Gary Saarenvirta on The Retail Code Podcast

Listen to The Retail Code Podcast

Daisy Founder & CEO Gary Saarenvirta joins The Retail Code Podcast with hosts Liza Amlani and Gary Newbury to discuss the key factors to successful retail tech implementation.

Original Source: Merchant Life YouTube 

The Retail Code Podcast Transcription

Liza: Welcome to The Retail Code. Today we have Gary Saarenvirta joining us who has an AI solution that we'd love to learn more about. And of course, we have Gary Newbury.

Gary N: Hello.

Liza: So, both Gary's. Gary Newbury, can you go first? Introduce yourself, tell us a little bit about what you do. And then we'll tell you a little bit more about the other Gary.

Gary N: Cool, we're gonna get started here today. So, great. I'm a senior exec on call specializing in retail supply chains and the last mile. My purpose in life is to help inspire retail business leaders to think big, be bold, scale, adapt, win one epic supply chain transformation at a time.

Liza: Awesome. And Gary Saarenvirta, tell us a little bit about you.

Gary S: Yeah, I'm the founder and CEO of Daisy Intelligence. Our mission is to improve people's lives using machine intelligence. And our goal is to elevate employees at the large corporations working retail to really allow them to focus on their mission, deliver value to their customers, and deliver shareholder value. And so, we believe technology should make people's lives better, make people's jobs easier, and not be about replacing humanity. And so, I think we feel strongly about that. And we're here to support humanity, not eliminate it.

What are the Factors to Successful Retail Technology Implementation?

Liza: Great. I love how you put that in. I'll just give you guys a quick introduction on myself. My name is Liza Amlani, and I am The Merchant Life. My focus is really around helping retailers with their merchandising optimization, especially as they enable these digital practices throughout their retail end to end. And of course, I act as, let's call myself an interpreter, as I come from the retail insider perspective, and giving retail technology companies like yourself just an insight into what challenges retailers actually face. So, I'd love to go into the first question we have for you, which is really around the key factors contributing to the successful adoption, and maybe the not so successful adoption, in retail technology implementation.

Gary S: I can speak from the perspective of implementing analytic solutions, you know, so we have AI and our goal is to help merchandise planners with activities like promotion planning, price planning, and inventory optimization. And I think big challenges is, you know, the change management for us is the biggest challenge. And I think today, there's a lot of people who do these jobs today, you know, like every retailer plans promotions, and prices, and forecasting, and inventory. So, all of these activities are done. So, bringing technology into the picture, I think the other challenge is, how do you change the process and the roles of people to use the technology in the most effective way, and that change management is very, very difficult. And so, having executive support and buy in, you know, there's a financial value proposition that executives love, but then there's a user value proposition, what's good for the people who the technology, the merchants, category managers, and planners in our case. And so, technology companies tend to focus on the financial value proposition, and not so much on the user proposition. And so, we have decision makers who get the value, but the users who live it every day, that we need to have a value proposition for them. And so, our focus has been to try to make the job easier. We're trying to figure out what that means. And so, I think where we've seen failures in this change management, the inability to deliver value to the people who use the technology that they use day to day, whereas executives love to see the dollars and cents come in the bottom level line, but that doesn't excite the everyday employee who does the job.

What Represents Good Change Management?

Gary N: You know, there's a couple of things I'd like to, sort of, dig into. The first one is, you mentioned that, you know, executives want to see the financial return on investment, ROI. And I find time and time again, with things in my area, things like warehouse management system, travel management systems, don't get the air of day, because the weakness in the business case doesn't take a wide enough view of all the benefits and maybe the user proposition is definitely not on those. It's like how a pig is going to be feeling today, oh I don't know, forget that, we can’t put a financial figure on that. And the other one was change management, I'd be really interested in what you mean by that to just to get a deeper dive on what does that actually, you know, what's involved in it, who's involved, what's involved, I guess we're talking about stakeholder management. So, I’d be interested in your views as a company trying to get into retail and advance your solution form. When you're seeing these opportunities move forward, what represents good change management, and perhaps no change management, which I guess is the opposite?

Gary S: First, I talk about the value. I think, yeah, I agree with you that you need to measure holistic value at the level of the company. And I think that's what I've learned. If you're just looking at a slice of the company saying I'm improving this process, there's all these interactive effects, and if you ignore the interactions, then you may net out to no benefit at all. And so, we try to measure the impact on the company itself. So, you know, I want to open the annual report and look at the net income at the bottom of the P&L and say, there was value created and we can actually attribute it to the technology work. And so, we work hard to measure that. And I think if you don't measure that, then some of these business cases are weak, because they don't take into account that whole picture. So, I think that's how we try to address the kind of, value, and it has to be holistic, right? And I think there's very few companies that measure a holistic value in terms of technology players, because most of us are technologists, not business people. And so, you know, and it's something I've learned over time, and having failed at this and beating my head against the wall for many years. It was not through great wisdom that we've gotten to this place. And to answer your second question around the change management, you know, I think the challenge there is, one is the belief that technology can do what I do as a person, right? So, there's building belief in the human beings that that's actually possible. And I think there's this skepticism that we have to overcome, that's one big barrier. And that's the technology's responsibility, is to educate the people along with the technology change. So, you're not forcing something, you want to make them part of the journey. And I think the important part is there's a journey, and you need to start from where they are today, and meet them there, and then change slowly. I think, in the early days, we tried to push too much change too quick, to say let's go to fully automated from where you are today. Epic disaster because we didn't bring the people along. And so, we've, kind of, made this multi-step journey where, at every step of the way, all the people who are involved are coming along, learning with the technology, and buy into it in stages. And so, that attitudinal change is really about education and making sure that we do that. And that, you know, I think that's important, that the amount of change that you're requesting from the retailer is commensurate with the financial value. So, even though there's great financial value, if it causes chaos, then it's not going to happen. So, I think that's one element of change. And once you buy into that attitude, and the willingness to do it, then you have to change processes. You know, technology will impact the process, sometimes big, sometimes small, and change maybe job roles, if the technology is now taking on some of the job role. So, I'd say process change, job role change, and maybe even compensation change, the, kind of, secondary changes, that once everyone buys into, we're actually going to do this. And I think a lot of technology fails because we don't get the actual buy in at the beginning. And so, for us, it’s this crawl, walk, run approach that we've learned because technology people want to run, you know, we want to fly in rocket ships and everyone's walking. And that disconnect is way too much and we failed many times because of that.

What is Daisy’s Value Proposition When Implementing Technology?

Liza: Yeah. And I think you bring in such a good point because, coming from 20 years in merchandising, I can’t tell you how many times a technology has come in and changed the way that I work, but not understanding what are the elements of the activities that are part of my role and what that fear looks like. Well, if you're going to take away parts of my buying and planning function, what does that mean for me as a merchant? Do I still have my job? And on the flip side, in the consulting world, you know, working with a similar type of implementation where we're digitizing parts of the process of, let's say, the line plan, or product creation, or even category management, the challenge is that if you do not get those folks that are actually doing the work on board, and maybe even part of the decision of picking the software, that's where you get those change management and change of behavior, let's call them hurdles, where it is so difficult to move forward. So, how would you say, when you see something that you're implementing is not being, let's say taken positively, how do you cross that hurdle yourself? Do you say well, I know what you've been through, and I understand what you're saying, and this is how this is going to change your life and remove manual tasks? What is your value prop in that respect?

Gary S: I think in retail there's a lot of manual activity. And so, ultimately, we want to automate the low value manual activity. Our goal is to elevate the person, elevate the merchant, elevate the category manager. So, I’ll use an analogy, for us, our technology is like that in a fighter jet. The pilot decides where does he or she wants to go, pulling on the controls, but then the computer actually executes the detailed flight instructions and does 1000 control adjustments a second. That's what we hope. The human is the pilot, the human is the boss of the technology, and the computer takes care of the gory details. And I think that people don't want to do the gory details, like relentlessly deciding every week what should I promote this week, and next week, and next week, what prices in all my price zones, how much inventory in every store, technology can help with that, kind of, relentless volume of decision making. And so, I think the first step we do is, you know, look at what you're doing today and say, hey, how can technology, let's evaluate backwards what you did and educate you, what does the technology see that you don't see, and then educate you and say, hey, it's not doing anything crazy, it’s automating some of the low value tasks, get the buy in, you know, the retailer is definitely in control. So, backward step first. Backward step and education. Once you have confidence, merchants go, okay, let's put your money where your mouth is, let's use this in a forward looking way. Again, merchants, category managers come up with a plan, let the technology say here's the changes I would make to your plan and make recommendations, merchants still in control, humans still in control, and then they say, okay, well, let's try a few of those changes, let's see if it works. And all the changes recommended, we've learned about them, because we did this in the backward looking way, it’s all consistent, right? We try it, now it's working. Okay, now let's flip the role, let the technology make some recommendations and let the people be the boss now and oversight, you know, then great, that works. And then finally, could be full automation. So, those are what I see the four stages. And every part of the way the people are in control, the people are the boss of the technology. And I think if we continue to reinforce that story and really focus on what are your challenges, and we really try that. The most valuable time I spend is sitting with retailers. And really, you know, that's when I realized, oh, my God, we're delusional, you know, when you see the real people use the technology, and, you know, I really say, okay, wow, that was a great, you know, we learn and how do we really help the people. And so, I think, you know, that's how we try to overcome it, and how we've learned from, you know, bad experiences where we failed miserably.

Can Data Requirements Be a Blockage to Successful Implementation?

Gary N: One of the things, because you're a specialist in AI, so again, when I converted to AI in my own area, when you apply AI into a transport situation, and you're trying to help the routers do more effective job, the AI will come up with what appears to be incredibly randomized pattern of how you would move the vehicle instead of a typical route, which will be like that, that'd be intuitively correct for a router. But when you bring in other elements, like costs, and time of day, and all things, it will go all over the place if we try and meet time windows or a specific requirement. So, when, in the areas that you specialize in within retail, what are the reactions when you start running some models? And what are the other data requirements that they probably may or may not have at that stage? And whether or not they bring them in, or have a look at it, we don't have this stage. And so, how do we, is that sometimes a blockage to successful implementation?

Gary S: Yeah. I mean, retailers have the data that we require. I mean, so we look at transaction log, you know, sales data. With the data, it’s just maybe not managed the way we would like it ideally be. So, it's like the product hierarchy. You know, what we would like, you know, when retailers are planning promotions and merchants and category managers say I'm going to promote blueberries, right? But behind the scenes, a large retailer will buy blueberries from 27 different suppliers, you have 27 different SKUs for blueberries, and some of them are only valid at certain times of the year because you buy this. So, blueberries becomes a very complex thing. For the category manager, some will do it, somewhere behind the scenes the technologist will go and pull all this data together, but that's not managed. And so, some of the work, we need to do extra data management work to clean up some of the data, but most of the data does exist. It's just that the way that retailers don't require the data to be this perfect because they don't use it. But now that the technology is making the decisions and doing recording, the data does have to be much better for that purpose. And so, we spend a lot of time augmenting data. And then we ensure that we're using all the same inputs that the retailers use today. So, you know, using competitor information, looking at market data, you know, from, kind of, the market data vendors, you know, looking at some consumer research, because they feel confident that, hey, this plays part of our decision, well, your technology better use some of that data too. So, we've brought in extra external data sources that we didn't originally do that just because that builds confidence in the merchant. So, you know, a lot of data cleaning work, and then making sure that we're adding all the same inputs so that there's confidence in the outcomes.

What Happens When the Data Doesn’t Tell the Full Story and You Get It Wrong?

Liza: Yeah, I know. You got my mind thinking in a lot of different ways here. What happens when the data doesn't tell the full story and you get it wrong?

Gary S: So, we have a different approach. Our technology doesn't learn from the data, you know, so the scientific method, the way science has been done for hundreds of years, is you come up with a theory, you express it mathematically, second you go get the data. So, if Einstein waited for big data, we wouldn't have a theory of relativity and we wouldn't have quantum mechanics, and therefore, we wouldn't have these laptops that we’re talking to each other on, right? So, the idea of data first and then learn something, that's backwards. I think the computer science world has turned that backwards. And so, what we've done is created a mathematical theory of retail based on fundamental truths, which are customers buy meal solutions in grocery, they buy full meals, not items. So, customers buy solutions. In hardware, if you're going to paint your house, you buy paint, but you also buy drop sheets and rollers. So, there's a solution. And that's a fundamental input to the theory. And that means there's also cannibalization, because you bought one product, you didn't buy another, because I bought a brand of paint, I didn't buy another one. So, there's a consumer choice, that’s cannibalization. And there's pantry loading, because I have on sale, I bought a forward supply of it, just like retailers fill their warehouse with forward buying, consumers do the same with their pantries. There's seasonality, what you buy on a religious holiday is different than what you buy in the normal part of the year, what you buy in winter and summer are different. There's price elasticity, there’s promotional elasticity, competitors. So, those are the factors. And everything I said, every retailer goes, oh, yeah, I totally get it. And then we just turn that into a mathematical formula. And then we said, okay, where is the data about these factors. And that's what we get from the transaction log, from the promotion data, from the product master. And then we plug that data into our map. And so, we try to assemble a holistic theory of retail. And so, that's the approach we take. And so, the technology is not perfect, but it certainly adds incremental value to what the retailer is doing today. And we prove it, you know, we have leading indicators that say, hey, we can predict sales 12 weeks out based on our mathematical theory. And when we get it wrong, that's where the AI goes back and tries to recycle and learn. And so, the goal is incremental value, not perfection. And that's sometimes harder for people, people say, see, but you're not taking that into account, and we say, well, I know, neither are you today, we're not doing it either. But what we're doing is still better than what you did yesterday. And that that's a tough, that's a change management mindset thing too. So, that's, kind of, a unique to us approach.

Are Executives Resistant to Technology Implementation?

Gary N: We talked about making the guys who, again, actually the use the technology, kind of, let loose on an adventure of learning, as opposed to you're going to get your job done, you know, you're out, to help build their confidence. But as you’re talking to executives before you get the commission to actually implement, really start the implementation, what are the barriers? Do you find that some executives are, kind of, a bit, yeah, sure that works, but actually, what we do is really brilliant, because I know it? Whereas what you're offering is, well, what you're potentially doing, is saying everything you might have known in the past, that you thought was working really, well, we might actually see that it isn't working so well, and I can actually show you that there's an improvement, that maybe your boss might look at you and go, why have I employed you for such a long time as an executive here? Did you find that there's a fair executive level? Because many executives in retail are of a certain age group, certain demographic, and they may have been successful 20, 30 years ago, very successful when the world was a lot simpler, but the world has become much more complex. And as a result of that, there's more technology and more potential threats to, kind of, uncovering that they don’t know everything now,

Gary S: I think there is resistance we see. I think one thing we try to position is this as an empathetic sales approach. And say, look, everything you've done has been wildly successful. There's absolutely nothing wrong with category management and merchandise planning. It's been great, like, look how successful you are as a business. And you can keep doing that, because most companies are, but we're saying the world has changed. Look at the last year, definitely we can all agree the world has changed. This technology has been built to take advantage and help you through some of those changes. Now, we know change management is really hard. And you can keep doing what you're doing. But if you want to explore this together, you know, that's an approach we take, so that it's, kind of, not a competitive, hey, this is better than what you're doing. And then I think the barriers that executives face, you know, they love the financial value proposition. Everyone likes that. But it's, you know, do I have other priority initiatives, because the company only has so much bandwidth. So, we run into, hey, we're in the middle of a big SAP implementation, and my point of sales system is changing, so when that's done, then we're okay to talk to you. That's one kind of resistance. And that makes sense. You know, companies can only deal with so much change at once. Or executives don't have the technical understanding. So, a lot of what I described is very technical, you know, the fighter jet analogy and the Einstein theory. So, you know, everyone says, yeah, we have AI, I’ve got an analytics department, we're doing everything you said, we're doing that already. And so, because they're not the deep technology understanding, so getting into the weeds of the technology to explain the differences, that gets really hard. And so, executives sometimes say this is AI, well, we got AI, my people tell me I have AI, so we're already doing it. So, I get a lot of that pushback as well. So, then we don't have the IT resources. That's another big pushback I get. IT is overwhelmed. And they're thinking, you know, I'll put you in the IT plan and we typically end up waiting. On the worst case, I waited two years for IT to support us in one case. And so, you know, IT bandwidth is a challenge. So, you know, having those are the typical barriers we run into. And we try to, you know, we have approaches to try to address those. But, you know, things like that are typically challenges.

What is the Need for Technology Like Daisy’s?

Gary N: So, I think we've learned quite a lot from this conversation. The barriers to even have an opportunity to move the technology forward inside the organization. And then once it's inside the organization, some of the challenges, not necessarily obvious, but certainly methods of how to engage staff who are going to be directly affected by technology and its implementation. And trying to work out strategies of how to get them on board in an efficient way that allows them to build their confidence, that this isn't going to affect my job in a negative way, it's actually gonna save me some time so I can actually spend more time thinking about the things I should be thinking about, because I'm not having to worry too much about the transactional things. And, Lisa, you want to add anything to that?

Liza: Yeah, honestly, Gary, you had me at I'm gonna sift through the data for you. You know, coming from, big companies like Ralph Lauren, you know, back in the day, we had lots of data, we just didn't know what to do with it. I think we're going to end today's conversation around, you know, what is a big bold statement that you'd like to give our audience on the use and need for technology like yours.

Gary S: Technology should improve our lives. And I think for retailers, we want to make the job easier. Retail is a tough, tough business. There's so many moving parts. Companies like us, we're just here to help. We just want to help. You're the hero of the story retailer, we want to help you do a tough job, make it a little bit easier. And in that way, service consumers, service society in a better way. And, you know, that's our motivation. And it's about helping humanity and changing the world, making it a better place. And that's what it's all about. And AI offers that promise. AI done right. We have a philosophy of AI done right is about some of the things we talked about today. So, happy to have been on the program, happy to keep chatting, we'd love to chat more.

Gary N: Well, on the note of AI is going to change the world and make the world a better place, on that bombshell, I think it's a wrap. So, thanks, Gary, for contributing to this discussion. And, Liza again, and I look forward to continuing on this journey.

Liza: Thanks, guys.