Getting to the ‘So What’: Unilever’s Dan Cook on Democratizing Data for Actionable Insights

Data democratization is a term that gets thrown around a lot in consumer goods, but translating it into an effective strategy that best utilizes people’s time can elude even the most experienced teams. Dan Cook, Unilever Head of Walmart Food Sales, joined a conversation to share how the use of data, analytics and technology has evolved dramatically during his career, the KPIs that are really resonating right now, and what's it like to measure digital performance in the most rapidly shifting landscape in modern history. Lisa Johnston: Hello everyone. Welcome to “New Day, New Rules, Democratizing Data to Unlock CPG Growth.” My name is Lisa Johnston, and I’m a senior editor at CGT. I'm excited to welcome our guests today and dive into today's topic. We have two industry experts here today to talk about the new and different ways consumer goods brands are leveraging data and analytics as well as the role that artificial intelligence and other technology can play in gaining a competitive edge. With me today are Dan Cook and Matt Holland. Dan Cook is Walmart Foods Team Lead, Savory & Tea at Unilever. He's a veteran CPG and retail problem solver and has held various general management, sales and marketing leadership roles throughout his career. Dan's well known by his peers and colleagues as a very passionate and energetic business development leader, who is truly a marketer at heart, given his love for the consumer. He strives to create products and experiences in a way that always puts the customer first while simultaneously driving sales and business objectives. Matt Holland is head of the CPG sector at Hypersonix, where he helps companies transform the organizations to better compete with data and analytics. A highly experienced leader in technology, business strategy and analytics, Matt accelerates digital transformations, evangelizing and building new analytic capabilities and organizations that drive better decision-making. Dan and Matt, welcome. Very happy to have you here. I'm going to pass things off to you now. I know you two are ready to have a very candid conversation. Matt Holland: Thanks Lisa. Dan, I'm thrilled that you've joined us and wanted to have a discussion about your career arc, how things have worked for you in terms of data and analytics. I think talking to someone like you who has had so many interesting experiences — you were a Marine, a retailer (Walmart, Target, SuperValu), a vendor at Nielsen, working with me, a CPG manufacturer (Unilever, Mandolin's, and others). Thinking about this whole space and how you did business when you started and you were in stores, doing some of the blocking and tackling, the execution day-to-day, and now where you are today. How have things really changed in terms of data and analytics? There has to have been such a difference from then to now. I would love to understand, from your perspective, how things have changed and how things are getting better. Dan Cook: Really nice to be here with you today. There's a lot to unpack in that question, Matt. Let me try and make a start of it. For me, when I think about how analytics have changed over time, I think about how my roles have changed over time and how in the beginning of my career — I began my career in retail 20 years ago — I recall much of the focus was to understand what happened in the business. Whether it be over the weekend, last week to last year, we spent a lot of time trying to diagnose those trends. Obviously, it would be critical for me to be able to tell the story around performance and share that story with senior leadership. In order to do that, I pulled my reporting manually, and aggregated sales trends across a few limited KPIs and measures to determine how to optimize for execution. That's because there wasn’t anywhere near the level of technology that we're going to touch on in this conversation today. Then, I’d try to do some post-event launch measurement to better understand performance for corrections of errors for next time. I'd say aggregating and pulling all that data, again, it was just time and sense of a manual. You fast forward to today and it's still about that, execution is critical. It's still very much about that, but it's so much more. It's about what we can understand about the shopper, the consumer, and who's buying the product. It's about understanding their behavior and ultimately using that insight to get out of what we're doing, or did then, which was really that post-event launch analysis and get to that prescriptive insight of what we should do for the future. Holland: You mentioned a little bit about the speed. There's an expectation for you to go faster, do more, and offer a little bit more insight in terms of diagnosing what happened. Now what are you going to do about it? How have the data and analytics capabilities that you've used evolved to help answer those questions and where do you have gaps today? Cook: We all know the proliferation of technology. It's available for us to use in our day jobs now — AI, machine learning — it helps us solve things and drives speed-to-market, to drive those actionable insights that we're looking for in a much faster way. A lot of times these machine learning algorithms can be more accurate and more precise in addition to the speed-to-market, they can bring those insights. That's where the focus is, and the biggest change to what my role would have been 20 years ago, but I would even say a couple of years ago. Holland: You mentioned that there's so much more. We've been through a year of disruption with the pandemic. I assume disruption and even your own business, how have the expectations of leadership or sales leadership changed over the course of this time in terms of how they expect you to use data and make decisions. Cook: We know, again, that the world has changed, the KPIs that we used to look at as sales folks way back in the day as I'll call it, were pretty rudimentary. They were sales lifts and what did my business do? Did I do better than last year? A lot of times, if it did, if it hit a hurdle — 10%, 15% — the last year, we would just repeat it. Not knowing anything around the efficiency of a promotion, not understanding what really drove it, not really having that insight. I'd say today it's completely different. We do care about those things, but care about so much more now. I would say our leaders expect us to know those things. Our leaders really expect us to have up-skilled and use those better measures to understand the efficiency, what the shopper purchased, and what the buy rate was. We've stepped up our analytics to deepen our understanding of the value of that shopper: Did we bring in a new shopper with a promotion? Did we just drive frequency with our existing? These are things that we're expected to know, honestly, on a daily basis. We need that information not only to drive what we should do to correct that previous promotion, but what we should do next. It's about how we take those actionable insights to answer not only the what, but the so what, and what we should do to win. Holland: That makes a lot of sense. The deeper you can dive into what happened and diagnose why something didn't perform the way that you expected. You should be able to then understand what you need to do differently next time. It sounds exactly like what you're doing. Let’s talk for a second about speed and insights. It used to take a whole week to dig into dashboards and reports and understand what happened last week. Sometimes by the end of the week you still haven't quite figured out what happened last week before the new numbers come up. So, in terms of technology, we see a lot of improvement in terms of applying AI and ML, and being more data-focused, but it always seems there's another question to answer. It never seems to be enough, you’re always chasing additional insight. How does this manifest in your work? Cook: That’s a great question. Overall, the objective with the proliferation of technology and big data is about answering these business questions that you were just talking about. Whether it's trade optimization, why market share is down, or what trends are emerging in the market, and what the consumer or the customer is trending towards. It's about how we utilize this thing that is finite, that we all only have so much out of, in a better way. We are all busy — I feel busier now more than ever — and being busy doesn't equate to growth, or to delivering goals. You have to be smarter and work more efficiently with your time. At a baseline, you want to use the technology to give you more time so that you can do the things that are going to move the needle on your business, things that technology or AI and machine learning can't do. You can use machine learning and AI to get 90% of the way there, then from an art standpoint, you come in and do the last 10% — the most meaningful, valuable work. I'll call it, low value to impact versus high value to impact. I would say that the little value to impact work, we want to automate that. We want to move that out of our day-to-day jobs, spending four to eight hours a day on that stuff. We're getting that back from the technology, using it to act and take action on what the data is telling us, and gain insight from it. Holland: That's so key. You talked earlier about up-skilling a little bit. We had spent a lot of time working together and training clients on how to use data and analytics, how to use tools. You spend a lot of time usually with standard training, which is keystroke training, which is explaining how to pull data and visualize that data. With the advent of some of these new technologies around natural language processing where you can ask questions, the challenge today is how to train and up-skill around interpretation. Then, how to get people to understand what they're looking at and how to use those KPIs and numbers in a way that makes sense for the business and is actionable. One of the things I learned from my consulting past is that everyone is trying to make a quarter, or some sort of number, and there's a timeline to how well and quickly you can execute a lot of different things. So, focusing on what you can actually do and how you can move the needle is the most critical, then having the time to be able to dig in. To do that by moving some of these more routine tasks off your plate, has been something we've been trying to do for a long, long time. This is not new. But at the same time, the technologies finally exist to offload some of that stuff, even for folks who don't ever want to go in and pull their own data. The speed part is key, but the technology is really the enabler. I’ll go a little bit further on that then, who does all of this work these days? You're in a sales role, there’s insights and analytics functions, there's IT functions. We have this notion of data-driven decision-making, putting more data and tools into the hands of every user. I talked about ease of use and speed, a big notion these days is democratization of data. How do you get more information into everyone's hands, so that everyone can make decisions better? Do you agree with this idea of democratization? Or is this something that is a little scary? Cook: I think there's a nuance to this answer. Overall, I'm going to say yes, but it's going to have some caveats. The reason why is, you've been speaking to the concern that you and many leaders likely have, depending where you are on the spectrum. When you think about the democratization of data, it's because you want to make sure that those who have the data understand how to use it, how to interpret it, and that they have the skillset to do it. In terms of my view, everyone should be an analyst and have access to the data. They just need to understand how to use it. In the end, being able to share that data broadly. Meaning everyone has the skillset and can properly use the data, leading to better outcomes for the consumer and for the customer. Basically all of the teams have not just a piece of the puzzle, but the full picture of what's happening. They're understanding it, watching it, monitoring it, and they’re optimizing, testing and learning against it to continue to drive better results. Everyone in a function cannot be everywhere. We need to decentralize responsibility to be able to go faster and this helps us do just that. Now, we're seeing this with a number of topics in the industry, e-commerce is one Every CPG today, for example, is focused on the future and trying to future-proof their growth, trying to win at e-commerce. What does that take? That takes everyone to understand e-commerce. It can't just be a couple of folks in the company driving it, it’s got to be everyone in the company understanding the importance of e-commerce, how to optimize for e-com, and what that looks like. We're on a journey that everyone needs to go on together, and depending on your role and function, it may vary in the level of expertise that you need. Everyone should be comfortable using data, interpreting it for their role — it can be role specific, but having the right skillset to understand the business through the trillions of lines of data that we have today, is going to be very important, but even more important into the future. Holland: And it's accelerating, right? The pandemic has put e-commerce on a pedestal. It's very different than it was even at the start of the pandemic. It went from being this small thing that's growing on the side that you look at once a month, to something that you're looking at every single day. Now, we’re making sure that you're not out-of-stock or something's not happening with the buying process through your own, third party, or direct-to-consumer site. You talked about skilling and upskilling, this idea of training and the democratization of data and analytics. What do you think about motivating people to do that? How do you evangelize that data use differently than what we've been doing for the past 20 years in CPG? What's that role of upskilling to develop that next generation, business person? Cook: Like anything else, it's a business transformation. In general, sure there can be some apprehension, or some questions that employees have around how to be better or what this will look like. But that’s what companies need to get good at, explaining the multi-year journey. We've got to embrace it, this is going to be expected of you, but we're going to build that path together. That journey that you need to go on, you're not going to go on it alone, we’re going to go through it as an organization. That’s what's really important, explaining that it isn't about robots taking anyone’s job — and don't get me wrong, we know that technology is certainly replacing some roles, but in broad terms. It will be about how we repurpose or re-imagine someone's role at a higher strategic work value that a machine can't do. It doesn't mean you won't have a job, it just means you're going to be doing work that's actually more valuable to you. The work that you're going to be more passionate about in your role. It should be an enabler. It should help you be a better leader, better employee, to be able to deliver what you need to every day. Holland: Let’s talk about the process of what used to be a project in analytics. You'd spend a lot of time upfront, let's say it's a six-week project, you'd spend four of those weeks pulling, collecting, and collating data. Then you'd spend two weeks interpreting it. That's going to flip because the data comes faster and then the interpretation of what you should do is really where the action is. You've been in a lot of different roles over the course of your career, what has been your key to success in terms of using data and analytics to achieve your goals? Whether it's performance share or whatever it might be. How would you approach this and what would you say to our audience in terms of their own journey? Cook: Obviously, it depends on your business, your company, and your goals, but the KPIs that resonate for me, time and time again, are sales growth and market share. Market share is something, kind of consumer relevance, I would say market share, against competitors. So, in essence, brand against other competitors. Retail market share, of course you need to look at that and just overall category share. Again, going back to the competition piece against brands — definitely look at that. But then, there are these other metrics that have become more important, like consumer lifetime value. Did we bring in a net new shopper, what was their overall spend, or what are they worth to us? These different metrics really upscale the analytics that dimensionalize value to the enterprise. You're going to see sales folks of the future, marketing, pretty much every function, follow these deeper analytic measures that we traditionally haven't used, but that are the underpin of true value to an enterprise. You’re going to see more and more of that. Holland: That's great to use that example. One of the reasons we're looking at things more like customer lifetime value or consumer lifetime value, is that we have the data to do it. It used to be that the customer or consumer was this big, amorphous blob that we were tracking in a panel — you could break it down a little bit, but you'd run out of samples pretty quickly. Now we have all kinds of loyalty card data down at the transaction level. We have shopper data for e-commerce that runs across different sites and looks at attribution and journey in terms of the path to purchase. So you know who these consumers actually are, and you have the ability through retail media network buys — all of these different things — to actually go out and target them once you understand their value. It’s so much more actionable than it was because you have the granularity, the tracking, and the measurement to be able to put all of that in place and follow up. Cook: E-com is an excellent example of that, especially during the pandemic last year and even still today. Everyone saw their digital penetration increase, and not only that, they saw an amazing amount of sales shift from brick-and-mortar to online. As the head of the e-commerce business last year, I went from sending monthly recaps to my senior leadership, to really feeding the business daily sales information to measure performance. We were doing 20x in sales compared to pre-pandemic through channels like Instacart, and other customers shipped that I managed. It just underscores that point, the pandemic exposed the need for us to do this — we need it now. It's not something that I even want to do in the future, but the future is now. The acceleration that we're talking about is really something we're living now. The need for having those measures, having that information, like I'm talking about CLV and other measures within that environment. It just became that much more imperative. That’s where you're going to see CPGs double down. Holland: I would imagine it’s also pretty complex. I was going to take the next few minutes to shift gears a little bit and talk through what I refer to as the three AI imperatives of CPG. One of our partners at Hypersonix is Google. We leverage Google Cloud for a number of different applications and as one of our clients, our partners work with them, as well as AWS. They've recently did a really nice deep dive on applying AI in CPG and CPG retail, and have been able to find out that essentially if they understand some of those different applications and think about them as almost use cases for applying AI, that they've been able to identify $490 billion or almost half a trillion dollars in incremental value that CPGs could generate, based on leveraging AI and ML technologies in their business. That's really the prize. Then the question is, how do we get there? How do we capture our share of that growth and opportunity? I wanted to go through three things quickly here that I call the three imperatives. One of them is about data: The data imperative. As we think about all of the data that we have access to, Dan talked about increasing amounts of data and leveraging technology to wrangle data and gain insights from it. We see this as a best practice and as the future is data unification. There are a lot of companies that have built data lakes or data warehouses, but their data is still siloed in the lake — it’s not really a lake, but more like a bunch of little puddles. That doesn't really enable the kind of complete view of the business that you need to measure and monitor your enterprise today. As Dan mentioned, there's rapid growth in these newer e-commerce channels. The complete view is what will unlock growth. We've studied, and will continue to study, our core channels and core businesses very deeply. At the edge of the frontier are where the insights come from. You may see smaller brands pop up that have this all other bar growing and growing within your product share information. You may see when you model data, a breakdown of others or even errors, things that you can't predict or understand. With the model today, this is insight. Aligning how you approach some of these things is the key, but that unified data is the basis for truth. The second thing is that we have to detect the disruption before it happens. There's been so much disruption over the past year in supply chains, but if you survey supply chain or operations executives, they will tell you that while they put some new tools in place to help monitor the disruption and rebalance inventory, they did different things during the pandemic. Only half of them kept those processes in place, which means the other half is going to be flat-footed when the next disruption happens. To do this, it can no longer be something that humans can monitor — this requires automation, quick insights and quick response, and continuous monitoring in order to be able to capture and understand that disruption and do something about it before it has an overly negative impact on your business. Machine learning is the place to do that — to understand patterns and recognize trends and disruptions in data signals. That's what we need to put in place to sustain our effort, so we can monitor, detect and avoid the next disruption as much as possible. The next piece is about new data. So thinking about e-commerce, we're talking about customer lifetime value. These are new metrics, in some cases they can be new model inputs or drivers of sales, but there's a lot of different solutions out there. You need to take control of that data in your own enterprise because it's going to create a lot more complexity that you need to manage. Lastly, in terms of what your “North Star'' is, think about what's actionable, what you need to track. There's going to be a lot more here to invest in than you likely have budget for, and it's going to be a bit of a journey in and of itself. Focus on what you can actually impact would be our recommendation for that. The next one we talk about is the people imperative: How will future analyst's work change? It will fundamentally change. Think about low-code and no-code tools, natural language processing — there are platforms, I know of one at Hypersonix, where you can talk to the tool, ask a question, and it provides responses. That sort of interaction can even allow you to then build visualizations or simple models based on a low-code environment where you don't need to know Python, R, or any language in order to be able to build. These are going to fundamentally change how we work with data. It's going to advance our reach into areas that have typically only been specialized vendor areas, where we've worked with a marketing mix supplier or an RGM supplier because building new visuals on-demand is the future. We've gone through an era of dashboards and reports. A lot of companies are realizing that they can't build enough of them to keep up with all of the different questions that people have. The fact is, people don't think in terms of dashboards, they think in terms of questions. We need to best provide a solution that allows those visuals to be built on-demand when needed, then we can take some of our data scientists and refocus them — their expertise on innovation, on new things that we want to do with the data, on new things that we want to learn. The data science team is much happier with this and better enables the overall organization with these tools. This brings me to the next point, which is shifting; there's been a lot of specialization around commercial analytics over the past several years. If you're going to support brands in this day and age, you need to be more of a generalist, have control centers or some visual war room built into your organization that demonstrates where you can go in and talk about plans for the year. We're going to immerse ourselves in the data, understand what's happening with the business, and we're going to build something new and different next year to help achieve our goals. That requires cross domain expertise with people who can talk about pricing, marketing mix, and consumer segmentation. It also implies a decentralized insight and analytics team - individuals that are embedded in the business. There will always be a central function for analytics, but we've built these ecosystems and large functions for insights and analytics today. A lot of times, people who are working on projects in those departments are very far removed from the business questions and decisions that need to be made by the brands. That can lead to rework and inefficiency, a lack of understanding and actionable insight. Then, finally, the reality is when you answer a question, there's always another question. So, how do we do more with the same budget? As we get more complex in terms of fragmenting sales channels, different metrics, and KPIs to run our business, that complexity is going to drive more questions, so we need to be prepared. Finally, how do we put systems and tools in place that allow us to not just look at our core, but also cover underserved businesses. Customers who are not in your top 10, but still contribute millions of dollars to your bottom line, or brands that don't have the majority of the investment in terms of advertising but are some of your most profitable. How do you provide more impact to those businesses, which are underserved, but can often get more impact by investing in. I want to talk about the Tool Imperative, it's really a must. As I've been working in and out of the industry, selling to and working with different clients, there are various degrees of understanding and acceptance of AI. Some don't believe it and don't want to talk about it, others are very much involved in AI development and are champions and evangelists in their own organization. That spectrum exists not just within the world of CPG, but also more broadly. People talk about facial and image recognition technology, for example, and all of the different societal impacts of that. Facebook has already made the decision for example, that facial recognition technology is important and they're one of the biggest and the best at it — this has already happened, the decisions are already made. Why don't we apply these technologies to our business problems, and leverage the best of that technology to do more and go faster in the things that we need to do for our business? The rise of the machines isn’t something to fear, it's something to learn to trust. We have to let go — and believe me, I was one of the skeptics, one of the last to adopt — you need to deploy machine learning to surface the insights of tomorrow because there's just too much data to get through. You cannot deep dive fast enough, you cannot boil the ocean because it's too big. You need to allow the bot, the algorithm, to dig down in the information and let you know what's really important. Therefore, you can also answer a lot more questions because you're going to be taking some of these more routine Q&A's off of your plate. We can get rid of that expert as a bottleneck, that's been a problem since we've built some of these ecosystems over the past five and 10 years. The middle one is about barriers to adoption. No matter how fast the technology has become, no matter how easy it is to use, we still have a lot of business people who just flatly refuse to go in and pull their own data, look at the reports and dashboards themselves, and prefer things are fed to them. This is where we at Hypersonix firmly believe that the barrier to adoption is the user experience. How do you create a much more natural, conversational experience with some of these solutions, so you can ask questions, get answers, and do so in a more interactive way than any query tool can be. That's the answer, natural language processing. The answer is not more reports, it's not trying to invent the next dashboard that will answer every question. It's also creating a more complete tool. We've gone through an era of point solutions, and with those point solutions has come a lot more things to manage, more tools to learn. We don't need more tools to learn, we need more complete tools that contain more of our data that we can spend more of our time and really become experts at. The last one is something we've talked about during the course of this webinar a couple of times: Democratizing access. This is a slightly different take on it though. We talk a lot about leadership, and leadership being the multiplier. If you're a leader of a team of 10 people, no matter how hard you work, how many hours and what you do, you're never going to be able to equate to, or top the amount of output that those 10 people — all moving in the same direction with a common need and a common motivation — can accomplish together. This same notion exists with data. Democratizing the access, getting more data into the hands of users, with simple tools to be able to access it. This is the impact multiplier of the future. This will, because it's made easy, allow you to do more with your data and make better, faster decisions across the entire enterprise. It'll help in the process, work through and up-skill some of your sales and marketing people, because as Dan mentioned a few minutes ago, in the future, everyone is going to be an analyst. Johnston: Thank you, Matt, for that great information. Thank you both for the conversation. It had a lot of great takeaways, interesting perspectives for our audience. Dan, to your point, the urgency to re-imagine roles that are more valuable and that machines can't do, that's definitely, certainly a driving conversation right now among a lot of brands. Dan, how do you connect with your retail partners today to make sure how to best forecast demand? Cook: That's a great question. We are working closer and closer overall with our retailers. At Unilever, certainly, it's the case of every CPG today, to embed ourselves and process, and cross organizationally with the teams on their side that are forecasting the business and planning the business. It's become clear to all of us, the better you have that partnership — think about sharing data, sources of data, understanding things like supply, what motivators or drivers of our retail partners overall plan. The more we have that insight, the better we can forecast. Obviously as a CPG, typically what we look to do is bring a ton of thought leadership. We bring category insights and very objectively bring that category story. Retailers have a ton of data, they understand who's converting, especially if you think about — using e-commerce as an example — they have conversion rate, they know where people are browsing. All these data points. Think about putting those all into a data lake together to understand what forecast should be, putting the right inputs into that to make a more powerful, robust forecasting. That's really what it helps us do. That's a long winded way of saying, there's a lot that we need to do, and it needs to be underpinned by really strong technology and cross-organizational sharing in the future because that's what you're going to see happen for the CPGs that win in the space. Johnston: I have another question building off that. What do you think are the biggest opportunity areas for artificial intelligence to enhance brand-to-retailer collaboration? Cook: Biggest areas? I'm going to flip it to Matt to talk about as well, but overall there's this need for us to speak to the market, to be quicker, better, faster. You think about the signals that are happening right now, while we're on this call, someone is buying the brands that I sell, that multiple companies sell off of a shelf right now. Do I know if those brands are out-of-stock? Do I know time of day? Do I know what inventory I need to supplement at a store level to ensure that an outcome is driven so that we keep the shelf full and grow market share. You guys know what our goals are, that's what you're going to see AI solve. It's what it's already beginning to solve today and machine learning is beginning to solve today. You think about media spend, time of day, where should we serve the ads, what websites? All of those are things that machines can do way better than I could do. With our retailers, media and networks, how can we deploy that type of technology to figure that out quicker, better, faster for our brands? A lot of times, that retailer will have data, but we need to know what works best, like Matt was talking about, precision for targeting. We know a lot around what works best for our brand. So how do you start to marry those two together, to drive a more powerful outcome? This is where things are headed. Holland: It has to start with supply chain and inventory. As Dan mentioned, those are already being handled (in most cases) through algorithmic models, determining what those inventory levels should be, whether there should be safety stock and how much. It seems that's the biggest opportunity. Then, I would also agree, around marketing and ad tech, things around placement of digital ads and working with retailers on retail media networks. Those are the areas where you can partner not just to make sure the product is in the store, but also that you're delighting the customer and doing something that's positively moving the needle, not filling a hole, which is important. Taking your relationship with that customer to the next level, by getting them to purchase one more product, make one more trip to the store, buy one more item in your brand. Cook: People are sharing more data than they ever have with CPGs, certainly through e-com, but all of this online sales influence, offline as well, it's an ecosystem. We've heard this before, consumers are okay sharing data, as long as you do something with it. Think about personalization, customization, the need to deliver that better customer experience and satisfy the customer. People are expecting us to do more with the data that we have. That's the question that we have to ask ourselves: are we responsible stewards with all this data that's being shared? Are we really, really, proactively giving the consumer what they want? They're telling us a lot, it's up for us to determine what that is and build with that. Johnston: Matt, what are some of your views around best practices when it comes to data driven decisioning in consumer goods? Holland: In terms of best practice today, for example, data — definitely investing in not just data lakes because that's been going on for a while. It's the evolution of data warehousing. It's important to get all of your data in one place, it's also important to connect it together. People don't quite get past the creation of the lake; it needs to be about the unification of the data in the lake, and that then enables the analytics that we've been talking about. With tools and AI today, the best practice is definitely learning how to interrogate that data that you've just unified. How do you get simple diagnostics in place that allow you to come to an answer more quickly that you can all align around. What happened to this? Well, the weather impacted us. Traffic was down in the store, the promotion didn't perform as well as it did last year. You may have a feed of weather, a feed of sales, and a feed of traffic counts or transactions, all from a particular retailer or a particular region of the country. If those are all in different places in your enterprise, you may not be able to easily analyze them together. Putting something on top of that data lake that allows you to interrogate and dive into that data is why you built the lake in the first place. Then with people, setting a high bar. It used to be that you needed a PhD to learn how to model, but you don't. You can build a lot of models and do a lot of things yourself now, that pass the basic sniff test. They're going to give you decent answers, maybe it's not as precise as it could be, but it's also usually free or based on a tool set you already have access to. How do you do more with what you have, and by doing that also make sure that you put the right tools in people's hands, so that you have more people doing it. That's also how we get to a better understanding and interpretation of all of these analytics in terms of what to do with the business because that's ultimately why we're here, to provide advice from the data about what to do. If we can do that more, we can impact more situations, more brands, more customers and more people. Johnston: Those are great words to leave us with. Dan and Matt, I'd like to thank you for giving us your time and subject matter expertise today. I'd also like to thank Hypersonix for sponsoring today's webinar. Finally, thank you to our attendees for giving us your time today, we hope you found it valuable. Have a great rest of your day.

Related Topics