Walmart Inc. (WMT) Earnings Call Transcript & Summary
January 14, 2024
Earnings Call Speaker Segments
Linda Lomelino
executiveHi, everyone. Thank you for having us. It's an honor and a pleasure to be here. For those of you who don't know me, I'm Linda Lomelino, Senior Director of Product and Product Design at Data Ventures, Walmart. I'm joined by Erich Kahner, who's Director of Competitive Strategy and Insights at dunnhumby. We're really excited today to share with you what we think 2024 has in store for us in retail. What's top of mind for customers, specifically in the context of economic uncertainty, fierce competition and limitless assortment, and how can data and insights really help drive customer centricity in a competitive landscape where loyalty is not only harder to achieve, but even harder to sustain over time. Erich is going to walk us through what we think are 4 critical themes for all of you to consider as you think about your year. And also, I'm then excited to share with you, Walmart Luminate, which is Walmart's proprietary suite of insights, and to illustrate how our ecosystem of data and analytics has really helped one of our biggest suppliers use customer data to drive category growth, again, in a very fierce competitive environment. With that, I'll hand it over to Erich, who will walk us through the big themes for 2024.
Erich Kahner
attendeeThanks, Linda. As Linda mentioned, we're going to talk about 4 macro themes to keep top of mind as we head into 2024. I'll talk about each of the themes in more detail over the next several minutes, and then Linda will show us Walmart and dunnhumby solution for retailers and suppliers to collaborate together to combine a deep understanding of customer behaviors and beliefs at speed in order to find growth in a highly competitive, crowded and mature market. On this first theme, 2024 will be challenging. It will be just the latest in this long evolution the market has seen toward customer empowerment. For 2024, dunnhumby projects year-over-year food and consumables sales growth in the U.S. of around 1%. And this will be slower than the rate of inflation, which means another challenging year for units. And of course, some product categories and retailers will beat the market, but it's going to take tight alignment with customer needs in order to do that. So in order to understand what tight alignment with customer needs looks like at the 30,000-foot level, dunnhumby conducts its Annual Retailer Preference Index study. In this study, we model performance of the country's 65 largest food and consumables retailers on customer opinion of prices, promotions, quality of the assortment, quality of the digital and brick-and-mortar experience against outcomes retailers want to achieve, for example, better sales growth, stronger customer share of wallet, stronger emotional connection with shoppers. So what this tells us is which of these opinions is most important in driving financial success, and with that knowledge, we can essentially create a weighted brand equity score for the 65 largest retailers in the U.S., and we're able to rank them from the strongest alignment with customer needs to weakest. And the return on aligning on customer needs is high. If we look at 5-year sales CAGR, retailers who are in the first quartile in our ranking have sales growth that's over 2x stronger than retailers in the bottom half of our rankings. And this translates to about $2 billion more in sales over a 5-year period for your typical average regional supermarket chain of about 200 stores. So obviously, this return on customer equity is much higher for national chains. Driving better customer opinion has huge stakes, and we're going to hear more from Walmart today, which is speaking from a position of authority. Walmart's 3 largest banners, Walmart Supercenter, Neighborhood Market and Sam's Club, all ranked in the top quartile in our study. So why is that? Well, when we look at which customer opinions are most associated with financial success, perceptions on levers that save customers' money routinely comes out on top each year we do our study. And in the past 2 years, this is more true than ever. Additionally, the importance of digital touch points has been trending up steadily since 2018. And when we look at Walmart banner strengths, the commonality is saving customers money and a strong digital customer experience. And a quick note by what we mean about digital. There's a few customer opinions that really matter the most there, easy online shopping, strong performance of the app, making a shopping experience easier and having technology that makes the in-store shopping experience easier. And you may notice this last pillar, speed and convenience, that's been declining in importance over the years. And that's largely a brick-and-mortar construct. And it's been decreasing in importance, I think at the same time and because of the increase in retailers' digital capabilities and also, customers, when they're visiting stores, especially the past 2 years, are really taking the time to find savings. Another thing we've noticed in our study, 2023 was the year in which we saw opinions of retailers changed the most out of any year we've done our study. This tells us customers are taking a step back and really scrutinizing retailers. They're asking themselves, is this retailer still best aligned to my needs. This openness to new beliefs presents a big opportunity for retailers, if they make the right moves today and together with their suppliers. When looking ahead, we mentioned 2024 was going to be a competitive year. And as with most things, this will largely be consumer-driven due to the headwinds they'll be facing. Lower savings stores to draw upon than they've had the past few years, rising and still expensive debt, slower disposable income growth, rising unemployment. So with that, I'm going to turn it back over to Linda for the so what and the now what?
Linda Lomelino
executiveThanks, Erich. So let's click down on the customer a little bit. So customers, as we all know, are increasingly selective about when, where and how they open their wallets. According to SAP, U.S. customer loyalty to brands has declined over 14% in the last year and 59% of Americans surveyed would switch brands for a cheaper option. So you've got a very discerning customer. We ran our own survey at Walmart with our own customers and those findings echoed SAP's findings as well. 58% of our customers spend more time seeking values, deals and discounts. 54% of them are willing to switch at any given time if they can find a better value out there. 53% of our customers compare prices across retailers before they make a purchase. So you've got a very intelligent discerning consumer out there. And so given this discerning behavior, it's become even more crucial than before to make data-driven decisions that will help brands and retailers like ourselves continue to stand out. At Walmart, we obviously have a very distinct advantage when it comes to understanding not only the Walmart shopper, but the U.S. customer and the retail landscape. We have over 4,000 physical locations across the U.S., accessible to 90% of the U.S. population with over 5 billion transactions a year. Those are not items. Those are baskets. And we have a growing online assortment. And so we possess a very detailed view of the American customer, including preferences, behaviors, attitudes, loyalties, switching behavior, et cetera, which is why we decided 2 years ago to launch Walmart Luminate, a unique suite of data and insights products that makes our users smarter in their decision-making process. They can make better, faster, smarter decisions, always with the customer in mind. Our merchants and our suppliers are leveraging Walmart Luminate to gain a competitive edge and drive growth in these uncertain times, and I'm super excited to share one of those case studies today. Suppliers and merchants like Mondelez are becoming more customer-centric by leveraging insights from Walmart Luminate to drive their overall strategies, not only how they think about their operations, but also how they think about their customer. You guys are all familiar with Ritz. It's a very popular cracker brand, and it's extremely successful in the market today. But Mondelez was not going to settle for success. It wanted to explore additional opportunities and decided to find category growth in a category that's already highly competitive and highly crowded. Given the competitive nature of the cracker category, think about an aisle crowded with brands, crowded with options, you've got numerous flavors, value propositions, Mondelez decided to analyze shopper data to gain a deeper understanding of the Ritz shopper and the Walmart shopper and what they wanted to do with crackers. This becomes a critical assessment to do when you think about shelf space becoming more and more limited and shopper preferences continuously changing. So you have to be able to access data nearly in real-time to be able to make smarter decisions. And this analysis that they did leveraging the Walmart Luminate ecosystem, allowed them to not only evaluate different forms, flavors and sizes at Walmart, but it also helped the Mondelez company uncover the role of the Ritz brand within the portfolio of crackers and help the merchants better understand their customer. So Mondelez used Walmart Luminate to analyze customer preferences across the entire cracker category. They have access to category data, not just their own data. They found that 50.6% of their shoppers buy carrier crackers such as Ritz, which complements other snacks within the box. This led to increased interaction in other subcategories like saltines, cheese crackers and other comfort foods as well. Each subcategory, we learned through the data, fulfills a very specific job to be done, ranging from entertainment and charcuterie boards to snacking, to feeding kids and even comfort food. Furthermore, because of the large increase in both -- sorry growth in both pickup and delivery, 80% of the shoppers of this category were also loyal Walmart customers. After gaining a solid understanding of the entire category, the category of crackers, Mondelez then decided to double-click on their own brand, Ritz. It became apparent through the data provided within Walmart Luminate that the Ritz shopper is not only loyal to specific flavors, but also that there are different purposes for different pack sizes. When they were armed with this knowledge, Mondelez was able to address questions also coming from the merchant about whether or not it was valuable to provide all of these distinct flavors, all of these distinct pack sizes. And in fact, there was. So this discovery was significant as it allowed Mondelez to effectively communicate to the Walmart merchant how each pack type caters to a specific shopper, a specific customer need and a specific journey. And so it emphasized the role of all of these different flavors and varieties. But Mondelez didn't stop there. They continued their efforts to better understand the Walmart shopper and the cracker category. So they used one of the products within Walmart Luminate called Customer Perception and this leverages our proprietary community of Walmart customers to gain insights about the category. So they targeted customers who had bought the category using our hyper targeting capabilities, and asked them about their brand preferences, their brand associations and their awareness of this entire variety of SKUs. It turns out that Ritz is a signpost for Walmart customers. Walmart shoppers who buy Ritz, are not only aware of the flavors, but report enjoying every single variety. The power of the Ritz brand became super clear through the work that they've done. Ritz not only was cherished by shoppers, but there was a nostalgic flavor to buying Ritz crackers. And the different flavors, as mentioned earlier, were driving very specific needs. So for both Mondelez and for us at Walmart, it was really powerful to be able to quantify the actual value of this entire flavor block that Ritz provides and further solidify the role of Ritz as a category giant and influence a lot of other long-term strategies. Data from Walmart Luminate revealed that Ritz has successfully catered to various need states such as entertaining and serving as an ingredient in meals. And this allowed Mondelez to adjust some of its long-term plans with the Walmart merchant, prioritizing the customer and building future strategies to ensure that Ritz, again, remains top of choice, especially when it comes to Walmart shoppers. And again, in a competitive landscape, it's about also defending territory with data. The strategies that Mondelez decided to shift include the following: focusing on the omnichannel shopping journey, and they identified a potential growth if they focused on these customers that were doing online grocery and online delivery of over $200 million, using seasonal moments like Thanksgiving and Christmas and the upcoming holidays to really work on a purposeful connection with customers about being able to succeed in entertaining, hosting the holidays overall, and then worked with our sister team, Walmart Connect, which is our advertising arm, to drive very exclusive deals online as a way of activating customers and driving a lot of traffic. So the overall success of using Walmart Luminate's data and insights portfolio to drive deeper, better understanding of the category and the customer is now also leading to new work for Mondelez, leveraging Walmart Luminate in the category of cookies as a great example, where they have brands like Oreo and Chips Ahoy! So looking ahead at 2024, it's clear that it's going to be a highly competitive year and a fast-moving one as well. Customers are smarter. They're reevaluating their preferences and are more open to switching products and exploring different retailers, different brands, different flavors. Additionally, they're becoming increasingly more selective about when, where and how they shop. They've got a lot of different options out there. So in response to these changing consumer behaviors, consumer attitudes, it's really critical for businesses like ourselves and our suppliers to make sure that we're always making customer-driven decisions and that they're based on real data. So by staying attuned to these customer preferences, shifting landscapes, surprising trends that are viral and take us by storm, we can position ourselves for success in the upcoming year. So in the retail environment that's marked by a lot of economic uncertainty, strong competition, shifting customer preferences and a more discerning customer, a deep understanding of our customers is not only advantageous, but it's now become mandatory for success. Thank you.
Erich Kahner
attendeeWe have a few minutes for questions, if there are any.
Unknown Analyst
analyst[indiscernible]
Linda Lomelino
executiveI think the question was, how do you see [indiscernible] adding to the success of this? We -- actually, we already see some -- so thank you for asking that question. Walmart Connect is our advertising team at Walmart. And we've actually seen some phenomenal work being done as a result of leveraging Walmart Luminate. And so as suppliers better understand their customers and not only understand them broadly, but then understand the different segments within their customer base, they've been able to be very intentional about their media planning strategies, and they can focus on very specific use cases like bringing back lapsed customers or activating against switching behaviors. And so we've seen some fantastic work. And as part of our road map, we're going to continue to scale that work with Walmart Connect.
Unknown Analyst
analyst[indiscernible]
Erich Kahner
attendeeThis is a follow-up question about the annual study we do. Just, I guess, what are the key factors in defining which elements of the value proposition matter more for driving retailer success? And I mean, we saw saving customers money being the pillar that's most associated with better market share, stronger growth. And that's, as I mentioned, more true today than ever. But if we look within the levers of what saves customers money, we've seen momentum in the personalized levers such as personalized rewards from a loyalty program, more relevant promotions. There's been momentum in that lever in the past few years. And that's come at the same time as this increase in importance of digital elements of the experience. And if you think about it, it makes sense. Those elements of the digital experience enable more personalized ways of saving for customers. So I would say if we're looking for some momentum in those levers, too, that's one story that we've seen. And also something that probably the top tier has most in common tends to be a really strong private brand portfolio, and that's because private brand, I think, as you all know, fits on element of saving customers money, but also on addressing the quality side of the equation, too. So there's a lot of momentum behind retailers, and we've seen Target make some improvements too and Walmart as well. I think the largest brand in the U.S. is great value branded products. So strong private brand portfolios lead to success as well.
Unknown Analyst
analyst[indiscernible]
Linda Lomelino
executivedunnhumby is one of our partners for Walmart Luminate, but we also have proprietary technology that we couple with a lot of the dunnhumby analytics. So it's a combination.
Unknown Analyst
analyst[indiscernible].
Linda Lomelino
executiveYes, absolutely. So the question was, as our suppliers are starting to leverage Walmart Luminate, are they finding meaningful difference regionally? Yes. The answer is yes. And what this allows them to do is a much more tailored strategy. They have access to data they've never had access to before. Even if you think about different data providers that they may have gotten data from in the past and probably still do, it has never been at the granular level that we're providing. And so they can really shape their strategies and tailor them not only on regional differences, but do we have a more price-sensitive customer? What does your best customer look like? Are they shifting to a different brand within your category? Are you losing customers or gaining customers? Are you growing at the same rate that Walmart is growing, right? And so there's fantastic analytics that we provide within Walmart Luminate that's allowed them to do a lot more of a tailored strategy. Great question.
Unknown Analyst
analyst[indiscernible].
Linda Lomelino
executiveFor Ritz?
Unknown Analyst
analystYes.
Linda Lomelino
executiveSo the Ritz work that we presented today was more of an in-depth understanding of the customer. Now they're putting into action a lot of what they learned. But the beauty of it was really understanding that there was a purpose to every pack size and a purpose to every flavor. Sometimes when you have a large assortment, you're trying to better understand what drives that category? Is it determined by brand or pack size? And that's one of the fantastic insights that you can get from Walmart Luminate is better understanding the customer decision tree, but then also the role that every flavor, pack size or whatever drives your category, what role does it play for the customer.
Unknown Analyst
analyst[indiscernible].
Linda Lomelino
executiveYes. And I think that's one of the most exciting -- so sorry, the question was how do you drive execution with this many suppliers, this many -- and it's a great question. So what I think is most fantastic about the Walmart Luminate ecosystem is before, you would have suppliers bringing data to Walmart and Walmart having its own data. We're talking now about a shared one truth, right? And so you're essentially looking at the same problem together, meaning I think it actually removes barriers to execution because you can design strategies collectively. So we've heard fantastic feedback from our merchants that they're leveraging our platform, and then also from our suppliers, who're able to walk up to their merchant and say, hey, let's do something about X, Y and Z, right? And so I think the execution piece is fantastic. And then certainly, the connection point to Walmart Connect, right, being able to activate -- because really insights for the sake of insights is meaningless. But what we've seen as similar to the Mondelez strategy is they're really starting to tailor their strategies based on the data that they have and the folks that are doing it are being very successful.
Erich Kahner
attendeeOne other question from the crowd. Do you share category benchmarking information with your merchants like Mondelez? And if so, how granular do you get?
Linda Lomelino
executiveSo we show category transparency within one of our Walmart Luminate products, Shopper Behavior, which is done in partnership with dunnhumby. And the beauty of that is it actually drives extremely healthy competition. And I think principle to Walmart is always putting the customer at the center of the decision-making process. So it's less about who's right and it's more about what does the customer want and need. We've seen a lot of successful suppliers say, hey, I noticed that maybe if I delist a few items, I can drive even increased category growth or maybe there's a product that's not really fulfilling a specific need. And so we provide category transparency for that very reason. I think that's it. Thank you so much.
Scott Light
attendeeAll right. Welcome, everyone. Thank you for joining us this afternoon or late morning. My name is Scott Light. I'm the Vice President of Churchill Systems. Churchill is a software development business based in Southeastern Michigan, where we specialize in providing AI technologies to major retailers here in the U.S. and throughout the world. In addition to machine learning, Churchill also works quite a bit in the areas of expert systems, a number of different AI technologies, exponential smoothing, genetic algorithms, fuzzy logic and so on. Joining me on the stage here this morning, I have 2 of veteran retail experts with me. In no particular order, we've got Randy Salley, former Senior Vice President of Walmart's ISD Division; as well as Mr. Dale Cade, Principal Consultant of the Columbus Consulting Group, also a former Director of Walmart's ISD Division as well as the Fossil Group and The Container Store. Now, as most of you have seen, AI is prevalent just everywhere in retail today. It's really become just in every type from customer service to merchandise planning, supply chain, pricing and promotions, even generative AI and ChatGPT. And what we wanted to do today is take a few minutes and really discuss some of the practical and effective approaches to integrating AI into your existing retail planning systems. So without further ado, let's get practical. Let's talk about some of our experiences that we've had in integrating AI into all sorts of different retailers. So what we're going to do is go through a few different questions that we've come up with from all of our years of experience in the industry and try to talk about where to get started when we're talking about integrating AI into a retail environment, whether you're brand new to some of these more advanced applications or whether you're looking to upgrade from your existing planning systems and looking to provide a little bit more science and technology to the existing environment you have.
Scott Light
attendeeSo we'll start right off. And Randy, I'll start with you. Where do you see these AI applications really providing the most benefit to some of these retailers in the world today?
Randy Salley
attendeeYes. So Scott, as you walked around the floor here today and yesterday, AI is almost in every booth. I mean it has been without a doubt just kind of the hype, if you will, for this last year with ChatGPT getting exposed, et cetera. And I believe there's tremendous value for retailers by using some of the generative AI technologies. But I think the bulk of that will probably be back-office kinds of functions, potentially integrating with your e-commerce search capabilities, et cetera. But for decades, we've come to this conference and other retail conferences, it's always been about getting the right product at the right location, at the right time, at the right price. I mean we've heard that mantra forever.
Scott Light
attendeeSure.
Randy Salley
attendeeAnd so when I think about AI within the retail space, I tend to think about those kinds of opportunities, which leads to how do you do your demand forecasting? How do you manage your seasonal profiles? What do you do with price and promotion and the impact upon that against your products? And those types of implementations of it, I think, is where you really see the benefit because that's the challenge that we're all faced with every day is how do we continue to provide that product at the right place, time, quantity, price for the consumer.
Scott Light
attendeeSo you see it as leveraging the science to those planning systems, supply chain, pricing, merchandise planning to really try and pinpoint where you're going to get the most efficiencies when you're placing product, when you're deciding what pricing to do, what type of discounting to do, the effectiveness of your promotions, that sort of thing?
Randy Salley
attendeeWell, and truly understand your demand. I mean that sounds like a simple problem, but that's a really hard problem because there are so many components that run into that, whether it be lost sales or weather or promotion in the case -- I mean I grew up at Walmart so we're very promotion-focused, and we were an everyday low-price company. So promotion didn't play a big part of what we did, but being able to understand the impact of a particular promotion on an item and ultimately, what is that going to do to demand?
Scott Light
attendeeVery good. Dale, [indiscernible] yourself?
Dale Cade
attendeeSo Randy did a great job talking about situations where retailers need to take a lot of information, channel it to a decision. I think there's another opportunity where you want to expand that out. For example, you're building content for catalog, your PDP, your ELP, marketing where you're taking -- you have information, but then you want to figure out how do I take that and expand that out to a broader world? The generative AI is excellent on that. And then coming back again, once you've taken that marketing approach and you put that information out there, and you've got customers that have done something and you gather that information and you channel back again, and it becomes a new feedback to that replenishment loop. And so you have the ability to do that. The nice thing about generative AI is you can take a little bit of effort, you can create a different view, different perspective and you can quickly cater. You have these huge companies that have tens of thousands of SKUs. I want to get it to a person. I want that person to make a decision, and then I want to see how that person makes that decision. And so on that back end, you have the ability to very quickly cater content to customers. And you can do that with a little bit of effort and there's a lot of great products out there that do that. And even as simple as ChatGPT with prompt engineering, you can do that very, very quickly. So you got that whole -- it really covers the whole gamut of things within retail. Get value from, very quickly.
Scott Light
attendeeAnd certainly, we've gone from one extreme where, for many, many years, we just had no ability to capture data, and we're very limited on that data resources and the ability to use that data for pinpointing promotions, pinpointing exactly where we needed our product to replace to where now, there's almost too much data. There's so much data and now it's saying, okay, what do we do with this wealth of information, and how do we really make some practical decisions based on this enormous wealth of data that we have available today.
Dale Cade
attendeeAnd that probably is not going away any time soon.
Scott Light
attendeeNo, it's not. So we talked about starting an AI initiative. Some people are further along the curve than others. What do you see -- and Dale, I'll start with you. And when we're looking at starting out with an AI initiative, what kind of challenges do you see some of these retailers facing?
Dale Cade
attendeeSo the first challenge, and I think the greatest has nothing to do with technology. It's extremely human. There's a lot of fear around artificial intelligence. There's a lot of anxiety because there's a lot that's unknown. There's not a lot of legislation. There's not a lot of governance within the industry itself. And so -- and there's a lot of hype, a lot of fear. So when a retailer starts, they have to really -- and they have to address that directly. You have to engage the organization. You have to let people ask those questions. So you have to focus on transparency. People need to understand the process of what's sort of the process. The first question is, why are we doing AI and answer people's fears and their concerns. They're real legitimate. We were talking to a company yesterday, and they came and said, their legal department has forbid any artificial intelligence solutions in their corporation because of those fears. Now that's extreme, but that's there. So you got to address that first. Why do you want to do AI, address the human limitations, questions, concerns, and you can do that. Companies have done it. And then finding the right solution. That's difficult right now because there's a lot of solutions out there. So you've got to know the problem you're trying to solve and then really dig into that solution and know what they provide. And that depth can be, hey, I just interfaced with ChatGPT all the way to we have our own LLM that we build custom for you and then that vendor gives it to you and you go manage it. So those are the challenges I see people having to get ahead of.
Scott Light
attendeeAbsolutely. Randy, how about yourself?
Randy Salley
attendeeWell, I think back to starting the initiative, it all begins do you have the data to properly feed and train your models, and making sure you've got that historical data that it is in, not necessarily a format, but it's clean data, you know what you're dealing with and you've got that on hand and are able to deal with it. And when you're looking at that data, I think one of the other things is to understand the fidelity of the problem you're trying to solve, how precise do you need to be in your forecasting, for example. I know you and I had a conversation a few days back about a particular client that you've worked with that wanted a by day, by SKU, by store forecast.
Scott Light
attendeeSure. I remember that.
Randy Salley
attendeeAnd in our circumstance at Walmart, I mean, particularly in the general merchandise side, I mean, we had 100,000-plus SKUs and a large percentage of those SKUs in general merchandise sold less than 1 unit a week. So if that's my rate of sale, I mean, how precise do I need to be at a daily forecast? So it really wasn't necessary for us with a large proportion of our SKUs.
Scott Light
attendeeSure, where is the value in a daily forecast with those volumes?
Randy Salley
attendeeYes. I mean just being practical about that. And the other kind of tidbit about that, with the replenishment systems we ran at Walmart, roughly 70% of our replenishment decisions [indiscernible] send a case of product to the store was driven by presentation quantity versus need and rate of sale. So really understanding that particular fidelity, and then Dale mentioned also being clear on the problem, being specific about the problem that you have to solve and making sure that you are employing the right tool and capability and model for that particular problem. And part of what I have seen is just way too many people that have a solution that they're looking for a problem to attach it to and that's generally not going to be a recipe for success at the end. But being very precise about here's the problem, this is the outcome we're expecting, what's the proper tool to get after that.
Dale Cade
attendeeAnd I would absolutely agree with you that from what we found having a clear defined road map right from the start to say, what are the objectives of an AI initiative? What are the challenges that we are specifically trying to address because you don't have to take on the world all at once. But having those clear-cut defined objectives will really make or break the success of an initial AI application. And one thing I'll add to that as well, we talked about data for a moment. Having complete data is not always a requirement to get started with the AI applications. One of the -- as much importance in getting a successful AI application deployed is being able to take data that, in the retail world, is not complete. It's not always organized or it's, on the other extreme, just a big blob, and be able to have the expertise to organize that in a way that these different AI applications, neural networks, expert systems, things of that nature can process that, train off of it, and it doesn't necessarily need to be complete data. So you can start with what you have and then build from there.
Randy Salley
attendeeYes. And to Dale's point about the change management portion of it, coupling that with your comment on road map, you can start small with a particular category and kind of limit the blast radius, if you will, through that initial implementation and bring all the -- manage that change management for the merchandisers or the planners or whoever you have to interact with and then build on the success of that first implementation because managing that change management is a whole lot easier when you've got people in the organization say, man, this thing is great and working fantastic.
Dale Cade
attendeeAbsolutely. And user adoption change management, accepting that there is a major cultural shift when you're bringing these new technologies, the science to a group that in many ways have been doing this for decades. And we've all heard this blending of art and science, but it is a real challenge. So you can get the technology right, you can get the improvements from a numbers and a mathematical standpoint, but getting the culture to adopt it is just as important as any of the technology implementation piece.
Scott Light
attendeeThe people in your organization will very quickly realize if you're dehumanizing them to implement artificial intelligence.
Dale Cade
attendeeAnd the goal is to improve their work output and improve on the skills that they are currently bringing to the table.
Scott Light
attendeeVery good.
Randy Salley
attendeeAnd margins and sales.
Scott Light
attendeeSo there's so many different stages of where retailers stand today. Some are just working in the most rudimentary Excel spreadsheet based planning systems. Some are working with some of the most cutting-edge planning applications that we have -- that we see here at the show today. Make some comments, if you would, about retailers that already have a planning application in place, when they're looking to add AI to it, is it start from the ground up? Is it -- where do they go?
Randy Salley
attendeeYes. Well, Scott, to your point, I think everybody has some well-established planning systems. And it could be the end all be all, soup to nuts, all singing, all dancing platform for one of the major suppliers of that type of software. Or it could be a giant spreadsheet that they've been managing their business on for decades. But everybody has some kind of system. And it doesn't have to be a rip and replace in order to move into the AI and begin to take advantage of AI to enhance whatever it is you're doing. Just an example from my past, we had a price management solution, store unique markdown solution, but we really felt we were giving away markdown dollars at some times because we tended to do -- everybody gets the same markdown on the same day. And so we engaged with Churchill back in the day and brought them in with their neural nets, and we were able to take our historical data, sales data, and combine that with all of the price change activity we knew we had made, feed that into the neural net and understand what that price elasticity is. And we used that to help us on 2 different fronts. One was indices and markdowns, it's time to get out of the particular category for whatever that season is, but not everybody needs the same markdown on the same day because I am sitting with 150 on hand, you're sitting with 10, I need a big markdown, you probably don't need one at all, and how do you manage that? And the only way you can get there is knowing what that -- how that price change is going to affect that particular demand. On the flip side of that for just our regular price changes, side counter items that and I'm going to take a price reduction on is, is that price reduction going to generate additional demand or not? And when we did see big signals of demand signal change, we were able to take that and feed it back into the replenishment systems so that we could replenish additional product in advance of the price chain so it's there when the price happens -- the price change happens and we've got inventory in stock. And -- for us, it didn't require that we radically change our existing systems. We just set the model to the side, fed it data, got it properly trained, producing the right kind of output and then fed that back into our existing systems to get the kind of outcomes we were looking for around markdown reduction and inventory management.
Dale Cade
attendeeAnd to follow on to that, I think the best markdown that you can take is one that doesn't have to be taken in the first place by looking upstream from your season and getting that initial allocation right so that you have the right product in the right stores, especially for some of these shorter seasonal items that you may not have time to make a secondary or a third replenishment push and so getting that initial set right is just as valuable as what markdown you'll need to take to reduce the markdowns to begin with before you even get to that point.
Randy Salley
attendeeAnd the other additional insights we were able to gain from that was not just the impact on that particular SKU, but the impact on that particular category through...
Dale Cade
attendeeThrough the cannibalization.
Randy Salley
attendeeCannibalization portion of it. So taking this particular price change, drive volume to that particular item, but I'm going to cannibalize other items and being able to see that across the total for the category or fine line or however you define your merchandise hierarchy was super important to us.
Scott Light
attendeeDale, how about yourself?
Dale Cade
attendeeI'll harp on this because I've learned it the hard way is, when you're going to make this decision, you have to do some internal assessment. If you're going to do -- or you want to do a rip and replace, there's no reason to check up on that, have the AI assessment part of the process. But at the same time, if you don't want to do that, there are solutions, good solutions. We've seen a couple so far that really do a good job of just layering on what you have, if that's what -- if that is what you need. I just encourage folks, AI as a technology, it's not cognitive. It doesn't think for itself. It will only do what you enable it to do by decisions that you make, and those decisions are very broad. But I was thinking of an example in the previous role as well, Randy, where we had a wholesale part of our business, and the account managers for the wholesale department pulling in the sales numbers, [indiscernible] 165 different portals to pull data, consolidate that into Excel, massage it, then pass it back in upstream. And that took 3 days. It would go out Wednesday evening, a gentleman down there nodding his head. So they have basically one day to analyze the data, and then they take Friday and get that back to the wholesalers to say, based on 2 weeks -- based on a week old data, we think you should do the following. So we introduced a solution to where using AI technology, all that information was pulled, assembled, processed, pressed down Sunday night. The data analyst team would look at it Monday, double check and then they head it to the sales department no later than noon on Monday. They loved it. Ran great for about 2.5 months. We had a little hiccup. It's going to take us a couple of days to figure it out, and we were going to miss one of those loops, those cycles. And the VP called my desk, said, you need to turn that back on. I restructured my whole department, my whole office, we've reset expectations for all of our partners, that needs to run on Sunday night. So what do you do in technology. You talk to your team, order pizza, work 2 straight days to get it running and hit the date. But when we did that technology, it was 3 or 4 people working. No one else had to do anything differently. And that's the power of the technologies that wrap around. Yes, the rip and replace, sometimes that's what you need to do and you just -- you gird up your loins and go through the rip and replace. So the nice thing about it, with that -- you don't have to pick. You can use AI in both those solutions very, very easily.
Scott Light
attendeeSo just repeating, some people are looking to completely putting a brand-new planning system that may or may not already include AI, in which you can add the AI functionality too, whether we're talking about merchandise planning to build assortments, to get your initial sets, whether we're talking about those inventory management and replenishment systems so that you're getting the right replenishment to the right stores, whether you're talking about these markdown applications, these pricing and promotion applications so they can either be included in a new planning system, or they can be added to something you have existing. And as many of us know, over the years, some of these existing planning applications get tweaked and customized and band-aided and get to a point where they look nothing like they did when they were first implemented.
Randy Salley
attendeeYes.
Scott Light
attendeeVery good. All right. So one more for us while we have time here. If you had to prioritize one major quality when you're looking at a vendor to start an initial AI initiative, what would you look for if you had to make one choice?
Dale Cade
attendeeSo I would say you would want to understand the depth of that AI solution. And I'll just -- I'm just making this up. Let's say it's from 0 to 5 because there's some vapourware out there. You can quickly get through that. But there are some that just do -- their solution is, hey, we call the ChatGPT APIs. It's not bad. If that's what you need, that's a shallow solution. Buy that solution, exploit the heck out of it, make sure you get a good deal with the OpenAI, so it's -- you want as cheaply as possible, all the way to solutions that they'll come to your organization, build an LLM that's specific to you and turn it over to you to manage. That's a very deep solution. You have to have the staff for that, probably data science team. But that's what you want to look. Because if you just need shallow AI, don't buy the personalized LLM, most solutions don't need that. But understand that depth and get the depth that's right for you.
Scott Light
attendeeI think that goes back to our earlier comment about having a clear-cut objective in mind of what challenges you're trying to address, so you can figure out what AI solutions works best for your company?
Dale Cade
attendeeThe deeper you go, it's going to cost you more. AI solutions aren't -- you have to invest time. You have to invest staff, capital, OpEx, hardware, cloud vendors. So those have costs, you need to weigh them.
Scott Light
attendeeVery good. Randy, how about yourself?
Randy Salley
attendeeWhen I think about it, Scott, like anybody can have access to AI models. I mean APIs into them. They're readily available. Anybody can go implement them. But if I'm thinking about, as your question ask, a partner that's going to help me along that journey, while the technical capabilities and their understanding of models is critically important, I think the most important thing for me would be, do they understand retail business? And what kind of experience do they bring to the table that would allow me to be -- have the greatest opportunity for success? An example I would give is I had an opportunity to hear a pretty major retailer in a presentation several months back that had used AI for their bakery to help replenish ingredients into the bakery to go make the doughnuts or croissants or cakes or whatever. And as he told the story, they ended up totally missing the Easter holiday because they didn't replenish enough ingredients and make the stuff they needed to make. And I don't know all the back story on that, but I'm pretty sure that the guy leading that from a technology standpoint, really, really understood AI, but he really didn't understand the dynamics of the retail environment because missing a major holiday is sort of a big deal. And so being able to have somebody by your side that truly understands retail, that has lived through those examples, that know the pitfalls, that have stepped in the mudholes would be, to me, more important than just their technical expertise.
Scott Light
attendeeIt's a great plan. Thank you, gentlemen. Well, listen, that's about all the time that we have. I just want to take a moment to thank both Randy Salley and Dale Cade for taking the time to speak with all of us as well as NRF for putting on a wonderful show. It's great this year as it is every year. If you have any follow-up questions, or you want to have some further discussion with Randy or Dale, we'll be at booth 1119 at the Churchill booth down on the first level. And as always, thank you to Columbus Consulting as well for their continued support Thank you, everyone. Have a great show.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Walmart Inc. transcript — plus 250,000+ transcripts from 12,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.
Get the API View API docs →This call discussed
For developers and AI pipelines
Programmatic access to Walmart Inc. earnings transcripts and 250,000+ others is available through the
EarningsCalls.dev REST API. Plans from $24.99/month — full transcripts, speaker segments,
full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.