Let me spend the next 5 or 10 minutes or so telling you about some radically new datasets in marketing. So, what I've talked to you about for the first 20, 30 minutes or so here are kind of traditional TV data Online data, etc. Let me talk to you about some radically new data sets in marketing. And let me encourage everyone listening to this, to think about how you might be able to use these kinds of data sets for better decision making. So, one of the favorite studies that I talked about and I'll talk about again in the future is imagine you could track the following data. Imagine you could take data. Imagine, let's take a supermarket and a store. Imagine you could collect data on people's Intentions. So, what did they intend to buy prior to going into the store? Now, why would you want to know this? Well, then you could compare what they intended to buy with what they actually bought and then you could see how much unplanned purchasing happens. Imagine you could collect shopper path data and a radio frequency identification data? In other words, you could track where the customers in the store. Now, that's really valuable because again let me go back to the example I gave before. Let's imagine you're a manufacturer of children's cereal and the reason why your sales are low isn't because your product isn't liked but because no one goes to the store where your product, part of the store where your product is located. Well, you can change that problem by buying different shelf space and moving your product throughout the store. Field of vision. Imagine you can actually have eye-tracking data where you can actually measure what products people are looking at. Because imagine for example, your soda, you make soda and your soda's sitting there on the shelf, but nobody looks at it. Well, if they don't look at it, they can't buy it. Last is purchase data. So, the exciting part about analytics today is again, imagine having all of these data sets at the individual customer level and linked between them. So, let me give you an example of a project that I worked on about 5 years along with my colleague Peter Fader and a former doctoral student of ours Sam Wi. This was a data set called Path Tracker where we actually tracked people moving around stores and you can see a little supermarket cart sitting there and you can see these red concentric circles moving out from there. That's meant to represent a silent ping. You can't hear it, but it's a silent ping that's being sent out to the different scanners around the store. That allows you to triangulate within a foot where the person is located through out the store and so this is the kind of data sets that are available today in application. Now, 5 years ago when we work on this project this was done by attaching little devices to the bottom of super market carts. That was the best technology, that was the Golden Age of Marketing in 2009, 2010. Now, 5 years later, most of us, if not all of us have cell phones in our pocket, and you should know, your cell phone company knows your GO spatial location at any point in time, and they can monetize that value of the GO spatial location. Let's imagine you have a data set, where you now not only know what people buy at the check out counter, remember going back a few slides, that was 1980s data, the scanner data. But now, imagine I could know where you are physically in the store. And by the way, this isn't just for physical instore data, this could be applied to website data I not only know what websites you went to, but the time you went to those websites, how long you spent on those websites. Just think about this as being spatial data. So, imagine a world where customers could be tracked inside the store. So,this solid line that you're seeing now in front of you, that represents a customer's path, one customer's path throughout the store. The little black squares you see represents their physical location every five seconds. And the red squares represent the products they purchased at various locations in the store. And notice most stores you pick up your card on the right. You go around counter clockwise, and then you check out, which is typically at the end at the bottom center of the store. So, this would be a typical path of someone throughout the store. How do shoppers move throughout the store might be one question you're interested in answering? I remember as a child, this was well before I was born, but there was a famous show on Leave It To Beaver which talked about Aashow that was taken place in the 1950s, was meant to be an all American family and they always would show the person going up and down the aisles of the supermarket. Up, down, up, down, up, down, up, down. It actually turns out and actually it's only through analytics do we know this today. It turns out, that's actually not how people move throughout the store. People do not go up and down the aisles. As a matter of fact, if you go in one side of an aisle and go out the other side, that's called a traverse. If you go in one side of an aisle, out the other side and back around the endcap, that's called a zigzag. Where you go, let's say, aisle 5, around and then up aisle 6, that's called a zigzag it turns out. Most customers make only one traverse, meaning they only go down an aisle once. It turns out most people do what are called excursions, they go inside an aisle and then out the same side they came in on, most people do not go up and down the aisles and we now know this because of analytics at the customer level. So this is not how people move. What I'm now showing you a picture of is a store planogram where you can see where most people move. And notice in the center of the aisles, you can see the center here, nobody ever goes to the center of the aisle. Most of the time, people only go to about a third of the way, and they come back out the same side. So, why is that valuable. Well, let me go back to my emergency room doctor example. I'm a manufacturer of a product, I'm selling it inside a supermarket and my sales aren't what I would like them to be. And now I start to wonder, I wonder why? I wonder if it's because people don't like my product or I wonder if it's because when they go in the store they don't ever visit my product. Well, customer analytics now allows you to do this. What you can see from this store level layout is that people just don't go to the middle isles that often. They're pretty much only doing whats called the racetrack or the outer ring of the store. They almost never do the interior isles and the basic idea is that's bad shelf space. So, you should not be paying for shelf space where customers do not go. And as a matter of fact you can show that, and we've analyzed data from thousands and thousands of different stores. This may come as a surprise to you. In physical stores, supermarkets, clothing stores, sporting good stores, etc, most customers on any given visit only cover about 25% of the store. That's remarkable so 75% of the store of any visit they do not go to. And that's an opportunity. It will be valuable to know that. What I'm now showing you here is what's called the heat map of the store. Red is hot. Green is not as hot but still pretty much hot, yellow is less hot and then blue is cold. Notice where's really, really cold. The middle of the aisle is cold. The outer ring of the store hot, the race track which is why you see firms paying lots of money for the outer ring of the store. They want to be on the end cap, they want to be on the end aisles. And by the way, the only way you know this, now, could you actually have humans sitting in the store, watching where people are going and recording it, you could. Very expensive can't be scaled measured with a lot of error. Could you put cameras in the store, which has been done for years? You could, but it's very obtrusive. Now all of a sudden, people know they're being watched. Maybe they behave differently, plus you have to take that video data and you have to make it quantifiable. What customer level analytics has done is, I can now track customer by customer where they go inside my physical store, and I can actually tie that to there purchasing data. So now, I can target them even better with products and services.
