 let me start with the story okay there's a man coming home from us and he's tired and it's a rainy lady in there and he walks into his neighborhood supermarket store he goes there and he looks for aisle five and he's looking for that one unique breakfast cereal box that is five years old is agreed to eat for breakfast tomorrow he walks down to the shelf looks for the box and doesn't find it there that is his moment of truth and the challenge facing the retailer is how do I ensure that I get the right product in the right channel in front of the right consumer at that right moment right so my name is Amit Kapoor I've been working as a consultant for a company and I've been working in retail and consumer clients for the last 10 years trying to solve some of these problems what I'm going to try and talk is what are the opportunities for retailers to try and address these consumer performance using data okay let me start with a confession I am not a big data guy I know a lot of you are techies a lot of you are statisticians I'm actually a business guy right and a lot of problems that I have solved would be actually solvable using XR to a large extent the fantastic tool that we have and the approach that you would see that I would adopt to solve is what I would call strategic problems right problems that are into some extent wicked in nature there are problems which have fuzzy context there is undefined data or very little data there is fuzzy data there is undefined data there's a lot of correlations there's a lot of causality that's built in how do you actually take those out and build it build it into a solution and the approach that what I would have taken in a number of these cases and what you would find on my limb piece of real estate on the web is how do you structure that problem and I'm not talking about structuring the data how do you structure that problem to understand what are the hypotheses that you're trying trying to solve right how do you synthesize the data to try and address some of those uh some of those hypotheses and then how do you solve it and I when I solve it in a very abductive way in a very hypothesis driven way in which both the problem and the solution and I will merge this and change this and then how do you actually tell it into a story so that then you can go and communicate it to your stakeholders to your business and convince that this is the case for sure and the argument that you need to do in all these cases is how do I solve that problem to make that change happen right that that is the key thing you need to keep in mind whether we use big data small data that's the ultimate goal that we're trying to accomplish let's start with a little bit of context um I should have talked a little bit about it for me retail is a very personal right we talk about the moment of truth it's the only industry where every retailer has the first moment of truth when you actually go and use the that's the product when you can actually pick the product most of the industries it's the second moment when you actually start using the product with what you're kind of getting through any retail is the only one where that first moment of truth comes through and where whether you are walking into a dirty IOC of Petropa or into a public sector firm or into a supermarket they're all retail experiences that you have right and they're all moment of truth for that customer retail is also inherently local right if you want to go and shop for grocery down here in Bangladesh you can go to a spa you can go to a big bazaar you can go to a star bazaar you can go down where you finally get in football but that's pretty much your universe right and that's also the universe for the retail because that's what is what is the likelihood of catchment of people that are going to come here I think you saw yesterday somebody talked about pink code based analysis for doing big catchment analysis that's reality for retail because that's their dataset that they're going to look at the third thing is retail is evolving she talked about 470 billion dollar of detail out of which roughly five percent is organized and even though this we've been talking a lot about internet retailing internet retailing in India is five percent of that five percent right so bulk of the retail is actually happening when you go to a kirana store which is the unorganized as you would call it even though I think they're quite honest or the big supermarkets which you go and buy the food or not the products the second thing that is tremendous potential for retail right because retail in India it's five percent if you go to another market like UK and you look at grocery you'll find 65 percent is with four big leaders as those in supreme asda and versus and if you take the top 10 you have 85 percent so the beast that for the potential for growth is amazing and not only that it's a new marketing market you can see retailers right now experimenting with formats with customer service with a lot of different things which they have trying to figure out how they need to do our work and it's the same to outside where you look at people experimenting with big baskets or big box retailing now move to what you would call showrooming as the apple would call it or maliki channel experiences and all that so there's a lot of things that are evolving here how does a retailer look at is if you look at a profitability lesson say how does a retailer look at a customer or the revenue the top left right you'd go and look at it's a simple guess you a lot of there is a food for that people are walking to the store and you want to convert them you know because you want to convert them into a customer right so there's a conversion and then you want them to literally hold a basket want them to hold a basket put stuff into it and hopefully they put enough number of items at a high average value so that they walk out with a big basket that benefits right that's the old mechanism that you're trying to optimize whether it's an online to something or definitely another topic and on the past side if you look at it the bulk of the cost is budget that's because that's what a retail is trying to reduce that is the 60 to 70 percent of the cost in a retail because he's a trader he's buying the product and selling it to a large extent the next big item store space and store productivity is one of the key methods that a retailer would look at cross margin per square feet or if you were to call the equivalent of velocity like we talk about velocity in data the equivalent of velocity in a retail is how how fast is my product flying off the shelf and am I making money out of that those are the two big factors just two metrics to just give a sense of how hard a business is if you look at gross margins for a retailer the average for the top 250 retailers would be only 3.8 percent right so that's how thin the margins that a retailer works in none of them are in India none of them in India because of the top 250 your gross is actually even lower 2.8 if you're actually hard lines which is kind of DIY punishes and all you'll be 5.5 your fashion yes you make more money 7.8 yeah so this is net margin so gross margins no gross margins would be 30 25 to 30 but net margins after you've taken out the fast for a store space employees and central to be actually the other if you just look at few ratios which is just the market cap by asset that the retailer has the average for the industry is 0.8 right so you could actually take the buyer retailer right sell up all its asset and still make money right if you look at for a fashion retailer yes it could be a little high 1.6 if you look at people who are very well the amazon is the world the apple to the world the indian tax the estimate of the world you get to more like 4 and 5 so they are really value value on top of their assets but for the rest of them it's pretty much point 8 so not only is it a low margin business it's a very asset intensive business so when we talk about big data or any data the ability for our data or any analytics to do under the radar to pump up our margins is very high it's very high as well the ability of this data to price comparison websites and retailers to actually pump it down is also very because it works both ways so what are the three what are these three two or three things that your retailer can do to actually use data in a more effective way right and then there are lots of jargon we can put it at the end there are three basic packets that you can do that right you can either make the supply side more effective right so these inventory that you hold in the cost that you have of your products how are you making more effort you can either optimize the demand right or you can make the customer part more effective where you come in with this loyalty part that you can talk about the merchandising the loyalty part the resell part and all that he's covered by very much right so we'll talk about the first two right how should we do that right so let's and we kind of think of this big data as something dramatically new where it's come up it's actually not really new data has been there for a long time right if you look at let's start with the story right it's the second story you want to tell me three stories this is the second story right 1948 there is a local food store in Philadelphia which comes across with brexit these students is i want a way to actually classify my products so that i can check them all fast so two young graduate students well pretty much like that you know young developers there worked on that problem and patented with the us pto patented number 992 you can actually search it on google to get this image this is the patent that the file is the art of classifying items to identify patents sounds very detailed right art of classifying items through the use of patents right and this is technically what we would call the first parkour well it's not really parkour so it's a concentric circle those are which is how you would say and it took the industry another 20 years before they standardized on this which is the new pc the universal product for which is used now everywhere to scan products and make them check out right it's using details i can use in many other industries right and if you look at the concerns that have exploded in 70s the first thing that came out was a 10-pack between 467 cents was the first item ever scan and it never picked up it never picked up for five years people wrote obituaries of the scanner data people wrote business we've had a cover saying the scanner supermarket experiment is dead right and never picked up and there were concerns on data integrity and privacy customers were not willing to buy this a concept because they thought if we can't see the price but we see a scan for what we're throwing in the retailer to change the price so there were senators trying to repeal this and say we don't want parkour here right some pretty similar to what a lot of the concerns we now hear about here right and it took the industry another five years before this very picked up and it really picked up once the one much of the world and it came out to the world started to use this to make the products more efficient and then came to the point of sense right so that is the genesis of point of sale the point of sale data that picked up at that time is really that drove the scanner data which was first to the retailers and then with the consumer companies coming through the retail links that Walmart's the world established which then kind of flooded into the industry and then was started to use to become more efficient right so that's the genesis of where we are on this efficient and what is the next step where the big data can take us what does this efficient consumer response that has been talked about they've been RFID been talked about for a long time which has been already taken but the future of this far of big data is really coming into how can we take this similar data which is skew buys inventory level and convert it into a demand sensing platform where we can actually sense the demand in a very short time we can then in hope that we can actually service that you know and not only do it at a company level but the problem is to solve it across what I would call a multi enterprise level so not only the retailer solves it but the retailer works with its consumer company and solves it and if they do it you can actually optimize the inventory across the entire chain and make it much more effective and then that's why you have this concept of demand and supply chains which is really an art when the retailer picks up a part of sales stock from this tell can that trigger the demand so that people can actually supply the supply and then supply the product and I have my moment of truth where I'm not facing a stock there are companies that are doing this well the uh fairer technologies that actually does a very good demand sensing tool which is adopted by a bulk of different consumer companies and they really use partner identification and a lot of algorithms to kind of manage the demand sensing partner integrity but the challenge here is actually integrating not at a company level but actually doing it across your chain because until you do that you don't see the next level of benefits that you would anticipate you would want what's the second problem that that the retail uh that the retailers face right and uh this is another fun story which talks about uh there are these two little fishes that are walking and swimming in the swimming walking the system walk they're swimming in the sea and an order of fish passes by and I'll ask them well how's the water the two fishes swim along and then one of them turns to the other and says what the hell is going on right so there is this other data which is all around us which is very observable which is can be actually gathered by observation and this is really how consumers how consumers really interact with the store environment so I think the google keynote actually talked a lot about the red riding hood story which talked about how all that data is a process and there are people who have done this there's uh people who looked at how people interact in these old small spaces the size or the art of shopping and how people actually move in this back to our business in 1977 and he actually looked at really people observing people right now how they're interacting in the retail environment his background was for and for college and he looking at that data which was tons of data which started to really even walk through noting now every aspect of how consumers walks in such as the retail environment interacts with the environment looks at the data doesn't pick it up does pick it up tons of this data which we would now call potentially like the access to the data and actually started to analyze and the in-store environment which was really what picked up built this science of shop and you had all these null insights like coming through which now are being validated again by the data uh how does customer react when you walk into this store and there is this any reason for like really doesn't actually interact with the environment in the store there is the fact that we need shopping class to actually make a specter right this wasn't discovered for a long time the fact that we actually walked on the right when we look at a grocery store we don't really look at we don't really look at we don't look at the stuff that's on the left side we're not actually walking through the right and how intersection actually increases your ability to buy a car when somebody walks into a retail store and a person comes in and says can I help you it's actually because they figured out that actually increases the propensity of purchase and there were many such insights that they picked up in terms of what men don't like what pricing what women like what pricing of senior shoppers and all this was done in a way which weren't really required it here at that time it was done through manual observations later through video cameras being transcribed and the future of this now is the use of big data is how do we get these in store insights how do we get these shopper insights using analyzing the video data analyzing the mobile data that is there in the as you walk across and their companies are doing this retail night's customers that looks at all the video data comes in and triangulates with the mobile data that you have and places it in their companies that give you a reward to actually walk around the stores place customers to actually see that whether you can we can gather the food for data that will actually help us generate the heat bath that are required to actually see whether it's part of the stores or the customers walking in which part of it right and this links into normally the installed customer aspect of it but also how employees yeah so what should be the design of the store how should merchandising be kept and how what are the then the options of operation improvement that can happen which can take over on this process so those are two examples around how what we've been doing in detail for a long time can be taken to the next level by layering in this layer of detail right there are others I think which we which I won't dive into but the next best offer which is I think she's a talk about how do you get the next best offer to the customer the moment he walks in or when he makes a purchase I know about that how about multi-channel big question for a lot of retailers you won't see any Indian retailer right now offering two channels most of them either online or are they offline how do you actually integrate these two make that happen and then not only this pretty much so how can you change pricing promotion and optimization to actually and a challenge for retailers at report to actually do that one is privacy so there are huge concern and privacy the moment I start observing your offline data activities you are most customers are okay or they don't realize that they've been tracked online but the moment I have a video camera and I start tracking you and I kind of start using that as a case big issue target did this recently when they talked about how customers behavior change because at key moments in their life if you can predict those key moments of life you can actually get better one of the moment was when first when a woman gets pregnant so when we are about being out of the family then that's a moment and that raised huge privacy concerns how if you predict that moment we may not be comfortable being low being comfortable being told that somebody's actually so there's huge privacy concerns in this aspect the second is legacy it and and the mindset around it I welcome number of retailers where we've asked for we have these store data can we optimize the store range to more than each store so personalizing at a store level not even at a consumer level and a lot of the challenges that process that we've built our IT system it's already hard-coded into nine categories of stores we can't really move online so this whole investors require to layer in this legacy or either we remove the legacy or layer something off is a hard question for retailers especially when we link it back to the partner they don't have high margins to support all this kind of investment and the third is obviously talent um everybody's excited about working on the cutting edge uh complex for the week for uh for a lot of internet startups you'll find very few people interested in going to a retailer and say I'm gonna help you to all the next time I mean I think in the Carolina statement that they're not really caught on to the data I think there are a lot of them have have good loyalty programs if you look at where the big money that they wanted uh where there's a big money you may or big potential improvement it's actually more on the customer side stuff that she had talked about and I think they are they're starting to do something they have loyalty programs they're starting to find it and start a division but if you think about challenges about in-store design I never had a good experience shopping into one of the uh reliance list of the big bizarre and uh or any of this intelligent environment and if you look at demand supply chain I think just the supply chain is not that elaborate enough for them to start talking about this okay thank you