? So, they came up with something unique with a campaign that started with marketing the real women of Philadelphia. They create a website. They differentiated by creating a community around their product and the people in the community developed recipes. And what did that feed Philadelphia? If that information about how their products being used in the market from the actual consumers is really cool. And they are able to use the messaging from those consumers that they’re marketing actual videos submitted; recipe submitted.

It’s true engagement in consumer development in what they did, and it gave them the information about what products to produce. And so then they’re able to be unique in the markets with new products that they’re developing and differentiate from the store brands by bringing on new products all the time as they’re hearing their consumers, hearing what they’re saying, as they’re delivering new recipes, delivering video content from them they’re looking at that, they’re mining it for what the trends are, what the demand is for and developing new products. So, they stay current with their consumers and I think that’s just a beautiful example. Now, I know that that comes at a cost and not every brand has the marketing funds to be able to do something like that. But I think it’s a good example and you can do it on a micro level as well.

Quick poll

So, the next thing is, you really have to own the space that you get in the store. You can’t assume that the store has your best interest. And you need to make sure that your space is being managed and you have to measure to do that. And so, are you getting the shelf space that the buyer is promising you? Are all your products on the shelf, right? Are you at a price that’s competitive to the consumer? If you’re doing promotion, is that temporary price reduction being passed on to the consumer to give you a lift to attract new customers acquisition? If you’re on end cap, did it get billed out? If you’re doing other disruption in store displays, hangers, whatever it is, you know if you’re product is supposed to be at the cashier, is it really there? Because if it sits in the back room, if it doesn’t get executed on the floor, you’re gaining nothing.

In fact, you have all the cost that go into those promotional efforts from your marketing team your design, you know, building, shippers, whatever it is – you have to manage that at the store because you’re competing with, you know, national brands that have their own staff doing DSD. You’re competing with the store which is the same situation as their staff making sure that they’re getting their shelf space, that their products are out on display, that their promotions are happening the way that they’re planned, right? You have to do the same thing. So, this is very key here. Differentiate innovate, manage your store execution.


So, a summary from what we talked about today. Private labels are growing. It’s just a fact. You know, with the disruption in retail, the stores are having to do something to manage the shopper behavior to manage attraction to keep them coming to the store. They can keep them coming if they attract them to their own store brands. You have to innovate, you know, understand the consumer, understand changes in the market, change your products, come out with new products, advance your offering, advance the marketing, the packaging, make sure that you’re speaking directly to those customer segments so that they will choose your product when they’re there in the store.

Remember 70% of shopping, the decision happens there standing in the aisle looking at the products on the shelf speak to those consumers. You have to stand up. And then, remember that retailers are working to convert all those consumers in the store. Your loyal customers to retail loyalty. You know, to be loyal to that retailer away from the brands. And so how do you keep your customer? Well, you better be on the shelf. If your consumer goes into buy your product and it’s not there, whether you’ve been pushed off the shelf, you don’t have the real estate you think you have, your slots are missing, or whether you’re out of stock, they have two options, they can leave the store and try to go find your product in another store, and that does happen sometimes. Or they buy an alternative product and it’s going to be one of the normal national brands that you compete with, another consumer brand, or is going to be private label. So, you need to manage the store execution for your products. You have to take responsibility for that. Nobody’s going to do it for you.

Get ahead with Observa!

So how can we help? So, a little bit about Observa. We help a lot of companies with competitive analysis. Understand what’s happening there out in the market, what’s happening with the brands you’re competing against, whether they are the store brands. And we go do competitive analysis across the country and all different kinds of chains. Gathering information so that you can take it to the buyers and better articulate your differentiation and how your brands and products would do better on the shelf than your competition and it’s really smart to go take unique data to a buyer when you’re going to go meet with them. I mean, they have tons of data from their store systems. They get market data.

What can you take to them that’s unique? And by doing competitive analysis, you can take them information that they don’t have. And then manage the store shelf, like I said before you need to insure your box to where they’re supposed to be. You need to make sure that you’re managing that shelf space. Nielsen says that planograms will grow at 10% a month. You know, and we find in about 40% of the locations we measure there are problems with our customers products being on the shelf and that’s a really high percentage. And so, we help them with that unique information and fix those problems. And often it’s in conjunction with the buyers by making them aware you can help them, you know, by holding them accountable for what they’re responsible for too. To work with the store managers and so forth. So, we help companies measure store execution and the competitive landscape.

Observa information

We love to help if we can. I want to thank everybody for attending today. If you have any questions please submit them in the chat, but you can email us at sales@ We look forward to speaking with you about your unique situation, about what’s happening in private label for you. Maybe the buyers are asking you to pack private label. I know in my past, that’s been something that’s I’ve run into where you go in with a brand new product and you’re trying to sell in beside the other product you have with a retailer and they ask you can you pack this private label for me. I’m not going to give away my innovation straight out of the gate. It’s about building your brand. That’s where the real value is. You want those sales. So, tell us your unique story. Send us an email. We’d love to talk, we’d love to help you. Thanks for attending today.

Avoiding Voids Webinar - Transcript

Title slide

Alright, I think we’ll go ahead and get started here today. I want to thank you for joining Observa. My name is Hugh Holman, and I’m the CEO and co-founder here at Observa and I’d like to tell you a little bit about our company first and our presentation today. Our presentation is on avoiding voids and we’re going to talk about real-time analysis in action and how to get real-time information at retail store level.


So, a little bit about Observa. So, we provide real-time insights into how your products are actually being perceived by consumers at retail locations and we do this through retail audit. We have mobile apps and thousands of people across the country to collect information on your products. So, or mission here is to provide brands, you, with the means to effectively gain greater marketplace visibility and to ensure that your products are actually where you expect them to be.

Hugh Holman

So, a little bit about me: I spent on my career and technology and in retail. I was the chief strategy officer at AquaStar and ran sales and marketing and that organization was doing about close to half a billion dollars in retail sales. I was also with brand strategy for Savers which is the largest for-profit thrift store chain in the world with 340 stores across Canada, the US, and also Australia. I spent some time and in the financial services space in banking and technology services and so on. A fun fact, last week I was one of the hundreds of thousands of people that were caught in the CES blackout in Las Vegas. It was very interesting to see them run out of power and in such a large trade show. We founded Observa years ago and have been providing value to our customers since then growing our base. In fact, this company grew at 5.7 times our previous year’s earnings last year.


So on to our presentation on voids. So, let’s start off with what a void is. Kind of to start off a really base level, when you go and sell to a retail chain you’re expecting your product to make it to the shelf, but it doesn’t, so you get into some sort of a commitment. Maybe you’re paying slotting for that real estate to get your product on the shelf in a certain number of stores. And you’re often buying for a family of products and so maybe you have 3 SKUs, 6 SKUs, 12 SKUs that you’re expecting to make it onto the shelf and so you’re expecting slots or facings for each of those and sometimes it’s multiple facing per product. And in the case where you go into a store, your customer goes into that store, and they’re unable to find that product on the shelf: that’s a void. And so, it’s where you expect your box to be where it’s not.

Different from Stockouts

So, this is sometimes confused with being out of stock or having stockouts. A void is where you don’t have a slot on the shelf, so there isn’t a shelf tag, there’s no real estate or space for your product and this is different than of course having the slot allocated or the space there, but not having any inventory on hand for the consumer to purchase. And these are the number one and number two issues that our customers run into: voids and not having the product there for the consumer to buy is the biggest restrictor on our customer sales and number two is not having adequate inventory on hand with stockouts.

Where is your product?

So, if your product isn’t on shelf in the store, if you haven’t been slotted at the retail location, where is your product? And so, this could be a couple things can happen. So, one is that it was never put on the shelf at all. Your product is sitting in the back room there. And the execution at that store level just never occurred. Second is that your product made it to a central warehouse and distribution warehouse and it just never made it to the store. So, it wasn’t ordered in for that retail location and so it could be at the warehouse, it could be at a regional distribution center. The other thing is that it could have been ordered in and then if it wasn’t re-ordered, you’d end up with a blank space on the shelf and then it stopped being ordered in. So maybe it was ordered in once, they never had any re-orders so it’s not being ordered any longer. So, your product, basically, is not in that store system any longer.


So, what is the cost to you? Well, I think that you know that it’s extremely high. So, it’s very challenging getting space in retail stores as it is, and you spent a lot of time and energy. Sometimes you’re selling in to an account for years before they’ll place your product and so it comes at a high cost to not actually get that space. And so, for most of our customers, the highest spend that they have is around trade promotion and shopper marketing and buying shelf space or slotting can be the highest cost. They may spend more on getting shelf space than they do on advertising for instance. And so, shelf placement is often the number one marketing vehicle to engage with consumers and let them know that your product even exists. So, it comes at a high cost for it not to be there in front of the consumers. So, consumers then can’t purchase your product, so with that they’re not having the experience that you’re expecting them to have and if they’re going into the store looking for your product, not being able to find it can cause confusion and ill-will and so it can be brand damaging.

The other thing is a lot of consumers will choose an alternative product. So, they’ll choose one of your competitors to buy so that’s going to give them a leg up relative to your product and brand in the marketplace. Also, if your product is not on the shelf you won’t be getting any reorders in this is actually where most of our customers find the problem. They do projections on orders, they find at the re-orders aren’t coming in at the levels that they expect and that’s how they know that there is a problem out there somewhere and they’ll engage us to go find out which stores don’t have their products slotted on the shelf. And then ultimately, if your product is not at one store, that’s not much of a big deal, but what we find is this occurs in about 40% of the locations that are customers will find that one or more of their products aren’t actually slotted on the shelf and the problem is that the buyer is measuring back at the central headquarters and there not looking its individual store activity, they’re are measuring how well your products sell relative to your competitors for their category. And if you’re not selling to the levels they expect, you risk losing your slots in all store locations.

And this happens all the time, and it’s quite sad because you’ll go through that sales process and it might take you years to get places on the shelf. Maybe that you finally get added to the planogram or the shelf design which usually refers to most retailers twice a year and then, you know, a few months go by your sales and reorders don’t come in at the levels you expect, and you don’t know why and then your slots get pulled.

How does it happen?

So why does this happen? Well it’s really pretty simple. Retail chains, you know, make decisions at a headquarter level and then all of the executions actually get distributed out to the stores so those work orders, if you will, make it down through a store’s management chain and so the store manager might get requests to make changes to the shelf, add products and then their category manager might become involved. Ultimately somebody, some hourly worker, is going to make that change on the shelf. Does it happen or not? And so, it’s really easy for it not to happen to a certain percentage of the stores. Maybe all the stores don’t have the same layout if they are not what they call built-to-suit where each store is exactly the same, so they constructed them all the same. Lots of stores go into existing locations, they don’t have exactly the same format and they might run out of space in the category, right? And so, you don’t want it to be your product that doesn’t get slotted.

And then, the next thing is, are they paying attention to your product? Why is your product important, right? And so, if they’re not tracking sell-through for your item maybe they’re not re-ordering it and then over time it’s very common for unused real estate on the shelf to get filled with something else. And it can happen because people in the store are trying to do the right thing for the store which is make sure that all slots on the shelf are filled for consumers. Or it could be a competitor, and this happens where people that do direct store delivery for instance, maybe it’s a larger CPG where they have their own work worse in the field and they notice that, you know, those slots haven’t been filled for weeks and they go and put their own products in there.

How to detect?

So how do you detect it? This is a really a big problem, and this is a lot to do with what Observa does. You know, it’s hard to do. How are you going to get people out to the store location? It’s quite costly to hire a field workforce, and even if you do if you hire one person in a city how do they make it into hundreds, maybe thousands, of points of distribution to find out what’s happening with the products. It’s virtually impossible. And the other thing is maybe you have data, maybe you subscribe to scanned data, maybe you get sales data from specific store chains. You probably don’t have it across all store chains, probably not down at an individual store level and so, what Observa does is we uncover where these problems are, and we give you quantitative data and photographic evidence and that’s one way to go about finding out but otherwise you have to send your own people out to the stores.

Why Scan Data Misses Voids

So, let’s talk a little about scanned data. So, if you do subscribe maybe you pay IRI or Nielsen for a subscription to their data. You probably get it delivered every 4 weeks and so you get a clump of data that’s aggregated. It’s unfortunate that it’s not down to a store level. You might subscribe to national data which tells you how you’re doing relative to competitors per item. Or you might have regional data which tells you a little bit regional preferences and maybe different flavors that you have, or varieties of your products sell better in one region over another which is great information to have but it really doesn’t tell you what you can do at a store level to drive an increase in sales.


So then if you are measuring, how often should you do it? Well I mean ultimately it would be great to be able to go out multiple times a month. If you send somebody to a store once a week, that’s fantastic. And we have clients that have very high velocity products, they’re able to do that and they have us go into the stores and we get wonderful data because we’re measuring week after week after week and we’re able to tell what the issues are in the trends over time and impact those trends. You can get similar value out of doing monthly measures. And we do that for a lot of clients. And we believe that this is kind of the best practices for most companies. And it really helps you ensure that you’re getting all the real estate you expect and that you start to uncover other issues and deal with them. You have to measure you know it’s be able to manage things and understand whether you have voids, whether you have out-of-stock issues, or maybe a misallocation of slots to the wrong product, so those types of things. We can help uncover those.

How to Remedy

So, how do you remedy it? Well, you know, the first step is knowing. You can’t fix something that you are not aware of. And ultimately if you have data, then you can influence the fire. And this is really what you’re trying to do. If you have an uneven even playing field with buyers, they have an immense amount of data because they have all the scanned data at the store level for all the items in the store. It’s unfortunate that they don’t share that more readily, but they have that data. And so, if you don’t go and share unique data, you’re at their whim and will.

And so, what our clients often do is collect information at store level, aggregate it, slice it and dice it and see what the issues are. And then they can go share that with the buyers. And so, it’s not an anecdotal response that maybe our products not on the shelf in some of the stores. You can say, “My product, my primary SKU is not in 43% of the stores in here and here’s the store list. My secondary product, SKU 2, is not in, you know, 30% of the stores and here’s the store list.” If you can give that kind of information, that data, your argument for getting more shelf space, for adding new products to the family that you have on the shelf, is so much stronger than if you go in with anecdotes. Present it to the buyer, share.


So that kind of concludes our presentation today. I want to thank you for joining us. Hopefully, this has been beneficial. I think understanding the retail landscape and learning how to better manage your relationships with the stores which are your partners for sharing your products with the general public and is very very important and having dated to be able to manage your business is I think a concern for most people. And this is where Observa is able to help, and we would love to be able to help you with your challenges. Let me see if we have any questions in our chat here. I’m not seeing any right now. I guess, please feel free to email us. You can reach me at hholman@ Just to reach out for sales department at sales@ Please feel free to reach out, we would love to help you out with any store execution issues or just help you with your general strategy for 2018. I want to thank you for joining us today.

Knock Out Your Stockouts - Transcript


All right we’re getting close to the time to start here. I want to thank the people that are on already. We’re going to wait just another minute here till 9 o’clock Pacific time before we start.

All right, it looks like it’s 9 now, so we’ll go ahead and get started for today. And I want to thank everybody for joining us and Observa on our topic of knock out stockouts. And so, we’re happy to be sharing a little bit about Observa and our knowledge on the retail space. I’ll go ahead and start in on our presentation today.


So, I as an introduction I’m going to tell you a little bit about Observa. So, we help brands with real-time insights into their sales and marketing efforts and we do this by providing store specific insights and data including photographic evidence on how products are actually being merchandised in stores. And so, as part of our mission, this gives the brands that seek this unique information the means to gain greater marketplace visibility and grow their sales. And so really that’s what we’re about. We’re helping brands with information that helps them win in the market.

Hugh Holman

So little bit about myself, my background, I’m the CEO and co-founder here at Observa and I have a background in technology management and business strategy. It started with joining an artificial intelligence start-up when I was still in college and working through various positions in financial institutions and technology organizations as well as CPG and senior positions like chief information officer, chief strategy officer and I found myself at AquaStar which is a reasonably a large player in the seafood business running strategy sales and marketing. And that company at that point was doing nearly half billion in revenue. I also happen to be an early investor in and on the board of a company called Rad Power Bikes. Kind of an interesting fact, they happen to be the largest distributor of electric bikes in the United States now. Of course, we’ve expanded to Europe and Canada. And I co-founded Observa here in 2015.


So now on to our topic of stockouts. So, what is a stockout? Well, they are also referred to as “out of stocks” or “OOS” if you’re reading articles about it, but it’s basically where the product has sold through in the store. And so, it’s a problem for consumers and for the brand because they’re unable to buy your product at that point.

Different from voids

So, this is to be differentiated from voids. So, a void is where you have a slot, or where you don’t have a slot on the shelf. So, if you look at the image shown on this slide you can see that there’s missing product on the shelf. Where there are shelf tags in their space allocated for your items that’s a stockout, that’s where you’re out of stock. And the voids are where you no longer have any real estate in the store, so shelf tags are missing, there’s no more space allocated for your product. And that’s a big issue as well. These are the two largest issues that our customers face that reduce their ability to sell product and lower their sales capabilities over time.

How Does It Happen?

So how does this even happen? So, there’s a breakdown with retailers in the fact that you know all the work is distributed. Decisions tend to be made at headquarters, but it’s handed down to hourly workers in the stores and so the main issue is you usually don’t have enough space for your products and so when they’re restocked on the shelf that inventory sells through and then there’s no longer any space, no longer any product there for a consumer to purchase.

And so that leads to the next problem which is that the shelf isn’t being restocked often enough. So, the first is that you don’t have enough space, second is that it’s not getting refilled and then the third thing is reorders. And this is a real problem space because reorders are often determined based on sell-through. And so, if you sell-through 100% of your product and they’re seeing that that’s the amount of inventory that you’re selling they often don’t take into account the amount of product that you would have sold during the period that you were out of stock and so they’re basically reordering based on history. And unless you make an argument for increasing the size of the order they will continue to do that over time. And then lastly the warehouse doesn’t have enough inventory and the problem there is that stores may be ordering enough inventory but they’re getting partial shipments from the warehouse.

So, these are all places where if you don’t have information on the problem you may not be able to figure out what to do about how to attack it to solve the problem. So, the other side of this is once your product is sold through on the shelf what happens with that space. Well, if you’re lucky the store might put some of your other products in its space. So, let’s say that you have a half a dozen, or a dozen, flavors or items in a family of products. Maybe they’ll fill that empty slot with one of your other items. Well that’s great if they do that but sometimes a competitor will fill that space as well.


What is the cost of stockouts? Well, it’s really high. For many brands, their shelf placement is the marketing that reaches the consumers. I think the statistic that Nielsen publishes is that 13.5% of revenue is the average spend for shopper marketing and trade promotion spent by CPGs. And so, what we know is the smaller CPGs spend an even higher percentage, so buying that shelf space, spending money on promotions in the store is a very high spend. More than advertising for most brands. And it’s your best marketing. It’s what allows most consumers to see your products. You need to have your products on the shelf. So, the next cost is that your customer might be upset, they go to this store, they can’t find your product and it can be brand damaging because they might be a little ticked off. At the same time, they’re not able to purchase your product so it’s not driving sales, it’s not driving velocity on your product and they may end up choosing an alternative and you can see that quote down at the bottom that about 23% of the time a consumer would choose an alternative product, your product is not there to purchase. And then that leads to smaller orders, reorders, from that store.

Quick poll

And then ultimately the big risk here is that you actually lose shelf space. And if but you’re not selling as many units as expected by the buyer, you know, maybe you’re not meeting the projected sales, you can lose one slot for some of your items or you can lose all your slots. And so that’s the risk over time.

Cost equation

So, how should you look at this from an equation perspective? Well, you’re usually losing partial days unless you’re completely out of that store and they’re unable to get additional stock in from the warehouse. So, it’s the days times the amount of product that would sell through. So, what is your velocity there? How many units are actually moving off the shelf? And that’s multiplied by your price. And so that’s the basic equation there but what you need to be thinking out beyond that is conversion of your customers to other brands. If you’re not consistent ensuring that you have product for them to buy, there’s a higher likelihood that they’ll move on to alternative products. And so, it’s really important you manage down to the store level to have adequate inventory to meet their expectations.

How to detect?

So how do you find out how big your out-of-stock issue is? This is very expensive to figure out. And it’s very hard to do. So, unless you’re one of the major CPG’s you don’t have a field workforce that’s in the stores daily. And it’s very expensive to send employees out to the stores to, you know, walk the floor and figure out what kind of an out-of-stock issue you have. And so besides that maybe you get some data from the stores or maybe you’re getting scanned data and your data might tell you some of the story, but it really won’t give you the details you need. And so, you can start to compare, for instance, reorders from one store that’s similar to another store but now you’re thinking about how much traffic a store gets, et cetera et cetera, and it’s really hard to actually determine the size of the problem unless you have store specific data. And another fun fact we have here is that in the work that we do with clients we find that we have stock-out issues and we uncover those problems in about 40% of the locations that we visit. So, this is a massive problem.

Why Scan Data Misses Stockouts

So why not scanned data? Well, scanned data is usually accessed or received in four-week chunks. And so, by the time you get it its stale and the next thing is it’s aggregated. And so, you’re getting national data, regional data, you know, you’re probably not getting specific individual data down to a store level that will allow you to see what you need.


So, if you’re going to attack this problem of out-of-stocks, how often should you measure the problem? And this is kind of a tough question because we can tell you what will be best for solving the problem, but you need to make a business decision based on what the return is from this. So, you need to look at the velocity of your products. So, if you have a very fast-moving product, then it’s important to measure more often because the cost based on the equations we went over just a couple of slides back is much higher for being out of stock. And so, optimally, you would check multiple times a month. We have customers that check weekly and let me give an example on that, for one of our customers, over an 8 week period we measured the same stores and when we started they had a stockout issue in 59% of the locations and after working with staff and the buyer and other people in their distribution chain we were able to help them get it down 17%. S that’s from 59% problem to 17% problem in an eight-week period. So weekly is expensive though, and so what we find for most customers is that measuring monthly allows them to correct majority of the issues and really understand deeply what the size their problem is and correct it took to a really good degree.

How to Remedy

So how do we actually fix the problem? Well, you can’t manage what you don’t measure, so actually measuring is the first step, right? So, what are you doing that for? Well you can figure out that maybe you have inadequate orders at specific stores if they don’t have back stock which we often check for customers whether there’s backstock to fill out of stock issues on the shelf, you can work with the buyers to increase or size for that store and therefore the amount of back stock as well. If it’s a problem where they’re not filling the order back to the store from the warehouse, maybe the warehouse is running out of stock and you need to actually order more product in the warehouse or increase the size of those orders. And then after that you need to make sure that the product is actually getting to the shelves. So, in that case you need to work with the stores or your distributor, whoever is actually responsible for that, on the frequency of shell restocks.

So, the first item is something that you can attack and correct pretty quickly. The second one takes a little more time and effort in changing behavior. The third item is actually the longest one but can have the greatest impact for you over time. And that is working with the buyer and making sure that they are aware of the problem. And what we found is a lot of buyers are happy to receive this unique data because it’s something that complements the information they already have. They already have quite a bit of data, but this is unique data, it shows them both from a specific store’s perspective where the problems are occurring as well as by showing percentages across stores where they have regional issues, for instance. And it really put you in a better position to carry on that conversation over time and negotiate for more space for your items where you see a higher out of stock rate for your higher velocity items, you’re going to negotiate for more space and you can make a much stronger argument with the data.

The other thing that it helps you do is if they’re not going to give you more space for instance if other brands in the category might be doing better than you are or might be higher velocity than your brand at least you can work with them to reallocate space so that your highest velocity items have adequate space so that you can drive the most sales out of that category.


And that brings us to a conclusion on our presentation for the day. I want to thank you for joining our webinar. And I’d like to ask for any questions. Okay, I don’t see any questions coming through. So once again please reach out to Observa. You can reach me at hholman@ or our sales department at sales@ We would love to help you with your specific issues. Observa measures store execution in real time. We have tens of thousands of people across the United States that use our proprietary mobile applications to collect store specific data and help you drive increases in sales. Thanks again for joining our webinar today.

Leveraging Data to Increase Retail Sales - Transcript


Good morning. I want to thank you all for joining us today for this webinar. Our topic today is leveraging data to increase retail sales.


I’d like to start off by introducing Observa. So, we provide unique retail store specific data for customers and our mission is to help brands and people with products and retail to gain greater visibility and improve sales performance through the use of our platform.

Hugh Holman

A little bit about me, I spent most of my career as a technology manager and a strategist. I’ve worked for international retailers, significant consumer product companies. I’ve sold into both large chains, independents, private label, branded goods. Something fun I’m actually taking my family on vacation next week and really looking forward to some time away from the office with my wife and kids. And let’s go on into our presentation today.

What data is available and how you can get it

Okay, so what data is available and how do you get it. And I love this quote that’s up here from the management guru W. Edwards Deming: “Without data you’re just another person with an opinion.” I think that that is so true, and I think that this is a real challenge that’s actually growing today in that people expect you to come with data. They expect you to come with information and make an argument on why your products are better than the competitions.

Pre-Packaged Secondary Data

So what types of data are there out there? So, one that we’re all aware of is what we’re calling prepackaged secondary data. This is data that’s available to anybody out there. It’s not unique in the sense that it’s available to anyone. And syndicated data is the best example of this. So, you can purchase this data usually through a subscription from somebody like Nielsen, IRI or SPINS. Or you can find this type of information that’s published through reports that come out through various organizations that produce category views and dig deeper into a particular retail space and they usually cost a few thousand dollars apiece and they can really help you if you’re trying to go after a new segment or better understand the one that you’re in.

So, what’s good about this? Well you know where to go get it usually. It’s easy to find you have kind of a defined cost, the data quality is good because they’re selling it to a lot of people and it’s immediately available. But once again it’s not unique data. It’s something that everybody has. One of the biggest challenges that I’ve struggle with in my career and purchasing syndicated data especially is that is comes in a chunk. It’s not real-time. It seems to be a little bit of a carryover from previous times in the fact that you can’t move very quickly with it. You’re getting a four-week chunk of data usually which is great to understand how your products are doing relative to your competition and to see some trending over a period of time and then you can compare that four-week chunk to the previous four-week chunk and so on, but it doesn’t really allow you to react. And so, it’s really a challenge in that sense.

The other big problem I’ve run into is making a consumable by others in your organization. It may not align with how you talk about your products, how you talk about the category in and sometimes it takes a lot of work to get the information from syndicated data to make sense for your people to be able to consume it and make business decision, to change your business based on having that data. And that’s really why you’re purchasing it is to have an impact on your business.

FREE Secondary Data!

Sometimes you’re able to get this kind of secondary data free. Maybe you work with a major retailer – Walmart, Target, Whole Foods – they have a system that you can log into, retail link from Walmart is a great example of this and it will show you information about your products and sales data, where things are happening and maybe where things aren’t. Which I think that’s great.

Another place is with your partners out there. Maybe you’re using a broker which is you outsource sales and they may be subscribed to data which they can share you or maybe they’re going in and getting data from these retailer systems and they’re sharing that with you, so they take some of that burden away and that’s great. It’s great if you get it and hopefully it’s timely. Another thing about distributors before I forget is that they also have the information about what they’re selling on to the retailers, and they should be passing that on to you, hopefully they do. Because you can only have information that what you’re selling to them as opposed to the next step in the distribution chain.

So once again, if this is free, it should be immediate. Hopefully you’re having this information shared with you, but it’s limited in scope because it’s only showing your products not your competitor’s. And then if you’re getting some information from a Walmart, maybe a little bit of other information from your distributor, maybe some from a broker, maybe some from another retailer, if you want to actually look at it all together, that’s very hard. Now you got to combine it. You have to normalize the data, make it all similar, and then pull it together to analyze it and then to do that in a repeated fashion is challenging. Once again it tends not to be unique or other people have this information as well.

Primary Data

So, primary data. So how do you get a strategic advantage? How do you get information that unique to you and how do you get information this more actionable so that you can do something to impact your business right away? And so, in this case, we’re talking about primary data is data that you collect so how do you do this? What if you use your own staff? Now this can be challenging because you actually have a staff, you actually have to have people out the field to collect the data or you can use a third-party. Observa’s a third-party. You can choose others out there. We pride ourselves on our speed and our ability to cover large areas. I’m sure other people tell you that as well.

Also, customer service, going directly to the customer, the consumer in this case is key. Just understanding how they experience your product, how they look your product relative to your competition. And understanding where you were unique or where they may not be understanding your positioning in the market. That’s really great information to get in and I highly recommend that you engage with your consumers somehow in a recurring fashion to keep them in that conversation to help you understand as you move your business forward how they interpret changes, whether we’re talking about packaging or pricing. The size of your package to reach a certain price. How they see it in your category relative to your competition. They actually see your unique value as being unique or if you’re being viewed as more as commoditized.

So, what’s great about primary data is that it’s yours. And so, it is as flexible as you want it to be. You can collect the information that you want, and you can tailor the use of that data to your business. And using the terminology that you use and for the people in your business to make it easy to consume. Also, you can collect it as needed, it’s under your control, right? And so, you can have your people go collect it on a recurring basis or a third-party of course. You can have projects where you decide every quarter when we have promotions we’re going to go find out what’s actually happening in stores. So, you can do it as you need to, so you’re not committed to the same type of expense necessarily that you are with syndicated data. So, it’s unique, it’s yours.

So it’s very expensive, this is a con right, it’s very expensive to build a staff especially out in the field and to manage those people using a third-party relieves a lot of that and it can be slow because do you actually hire a staff that’s big enough to cover all the locations that have the geographic breadth to actually collect the data of the way that you need? Most companies don’t unless you’re a Procter & Gamble or Frito Lay or somebody like that, you probably don’t have that field staff. But a third-party can do that for you just to provide you that breadth and coverage and speed.

What you should ask for in the sales process?

So, once you have the data, how does it actually help you sell more of your product? This is the key. I mean the whole goal of having the information is to help you in understanding the market and then taking those arguments to the buyer. So, let’s talk about that.

New Sale

So, if you’re going to a chain and this is an initial sale, it could be a new product or maybe you’re on the market in other stores but you’re not at that retailer, you’re probably going to talk to the buyer at headquarters. Unless you’re working on an effort to grow your space within independents or through chains where maybe the product is authorized but it’s not pulled through except for if it’s ordered by the store. So, most of the time you’re going to the buyer at headquarters. So, what are you trying to do there? You are trying to show that you are part of a category that’s hopefully growing and having information whether it’s an industry report, whether you’re using syndicated data, to make that argument that it’s a growing category is good. Hopefully they already have that information, I would think that anybody that is managing a category at a retail chain would, but it’s good to go in with your own information. You want to show that your brand specifically, if you can, is growing within the category.

So somehow, you’re doing better than your competition, whether it’s unique value associated with your product that the consumers are engaging with or that you’re doing something unique, there are better qualities within your product or maybe you’re just doing a better job in marketing your sales and you’re getting your product out there and you have more exposure on the shelves and you’re doing better than your competition within those stores. You need to show that your sales are strong somehow relative to your competition because ultimately, you’re working to push somebody off the shelf or if the category is growing, maybe there is more space for you. But most of the time, you’re really just taking somebody else’s place on the shelf.

More Shelf Space

So, if you’re not going in with a new product then you’re probably asking to increase the amount of space allocated to your product. And I love this picture of the fish bowl, we all feel compressed on the shelf and so it’s challenging going in asking for more space and often space on the shelf is the best marketing for your brand. A lot of companies don’t spend a ton on advertising and so getting your product on the shelf in front of the consumer is very very important. So, once again you’re probably talking to the buyer at headquarters and you’re making an argument about why your product needs to have more space. So, you’re probably talking about the sell through of your goods in that there’s high velocity of movement on the shelf. You want to demonstrate that you need more space to combat problems like out of stocks, right?

So, if your slots on the shelf are running out of space, you know, halfway through the day and they’re only stocking once a day, that’s a problem. Because you could be selling twice as much goods. Need to be able to make that argument though and you need to have data to show that you’re running into that issue. Also, another problem that we don’t have here listed is voids. Voids are big problem now out there in the field where the retailer may have allocated this space to you, it may be in the planogram, but they didn’t actually slot you in some of the stores and identifying where those are to get actually slotted where you should be is important.

So, getting out there measuring to find that out to take that argument to the buyers is key as well. You need to show that your sales are increasing versus your competition and you need to manage this space in stores in order to do that. So, solving your void issues, identifying your out of stock issues and solving those so that you can make a strong argument about the strength of your brand in the market to that buyer.

More Stores

And so, then the next space is you want to go into more stores, right? So, if you’re working with a chain often you’ll go to the buyer and they’ll say, “Hey we’ll try your product. We’re going to put you in X number of stores” and you know maybe it’s a quarter of the stores that they have or some percentage like that. And so, they’re trying you. They’re putting you in which is great. But then you want to go back and argue for more space.

So once again you’re probably talking to the buyer at headquarters. You have to show them that your sales are stronger than the competition that you deserve to go into more locations because once again you’re probably pushing a competitor off the shelf in those stores. You need to be demonstrating that you have increasing sales and that there’s some sort of consumer affinity or loyalty for your product. And once again you can take in unique data for this. Primary data from your own consumers surveys or panel information. Also going in with unique store data that you collect, or you have a third-party collect, to make these arguments whether it’s for putting you in more stores or give you more space in the shelf, it’s good to go in with your own data. They love seeing unique data.

Data Format

So, your format. The main thing here is the ensure that it makes sense, right? Know your audience. So, in formatting the data, it needs to be clear. And I have the eye chart here. I think it’s kind of funny. You know, a picture is worth a thousand words. Being able to tell a story quickly is important so you want to be clear with your data. It’s got to be summarized in a way that it’s shareable and concise, so you can get your point about what it means quickly with that buyer. You want to be using their lingo, so you need to personalize this story for them and you can look at a store information records, any information you have on how they use the language and talk within their organization, so you can build that into your conversation. They need to be able to believe it. So, you need to tell them how you collected the data and the best way to make it believable is to take them on the journey. If they are part of the process with collecting the information in their stores, that is a beautiful thing, because they become part of that process and they tend to – we find – they tend to enjoy that and getting that unique data and understanding what’s happening.

So, if you’re out there finding out where you have voids, you know, where you’re running into out-of-stock issues and that’s a regular conversation, so you’re not waiting for some quarterly or half yearly meeting with the buyer. You’re sharing information with them on a monthly or weekly basis and you’re taking them along that journey. It’s going to be stronger when you go in and ask them for more space or to be included it more store locations. Also, they are going to be helping you solve those problems all the time. So, it has to be shareable. So, the data has to be something that you own. You can’t go in with proprietary information that you’re not allowed to share, obviously. And it needs to be formatted in a way that efficient to share. So, you need to do that, or the third-party provider can help you with that if that’s the case.


So once again to summarize. You need to have data, you need to use the data. It’s expected at this point. And, it’s only going to go more and more in that direction, and just telling about the unique values of your product isn’t enough. Sharing what the consumer thinks about the unique value is wonderful to take him to the buyer. And sharing information about what’s happening with your product at retail is key. Once again you can also collect primary data on your competitors. You can have a third-party collect data on your competitors, so you can take that in too. Make your case with facts. Tell them what the situation is and make your argument. You present very efficiently to your audience using their language, if possible, and then leave the information in their hands and make it easy for them to make that decision to follow along with the argument that you’re making whether it’s to put your product on the shelf, to provide you more real estate by expanding your shelf present, or whether it’s to move your product into more store locations.


I want to thank you for your time today and for joining us here for this presentation. Once again, we pride ourselves at Observa in our ability to help with primary data, this unique dataset, about your products that you can use to manage your business or real-time to solve problems out there in your distribution chain. I’m going to look and see if we have any questions now. I don’t see any here. I think that’s it for today. And once again thank you for joining us.