Lost in the Supermarket

Why Food Retailers Need To Embrace AI

November 01, 2020 SupermarketGuru
Lost in the Supermarket
Why Food Retailers Need To Embrace AI
Chapters
Lost in the Supermarket
Why Food Retailers Need To Embrace AI
Nov 01, 2020
SupermarketGuru

On Today's Lost in the Supermarket episode, Phil talks with Jason Hosking, CEO & Co-Founder of Hivery, The worlds first category management optimization solution.

Show Notes Transcript

On Today's Lost in the Supermarket episode, Phil talks with Jason Hosking, CEO & Co-Founder of Hivery, The worlds first category management optimization solution.

Phil:

Welcome to Lost in the Supermarket. I'm Phil, Lempert your host. Now I've got to tell you something over the past, probably two or three years. U m, I have bought, u m, a device or I , I , I really w ant t o say devices that I can't live without it is Alexa and probably Alexa is going to be triggered b ecause it's sitting right over there and it's going to say something, but when I look at what the power is in these devices, whether it's for music, whether it's for information, my calendar timers, it's extraordinary. I wake up every morning, jealous that I didn't think of it. Well, it's all about artificial intelligence. And with me today is Jason Hosking, the CEO and co-founder of Hivery an AI company. And I want to know more about artificial intelligence. So that's where we're going to start. Jason. So tell me exactly what I don't know about artificial intelligence, and then we're going to move to the food business. And then we're going to hear about what the food business could do by getting more involved. And then I want to hear specifically about H ighbury, what you're doing. I know you've got some extraordinary clients, including folks like Walmart, u m, major CPG companies and how you're helping them. But first give us the one Oh one about artificial intelligence,

Jason:

Phil. It's great to be here and thanks for having, thanks for having me and, and sharing, you know, hybrid story and our knowledge about the sector. So jumping into artificial intelligence, it's, it's, it's a huge topic, right? Um, sort of broad ranging. And I think one of the things I would say, u m, you know, just to kinda kind of bring it to life is I think that the term artificial intelligence or AI as it's affectionately known, u m, it draws a lot of confusion when people start to talk about it. And I think it's because it's such, such a broad topic. U m, you know, if you think about it, I know t hese are the three things that I w ant t o kind of call out when we think about artificial intelligence. U m, one, I think historically we've always associated the term AI with movies. Um , a nd that was certainly my case. When I entered into the sector, you know, in the last five , six, seven years, you throw the term AI around and you immediately start thinking about, you know, star Wars, Terminator two, u m, 2001, a space Odyssey, you know, even looking back, you know, t he, the Jetsons. A nd, and so I think because of these sort of fictional kind of connotations with the term AI and these robotic characters, u m, it all sounds a little bit fictional to a lot of us, a little bit out there, right? And as we move forward into, you know, now in t he year getting towards the end of the year 20, 20 IOI, very, very much, u m, part of our lives. And so as we think about that, firstly, that association with movies and the future a nd robotics, I think the other thing we need to all acknowledge is that artificial intelligence intelligence is a very broad topic. You know, it ranges from things like the calculator in your phone , uh, to self-driving cars, to something in the future that might actually kind of surpass human intelligence and have a major impact on the way the world kind of fundamentally operates. Um, and so artificial intelligence refers to sort of all of those things. And I think that that reason , uh , such a broad topic and can kind of conjure a little bit of confusion , uh , out there. Um, I think thirdly , um, we use AI all the time , um, in our daily lives and we don't even realize it's artificial intelligence. You know, you think about you, you mentioned the Alexa. Um, I , I was actually driving into the office this morning. Um, and I, and I just noticed, and I hadn't noticed this for a while that there's no ticket machine. When I go into the car park anymore, there's a camera that recognizes my number plate. It knows how long I've been there. And it charges me on the way out. That's, that's a form of artificial intelligence and image recognition technology. Um, you think about, you know, one of the things that I often, you know, most simple cases is, you know, the GPS systems that we're all used to using. When we get in our car, they used to be the T om T om devices that are on our phones, u m, that t hey're in our cars. And w hen you think about it, you give that a command. You say, Hey, I'm here. And I want to get there, give me the best, the best way to get there. And you've got some options along the way. You might want to stop somewhere along the way to visit a grandparent. You might want to avoid toll roads. You might want to take the scenic route and you, you've just given it a n, an objective about what you want t o w ait, what you want to do, and it gives you the best outcome. That's another form of AI. You know, the thing is that's really interesting about it. That artificial intelligence has been around us for a long, long time. I think the first time , um, you know, John McCarthy who actually coined the term artificial intelligence back in 1950, 1956 . So think about that. That's getting on sort of 70 years ago. Um, he sort of, he sort of, he's quite famous for saying , uh , one specific quote, u h, that as soon as AI works, u m, no one's g oing t o call it AI anymore. C ause that's how pervasive it's going to be. U m, and so, y ou k now, it all sounds like this mythical future prediction, but today it's much more of a reality. And so I think there's just to kind of round out this, this point about AI is really three types of artificial intelligence that we should think about. And one is artificial narrow intelligence. It's where a machine and AI, just, just to kind of put that into context is, is a machine that's able to think in a way that similar to complete tasks or think in a way that similar to what a human does. And so when we think about narrow intelligence, that's kind of where we are today. Um, for quite a few decades , um, uh , a computer has been able to beat , uh, the world's best chess champion. And so you give it the objective, Hey, this is the rules of the game. And our objective is to beat the opponent and it gets very, very, very good at that specific task. Um, there's a great video when you look at some of the old computer games and you just give it an objective and you watch the AI learn that rule, and then you start to think about artificial general intelligence, which is where, you know , um, they stuck where the information that the machines are able to process start to become more expensive, start to behave more like a human brain. And then this is super intelligence concept where kind of some point predicted in the future. Um , and there's a lot of argument amongst the industry about if that will occur or even when, and what that will look like , um , is when that there is a super-intelligence . But I think today all of the examples that we talk about the field that hybrid operates in is very much in that narrow intelligence is able to compute large amounts of data, run far more scenarios that we as humans are able to do and do things in very unique ways.

Phil:

So what I'm hearing loud and clear is that we probably need to change the term AI or artificial intelligence into subsets. So if , if I say , um, you know, I wanna buy food, that's a huge category. Um, here in the States that comprises 40, 45,000 different products that are located in a supermarket. Um, if I say, I want to buy meat , uh, that narrows it down , uh , to probably about, you know, a hundred different kinds of meats, whether it's chicken, whether it's , uh , Turkey, whether it's beef and so on. So what I'm hearing is a lot of the confusion is exactly what you said is the terminology that that's being used by a lot of people. And it's getting lumped together in a way that that is confusing to consumers. So, you know, how do we properly communicate to the average consumer , um , who might have a device on their desk or in their homes that is that narrow casting device , uh , that turns the lights off and on, or, you know, plays music or whatever. You know, I guess my, my question is, is not only how, but should we be narrowing it down for them to further their understanding and increase their use or the better off, you know, just not knowing and just making it part of their lives.

Jason:

Oh , look, that's a really good question, Phil , that I'm not sure I have the perfect answer. I, I think not knowing necessarily the answer , uh , I think it back to a little bit that quote from John McCarthy, one of the founders is once it becomes so e vasive, you don't really know, you know, that AI is there, like all the examples that I talked about. I mean, a lot of this and a lot of what we do i s j ust very, very advanced mathematical computation. That's able to run a whole lot of, u m, different computations and come up with the best answer in v ery, a very quick manner, which is something that we haven't been able to do historically. And so, you know, do we necessarily need to acknowledge, you know, back to that example of driving into the c ar p ark or, you know, you know, you've now been able to control the temperature in your home by a device that can monitor the external internal conditions that are just things accordingly, all of those sorts of things that have a form of artificial intelligence in it. Um, you know, I think that the terminology, you know, you're right, I think we could start being a bit more focused. I think the other thing is that I think naturally one o f y ou on this, I think naturally as more and more industries and more a nd more consumers just get used to this. I mean, we talk internally here at h ybrid a nd i t's a very common theme that artificial intelligence is essentially like data. It becomes like a general purpose technology. It becomes like electricity o r like the internet. It's something that we tap in and we use, and we utilize through so many different services and, y ou k now, so many things, you know, s o some of the examples we've talked about just now a very obvious ones, y ou k now, when we start talking about the retail sector and some of the work that h ybrid does, u m, you know, it gets very, very niche and very specific to solving certain problems. But I think we're all, we're all interacting with this more and more in our daily lives , whether it be the devices we interact with, how we manage our homes and our lives, or the way things are done at work. And I think , um, just kind of rolling it up another level as well. I think one thing I'd like to talk about is that, you know, one of the things that comes across with th is v ery general and broad topic of AI, u m , i s a lot of the fear around, yo u k n ow, job replacement, u m , a nd how, u m , y ou know, machines are go ing t o r eplace us, u h , d oing all sorts of things now. An d o ne of the things that, one of the areas that we'r e ver y focused on a hy b rid is what we call where, um , art ificial intelligence and computer technology is augmenting and sort of elevating human decision-making, where it's actually the combination of the human skil l set and the way the mind works along with the technology and the, and the ability to process huge amounts of data to come up with a far better outcome. And so one of the things, you know, I remember speaking publicly, not that long ago w hen I was referenced, I was d oing a talk a little bit on a similar topic to this. And I talked about the movie Terminator t wo. And if you think about that movie, which was a very fun movie of mine growing up, u m , y ou had, you had Arnold Schwarzenegger, the Terminator was all about logic. He moves the machine focused on a very specific outcome, but then you had John who was, you know, brought empathy, intuition, understanding to the situation. I was actually how those two people work together. Wel l, t here's two things work together to achieve an outcome, but col lectively they ach ieved a far better outcome. And we see that today in all sorts of industries. And I want to bring that to life in actually the medical industry. There was a study done a couple of years ago that looked at, um , um, uh, mam mogram tests and being able to detect breast cancer in w om en about a year, about a year out from when they were formally diagnosed. And what they did is they ran , they got , uh, several thousand scans and they gave it to a, like a leading practitioner and they tested the error rate just by looking at the standard ways of doing things that day. I think the practitioner and I might be slightly out kind of error , right? Of like 0.05. So quite a smaller, right. But an error right down the list. Then they gave exactly the same scans to , um , an AI, a mathematical model, a computer, and said, okay, you , you look at these tests and see what the error rate is. Now, the error rate improves slightly. It dropped from 0.85 2.79 . But then what they did is they then said, Hey, how about we get the pathologist to work with the AI system to see what the outcome was? And the error rate drops 95% down to 0.05%. And the point was that the computer working with the human intelligence intuition and the human mind working in combination with the, the technology could actually get a faster period result. And that's exactly what we see in the world work that we do. It's sort of, it elevates the intuition that the external knowledge, the, I guess, the broader intelligence that we talked about, thinking about those definitions before to work with that narrow intelligence to come up with an optimal outcome and kind of focus more on a strategic Intuit , intuitive, innovative solutions that perhaps the data doesn't see . So I know we went down a different topic there, but I think that to your original question about how we should , the terminology is definitely confusing. Um, I think maybe as it becomes more pervasive, we'll start being able to pick the specific segments that underpin the broader topic and understand how they're utilized in our lives. And, and maybe be a bit more specific in the way we talk about AI.

Phil:

So let's, let's take on humans for, for a moment and not from a job standpoint, but one of the things that I hear a lot, u h, both in business and and friends, u m, is, you know, they're concerned about privacy as we see more and more of these devices a re being put into practice. They're fearful of their own privacy that you know, that they've got, you know, this device in their home that they c ould be listened to. What do you say to those folks?

Jason:

Look, I think it's a really genuine concern and they have every right to, I mean, I think the power Of the technology , um , is significant and even us technologists need to not just focus on technology for technology sake, but understand, you know , the rights of the individual , um, and the right to privacy. I think, you know, the reality is, and it is quite an interesting, you know, when you look at some of the studies in this space , um, there's quite a difference in the attitudes towards privacy and giving up some of your data in generations. I think that the generations that are coming through now with just the connected generation that used to, they grew up with Facebook and they grew up with Instagram and they grew up with the internet. And , you know, I think everywhere you go, now, there are so many services that become available to us that make things simple. And so without realizing we're actually giving up personal data , um, you know, I think about the airport example of not that many of us are spending a huge amount of time at airports right now, but you know , you're at the airport , I travel internationally, I'm on an international flight connected with a local Wi-Fi , you know, you give them your email address and then you give them a few things. And all of a sudden that information is stored in a data and they know where you are. And then probably the next airport, you do the same. So all of this information starts to connect. They know where you're going there . They know where you are searching the , you know, they, they, they can connect it to other databases. We all know now that, you know, you think about , um, we all have these pretty , um, on Kennedy experiences, how targeted advertising and information can be. I was looking at a new couch online. Um, you know, I was searching couches and sofas for my home. Literally the next pulled up, my phone pulled up, my Instagram, I'm getting served all of these gorgeous sofas in advertising . Um, that's the reality because we are all giving up our personal data quite regularly and we don't know it . Um, I think, you know, there needs to be , um, some broader education around how a lot of the solutions we use as consumers. Um, you know, and certainly in library women more focus on the kind of B2B kind of solutions, but certainly , um, it's certainly a good point bill. And I would say we should also be very conscious of it, but we should probably try to educate ourselves about how much data we're actually giving up and how much is available out there for businesses to see whether it's our credit history , uh, whether it's, you know, our search history , um, you know, whether it's our location, whether it's our loyalty data, you know, people need to realize if they sign up to a loyalty program at a retailer that you're actually giving, you know , wall that might give you benefits and discounts and promotions and special deals, you're also trading something well that, and that's your information, and I'm not sure how broadly the community knows. And I think that they really need to start to understand that a little bit more and understand the implications of it. And I think there probably needs to be some board of governance and broader rules, but one thing I will say I 'm playing in the data space. U m, there is a lot of, u m, you know, there's a significant amount of rules and r egulation and compliance around when businesses d o s top to have access to certain pieces of information. One thing personally that struck me in all of this is I think I' ve r eceived personally, nothing relation to anything we do at work, probably five data br each e mails in t he past six months, you know, from certain services that I've been, yo u k n ow, on e of , o ne of the hotel services that I've got sen t me an email saying, Hey, we've had a data breach, or I t hink one of the, you know, the, the booking system is where you book restaurants that I got an email about my data being breached. And , and my fear was when I saw that, well, Hey, this is just the normal. So we started off, you know, giving away our data without thinking about it. And now we just add that. And now, you know, the systems and processes that the business had to protect our data probably more easily penetrable when we realize, and then we get these infants , you know, they need to report when it is a breach and all of a sudden that stuff that happened. And I think, I think our defense, our sort of acceptance of all of this is it's just, God's not really often. And I know it was just a personal reflection. I didn't, I didn't think too much of it. I didn't do anything about it. I didn't complain. I just went, Oh, okay, well that was done. I guess I've got nothing on , so that's not a big deal, but that's , that's a pretty, that's a pretty complacent you to take to your personal daughter. I think.

Phil:

And , you know , I had something similar happened probably about six, seven months ago. Well, pre pandemic. Um, I get a call one day from M erck and airlines and, u h, I'm a frequent flyer there. And they said Marriott has had a data breach and we partner with Marriott. So as a result of that, all of your American airlines information has now been breached. So we've got to change your frequent flyer number, u h, to, to avoid problems, which is the first time I really thought about, u m, how all these companies are exchanging information working in partnership. And, u m, a nd, and to your point, what I found interesting though, is something that your folks sent over to me t hat s aid that IBM said that 90% of all the data that's out there has been collected just over the past few years and F orester t hen said, but only 12% of it is being used. Talk to me about that. W w we're collecting this. I remember when frequent shopper card programs first started hitting supermarkets here, and everybody thought it was the greatest thing ever. You're going to have all this information. And I'm talking probably 30, 35 years ago. And I was talking to a CEO of a supermarket and I said, yeah , tell me about this program and whatever else. I mean, he said, you know, follow me. We were in their corporate headquarters. We go into a room and in that room are boxes and boxes of computer pronounce. They had no idea what to do with all this data that they're collecting. Now we've changed that a bit. Uh , but, but the point of it is that we still have all this data that's not being used. So let's talk about , uh , and , and let's move it into the supermarket . What should supermarkets , uh, what should CPG companies be using that , the wrong term AI , uh, to, to discover about you and I, and , and their businesses.

Jason:

One thing I'd like to s et u p, I'm really glad you brought this point up because those two stats, ye ah. 90% of the world's data was created in the last two years. Yea h. T he other side of that equation, that organizations on average using about 12% of that data, they were two of the fundamental stats that actually made myself. And my co-f ounders Fra nkie, Matt and Mink ah is f ou nd out company. We just sat there and we looked, you know, Frankie and I were kind of the entrepreneurs of t he business side of the equation. And the other guys were a PhD level research scie ntist. An d you jus t, the first time I heard that, and I t hink, I think I use that stat almost in e ve ry time I speak, because for me, it really, it really showcases what's going on right now. Um, we have , we are sit ting at the, really at the beginning of a massive data slash digital revolution that will change the way we do business and the way we live in s om e of the things we've talked about already, but just , ju st to stop for a second and think about that. Um , 90% of the bolt's daughter was created in the last two years. That means t hat i t's just growing exponentially. And we look here a daughter coming out, besides just, it's just going to grow exponentially. At the same time. Technology is advancing super quickly, our ability, our ability to process such data and mining i nsights from it i s becoming more and more powerful by the day. And so, you know, and then organizations, so they're creating more, but they're not quite sure how to u tilize, how to utilize it, how to, how to think about, u m, running their businesses more effectively. And so sitting at the center of all that was actually how h ybrid started. We weren't necessarily retail specialists. Although we had a special partnership with the Coca-Cola company that allowed us to kind of have access to data sets and industry problems and subject matter experts. But we really b ought with a w rought with us, this innovative mindset that, Hey, this is going to change the way we do things and how can we help this industry become more effective and make better decisions. And I think one of the things I'll talk about just quickly, u m, is in the world of sort of data science and AI, there's sort of three kind of areas, u m, of data that I want to talk about. One is descriptive analytics. So if you think about that pretty much, we've always looked at descriptive analytics that tells you what has happened. So, yo u k n ow, your monthly sales report, u m , y our, you know, yo ur, your profit loss statement, it is a, it is a description of what is w hat has happened. And a lot of the time we use point of sales, the ho tter o r whatever, whatever data sets we use is , is telling us, u m , w hat has occurred then the next thing that has been a big focus over the last decade or so is, u m , p redictive analytics. How do you use that descriptive data to start to predict what happens next? And that's where it gets really interesting because you can stop doing, you know , all sorts of interesting forward looking full costs. You can really start to predict certain outcomes with certain degrees of confidence. Um, and then that's where I asked stuff office to get really interesting. And , you know, especially it's growing data sets and computational power, the accuracy, the detail of those predictions becomes really, really novel and really, really interesting. We actually focus on that next step, which is called prescriptive analytics, which is the way we actually use those predictions to prescribe the best action . So you're able to give the computer, you're able to give the software and objective and say, Hey, I want to do this. You know , I want to generate more sales. I want to reduce the amount of stock outs. I want to increase the days of supply or want to maximize the range. I want to, you know, whatever that may be. And then people give you the best outcome. It can tell you exactly what to do. Um, that's where it starts to get to get really, really interesting. If you start to think about , um , you know, I'm gonna, I'm gonna, I'm gonna reference this report here. Accenture , I think was in 2018 , um, did a report about , um, you know, the revolution in the workforce that was going to occur as a result of AI. And , um, and interestingly , um, the number one industry that t hey saw that was g oing t o benefit the most was t he CPG consumer goods industry. A nd I looked at it and said, u m, u m, I've just got a note here that, u m, i nvestment in AI and human machine collaboration was going to boost revenues in the coming period by 51% for the sector. And the next closest t o that was h ealth 49% telecommunication, 46%. So I think the message that I would like to say to the audience and to yourself, that there is a huge opportunity , um , in retail, in seat consumer goods , um , around investing in this technology. And we're seeing this more and more, more , uh , another , um, another study sort of said that , um, you know, by 2022 , um, I think this was Juniper research and were looking at another stat , um, that , um, retailers would spend in 2022 to $7.3 billion on AI. I think that's a huge number. And you compare that back to about $2 million in about 2019 in 2019. And so sitting in the middle of this is a huge, huge opportunity. When I think about those prescriptive analytics, things that I talked about, what it really comes down to is how do you elevate your decision-making ? How do you get more targeted, more specific in how you run your supply chain, how you distribute your products, how, you know, how you stock the shelves, how you, how you actually make sure that as , uh , as we work with our partners, our alcohol CPG manufacturers, or the retailers, how we make sure we give the consumers the best possible experience by making sure that every location that they interact with has the best assortment of products that actually is tailored to the specific consumers that interact with that location. And so I think , um, you know , we're very focused on that area, but there are endless opportunities I would just say, well, I can't think I can hardly think of an area of a retail business or a CPG business that cannot benefit from understanding how the power of data and utilization of advanced analytics can help them streamline and , um , actually skill up their workforce to focus less on the kind of the broad word processing data on spreadsheets, and actually get information that enables them to make far more strategic targeted type of decisions to really offer a better, better product to the consumers.

Phil:

So let's, let's get into more specifics and real examples. Um, on the CEO of a retail chain, I'm in the middle of a pandemic. Um, I have supply chain issues. I have customers that I have to limit the amount of people who can come in my store at a, at a particular point in time. Uh, everybody in my store has to wear masks. Um, everything that that COVID-19 has affected with the supermarket , um, you're in my office , uh , safely distance and with a mask and a pitch me, you know, what, what can hive redo for me , um, at this time where I'm just all over the place saying, I just want to stay in business. I've got to figure out how to do delivery. I've got to figure out how people order online. I've , I've got to, you know, take all these prescriptions, if you would , uh, that hybrid can come up and I've got to grow business, or I'm going to go out of business. So what's your pitch to me and why should I buy?

Jason:

It's a really good question, bill outside and look at the end of the day, what you know , for the retailers and the CPG factors, CPG manufacturers, Highbury has the system that enables you to employ all of those real life. Real life, data feeds real life, world constraints. And as you rightly called out, COVID has put so many of those fundamental assumptions in question, it has changed the way we operate. We can, we put that into a system that reads that data in real time and tells you how each location should be , um , optimize the maximum sales and the maximum profitability. Now, there were a couple of things I'm going to pull out three we'll just straight off the bat that , um, where the, where the human needs to interact with that technology, thinking about some supply constraints. So I was talking in Australia, spoke to the board and Kimberly Clark , um, a couple of months ago. And you would think of your business like that with all of the panic buying that occurred earlier in the year would be absolutely on fire. Um, I think the reality is yes, sales are great, but the margin on the products is reduced significantly because they're needing to run, you know , facility manufacturing facilities, 24 hours a day. They're not able to get their products distributed quickly enough . Um, you've got limitations in store, you've got consumers shipping, you know, the click and collect and online more and more. I was really interested actually. What else, you know, what costs that you did on the ivory channel about still consumers prefer to shop at a brick and mortar grocery store about 78% of them. It's really, really interesting. I think when you look across the world as well, there's no single response or single situation to the COVID-19 situation. And here I live in Sydney , um, we've been very, very fortunate. It's pretty much life as usual. Uh , very few cases we've managed to contain things. You know, shops are open, people are out shopping, but then you look in some places where there are dramatic restrictions, but I say that to you, like we were able to, what we are able to do is in a nutshell, is we're able to tailor your offering to maximize the opportunity for each specific location. And that's what AI, and that's what data and machine learning enables me to do historically. Um, you know, you look at retailers , you mentioned more from , uh , you know, several thousand stores across the U S and they're planning across those locations. They have to roll it up to a level that makes it manageable to the people that work in the sector to plan out the assortments, you know, to, to go and execute what we're able to do. And then the industry have moved into kind of clustering. So, Hey, these stores behave similarly on that similar demographics. And so we tailor the offering. We actually say that , um , in the middle of all that clustering and kind traditional methods, methods of segmentation only get you part of the way. And we understand by looking at that utilization of those past data sets that you're talking about, that every location is unique, it has its unique fingerprint, or it's a snowflake and unique snowflake, and you can use data to one s tand uniqueness of that. A nd uniqueness of that location is based on the behavior of that location. So l et m e sort of bring i nto a little parallel, you know, u m, so traditional segmentation methods would say, l et's, let's compare me to my neighbor. My neighbor is, I think h e's about five or 10 years, less similar age group a s m e, same sort of background, probably similar financial situation, u h, w ith W est got two young kids. Um, and so we live in the same post , same ethnicity, both got children, you're brothers, we're brothers, you know, I like tenants and he likes classical music, you know, you know, or he likes football. And so that , you know , and , and what had , how do you, and so the demographic zip code socioeconomic segmentation works to a point because you can draw a generalization and trends . But if you think about what happens in an online environment, the first time you ever go to an Amazon, the first time you ever go to an online school, you might not have ever put your personal daughter and you might not have transactions so that they don't have your address. You know , they don't have your credit card details or anything yet, but what they do have is they know what products you've operated. They know what you add to your basket. You know, they know how you click between certain things and then as your backs and what they know is your behavior. And Hey, guess what? Back to that data collection we were talking about, I know that about every single person that's ever interacted in the website. And so all of a sudden they're able to see, Hey, people that start hovering on things in the same way, they start to be able to serve up a profile and recommend products to you based on your behavior, rather than the information, you know, your demographics, all those sorts of things over time, you then might buy something and you may store information. You might want a loyalty program that just adds to the richness of data . So what we look at, we say, we can understand the behavior of the consumers in your stores, by the sales patterns, in your sales data, by looking at it across all those, you know , several thousand stores. And therefore we can understand how each one should be unique. U m, and we draw all sorts of interesting insights to tailor the offering to make sure that the consumer experience is as closely matched to the product offerings that they want in that store. And then also efficiencies that are able to drive location. What that means is incremental sales, incremental value. You'll have less out of stocks. Y ou have products sitting on the fringe that, u m, t hat result in incremental sales to put this into context. T he last three years we've been working in, probably s ure with one of the biggest beverage manufacturers in the world with one of the largest ret ailers. A nd we took the entire category, wei ght, w ei ght gr ew the category, incremental sales, lik e $5 0 million ev ery year for the last three years, by just understanding how each location should be different, but not just understanding, building a system that enabled that to be executed, u m , i n the field. Now, now back to your original question, how does that, so what a lot of where we operate is kind of at that , you know, the process for planning categories is kind of an annual cycle right now. You think about that. Um , you know, for example, Walmart takes about 40 weeks of the year to plan next year, it's sort of assortment, right? When we've been doing this, we've actually put a solution in vending machines as well. That responds much more in real time, because if the drivers going out every day and there's an opportunity to optimize and grow sales, you can make changes. We worked in Japan and you don't have to have seasonal hot, cold switch overs. And so there's a whole mechanism about how you can optimize that in more real time, but there was a strong desire in the industry, but that for that cycle to not be 40 weeks to get down to something like as close to real-time as possible, I mean, I think they would aim for 20 weeks would be great. Our vision is to actually bring that down to live . You've got live data coming from your, from your sales force. You can understand what you talked about. Hey, we've got some supply supply chain issues in this particular category. Um, we've got less consumers coming in. I think one of the trends that we've seen actually, a nd t his t o your podcast again yesterday is w hich s aying, u m, a lot of the retailers pushing more and more going into the next assortments for the lower cost items. And that's because of the state of the economy. T hey w ant t hat. T here's a drive to say, H ey, we're concerned about t he purchasing power. We want to provide offerings that are more affordable t o the consumer, given the nature of the economy a fter what we've been through over the last f our, f our months and more, that's what we're going to be going towards. But , um, w w where this data leads to is a far more dynamic, far more reactive nature of the way the retailers and the manufacturers can respond to those market dynamics. If you operate in a , in a, in a system where it takes a huge amount of people and a whole lot of manual process and 40 weeks of the year to , to figure out how to assault your store. Imagine you can do that in , in minutes, and you can respond in real time with live data feeds. And then you have, you just need to think about the, you know, the labor equation about how you change the stores, but you can be able to look at it , look at the map of America, look at where all my schools are located and go, Hey, there's a dozen of them that are just really out of whack. You know, we , we , we , we , we've got, we've got such that they're not optimized out wrong , that there is particularly impacted by the current market conditions. Perhaps we need to focus in on that. We need to make some changes. So we get the offering, right, right now, that's , that's a much harder thing to do. So , um, to pitch that CEO, we can help you respond far more dynamically. We can tailor your assortments to the very specific and unique requirements that the consumers that interact with that location, and we can help by doing so we can help grow your sales and drive greater efficiency . And that's at the end of the day, what we're all about.

Phil:

So, Jason , uh, last question, if you had to look into your crystal ball and, u h, pick a timeline, you know, five years from now, u m, what, what will the supermarket be looking like that, u m, has Highbury and artificial intelligence embedded into it from the eyes of the consumer?

Jason:

That's a really good question. You know , I think at the end of the day , um , the way the consumer interacts with, with the supermarket is this is him . You know, I think as I think about myself as a consumer, I get my trolley on a walk down the aisle. I still enjoy that as much as I'm an elect to pick what I'm going to eat for dinner tonight. Um, there's a secret world that hits behind that, around what actually it's quite complex, quite niche. Um, there's a lot of different solutions and , and , and different factors that actually result in that offering to the consumer and what, you know, and I think the consumer just expects to see , um , the products that they want when they walk down the aisle, they go down the cereal aisle and we're working with some cereal play . Uh , some of the cereal manufacturers, they expect their favorite cereal to be that, you know , they walk down the beverage aisle, same thing. I think, you know, there are a lot of trends happening in the supermarket sector from, and you've talked a lot about these , these podcasts around , uh, around, you know, the trends that actually the pandemic has put to a massive poll around the white people, you know , the dining experiences and the salad bar . And what have you, I think what ultimately what we're about is driving efficiency, right? So what I would hope as the result of poverty technology embedded in these , in the sector, what not, it'd be a very, very rare occurrence that you would walk down the aisle with a shopping trolley, or you would show up online and , um, and you would find that the product that you wanted was out. So that's the first thing that you'll find this all . Second of all, you might start to realize that the products that they offer instead of being standardized, start to become a little bit more nuanced, and you start to see things that you're really interested in that you might've seen before. I think that's a really interesting construct when you think about perhaps as a trend for smaller assortments, no , maybe we don't need 300 cereals , um , in the cereal aisle when we go in and shop the cereals, but then if we go to go from 300 to 100 or even 20, how do we pick the right ones for that location? And as a consumer, you want to know that when you go to your local store or your favorite store, that they've got the products that most, that you're most attracted to, they'll be less out of stock . There'll be more products that they really want. And they'll also be products that they're really that perhaps they haven't seen before that they want to trial. And so I think, you know, what, there'll be less of that disappointment when they miss out on something that they were hoping to get. And that will also translate into that , that five-year horizon that perhaps we're talking about it , that'll also be very seamless as the, as the retailers start to , um, well, they're not already, they've already started really streamline their offering and focus the multichannel , you know, whether it's pure brick and blown up, you know , online click and collect online. All of those things are very connected. Um, it's really about driving the right choice for the consumer and the right experiments and the right offerings and the most efficient solutions. I think in the world that we operate, what happens is the stores get a little small phone , um, contrary to popular belief. Actually, this is actually been up until COVID at least the number of stores are increasing, but the size of them were getting smaller. Um , there was more space allocated to experience, you know, you're sitting down and having a sitting and having a coffee with your friend at the local school, maybe sitting and having a meal, some experience with brand activation. So the space for the product range actually becomes quite a bit of a premium. And so when we think about what we do, you know, we call it hyper-local retailing, very targeted, very specific to that fingerprint that we talked about, you end up with a great shopping experience because at the end of the day, the consumers are surrounded by the products they want they're available. Um , and then that is, that is then combined with all the innovations that the retailers are bringing to enhance the consumer experience and ultimately result in shopping, being, being, being , um, the experience that we all love and enjoy . I think, you know, what the pandemic will do, I think will drive an acceleration of the adoption of these technologies. So we can be far more responsive and far more dynamic in the way we respond to real-time challenges. I think that the sector has been a bit slow to adopt these technologies and therefore respond to the challenges that we've talked about. Well, Jason hybri wealth of knowledge, thanks so much for joining us today on Washington, the supermarket. And if you want to know more, just check it out@highbury.com .