In this episode, Patrick Kelly, the Senior Vice President of Product and Design at 84.51°, talks about first party data and its insight into customer behavior, utilizing a collaborative cloud, how data influences your local grocery store, and so much more.
In this episode, Patrick Kelly, the Senior Vice President of Product and Design at 84.51°, talks about first party data and its insight into customer behavior, utilizing a collaborative cloud, how data influences your local grocery store, and so much more.
--------
How you approach data will define what’s possible for your organization. Data engineers, data scientists, application developers, and a host of other data professionals who depend on the Snowflake Data Cloud continue to thrive thanks to a decade of technology breakthroughs. But that journey is only the beginning.
Attend Snowflake Summit 2023 in Las Vegas June 26-29 to learn how to access, build, and monetize data, tools, models, and applications in ways that were previously unimaginable. Enable seamless alignment and collaboration across these crucial functions in the Data Cloud to transform nearly every aspect of your organization.
Learn more and register at www.snowflake.com/summit
Steve Hamm: Patrick, it's great to have you on the podcast.
Patrick Kelly: Thanks Steve. Uh, yeah, it's a pleasure to be here.
Steve Hamm: Yeah. So tell us about your company 84 51, beginning with the company's unusual name.
Patrick Kelly: Yeah, it, it certainly is a unique name and, and, uh, I promise to come back to that, but first I wanna set some context on, on who we are, uh, because it's relevant to the name. So we we're a retail data science, insights and media company. Uh, we're a wholly owned subsidiary of the Kroger company. Kroger's the largest national traditional grocer in the United States. Uh, they've been around for over 140 years, and really they're all about uplifting America through food, and it's at a very meaningful scale. Um, Kroger's the second largest seller of food in the us.
Um, and, you know, some fun facts. Number one, seller of deli as well as the largest florist and number one seller of, of sushi in America. So 84 51 helps Kroger [00:01:00] and the full ecosystem across grocery retail, namely consumer packaged good companies or, or brands as well as brokers, agencies, and publishers.
Ultimately all about creating a more personalized, a more relevant and ultimately a inspiring experience for shoppers all along their path to purchase, both in store and online. So to, uh, to get back to our name. 84 51 core to the identity of 84 51 is, is both data science as well as a, a profound understanding and advocacy for shoppers.
And that that understanding is not just a point in time, but more critically it's, it's overtime, um, or what we refer to as a true longitudinal perspective of, of shopping behavior. And so given the significance of that, that perspective for 84 51, um, is actually named after the longitude of the location of both Kroger and 84 51, uh, headquarter location here in Cincinnati.
Steve Hamm: yeah. Very cool. Hey, um, [00:02:00] so Kroger's a pretty well known brand, but a lot of its stores actually have different names on them. So tell us a little bit about that. I mean, what are the other brands that people might be familiar with in their neighborhood?
Patrick Kelly: Yeah. Um, as I mentioned, you know, Kroger is a, a national retailer, and you're right. One of the reasons why the name Kroger might not resonate with everyone in the audience is it is it operates, um, in over 25 different, different, uh, brands or banners. So everything from, you know, Ralph's or Fred Meyers in the West Coast as you move into, you know, king Supers in the greater Colorado area, um, across the Midwest, there's a lot of Krogers in the Kroger banner as well as Rounds and Maria's, which is a little north in, you know, Chicago and that, that greater area.
Um, and all the way over to, you know, Harris Teeter on, on the, the East Coast to, to name a few.
Steve Hamm: Okay. Very cool. Now let's dig in here a little bit. What data analytics services does 84 51 offer, and how [00:03:00] is it different than the other retail data analytics services that are out there from other companies?
Patrick Kelly: Yeah, we, we have a portfolio of offerings, um, really across loyalty marketing, uh, our insights business, as well as our media business that operates under the Kroger Precision Marketing brand. Um, all of our solutions are designed, um, as I mentioned before, about kind of the purpose of and mission of the company to, to help our clients and partners make connections with shoppers in a really relevant, uh, and meaningful way.
And, and our, our differentiation in the market really starts with the foundation of our first party data asset, which represents 60 million households across, across the US That's about one in every two households. Um, and it covers over 2 billion transactions annually. And so this, this data asset, as I mentioned, is really the foundation of, of differentiation, but when it's coupled with our cutting, Cutting edge data science.
Um, that's really what brings it to life through all of these solutions and offerings and, and that science, you know, it spans from descriptive and [00:04:00] diagnostic insights to better understand, uh, you know, customer and, and, and shopper's behavior, um, and diagnose. You know, again, you know why that's the case.
All the way to predictive and prescriptive analytics that helps ensure that we're not only understanding the customer in the, the best possible way, but, but really that, that information is, is actually, is actionable. Um, and we're, we're fortunate and proud that the, you know, the industry is really recognizing a lot of that value.
Uh, this past year our, our insights business was recognized as the number one in both shopper data actionability, as well as, uh, the number one grocery retail e-com data asset. Um, and, and our media solutions, uh, being recognized as number one. Targeting effectiveness and measurement capabilities and as well as return on investment.
All of those things, again, are, are really built on that, that massive first party data asset that's, that's so representative of shopping behavior across, uh, the US and then the, the science that's applied on top of that in order to bring, bring those, [00:05:00] those insights, there's analytics and the science to life, as well as ensure that it's, that's actionable and it's not just, you know, an academic.
Steve Hamm: Yeah, yeah, no, your, your media business is, is that, um, uh, is that ad placement or is it ad campaign diagnostics or, or,
Patrick Kelly: Yeah, it's, it's both. Um, you know, our, our media business is really, it's built for, for driving results. Um, it ensures the best addressability ensuring you can reach the right, the right shopper, again, through that first party data asset that it is actionable and you can directly connect to the shoppers.
And that it's, that it's meaningfully accountable, um, because the measurement is, is verified sales match back to the Kroger, you know, stores and online. Um, so it's everything from, you know, cross address, addressability, actionability and accountability. Um, and it is a closed loop loop system, which is, You know, only of increasing importance as we think about a lot of the, the dynamics and shifts in, in media solutions across the industry.
Third [00:06:00] party cookies going away. You know, we don't rely on those when it comes to our first party data asset. Um, and consumers expect a seamless, seamless commerce experience. Um, and advertisers expected data driven decisions. Um, and, you know, we're, we're excited to be driving a lot of the, the, the standard and, and kind of elevation of those, um, accountabilities and standards in, in the.
Steve Hamm: Yeah. Well, let me make sure I understand how your business works. So Kroger is your parent company and, and obviously you provide a lot of data analytics services to Kroger, but you also sell analytics services to the CPG companies that provide the goods or sell their goods through Kroger. Is that, is that how it works?
Patrick Kelly: Yep, that's right. Yeah. In addition to those, those kind of market offerings, um, we're very embedded into the Kroger, um, you know, operating all the way from kind of merchandising to digital to supply chain and operations. And, um, you know, where, where Kroger looks in turns, 84, 51 to really advance, [00:07:00] uh, its again, its ability to better serve shoppers is around providing sciences that, that really enhance that, that entire ecosystem.
And then, you know, to your point, um, we work across, again, that broader ecosystem, so all of the vendors and clients and partners of, um, Of, of Kroger to ensure that there's, you know, it's, it's, there's a kind of a full flywheel effect. Um, because every single one of those pieces, whether it's better understanding your, your consumers as a, as a brand that sells into Kroger, or it's activating, you know, um, coupons or promotions or, or media placement, all of that, every single touchpoint, um, is a part of, of the shopper experience.
And that, you know, that includes, you know, products being on the shelves. And so, you know, we generate millions of forecasts, uh, on a weekly basis to ensure that we have a, a good understanding of what's on, on the shelf and what needs to be reordered. Because, you know, fundamentally, if, if a product is on, is not on the shelf, all that other, you know, all those other tactics really become, [00:08:00] uh, pretty, pretty significantly minimized.
So it's everything from the foundation all the way down to that, the personalized experience of, of the shopper and all of the touch points that are necessary in order to bring that to life.
Steve Hamm: Yeah. Yeah. I can see how this would be really a valuable relationship for the CPG companies because Kroger, in a sense, is a proxy for their entire business in the us. I mean, it's such a, such a large market share that they basically can understand their US sales, you know, distribution, all that kind of stuff through that one source that has very detailed, uh, sales information and personalization information for them.
Right.
Patrick Kelly: Yeah, no, absolutely. You're spot on. Um, it, it's about getting it right for the Kroger shopper, but also partnering with, um, with that full ecosystem on, on their, their national strategy. Um, the scale, the scale and the robustness of, um, our, our data asset, um, is matched in its granularity. And again, you know, of course we're, we're, we're named after the longitudinal perspective that we have.
That [00:09:00] means that you can get a full view of, of shoppers that, that insight, uh, you simply just can't get anywhere else at the, at the scale and magnitude and representation that 84 51 offers.
Steve Hamm: I want to, um, take a step up, kind of 40,000 foot level.
Patrick Kelly: Yeah,
Steve Hamm: Traditional retailing faces a number of challenges that we're quite familiar with, you know, including that long term one of purchasing, shifting to online, and then shorter term disruptions from Covid. So I think it'd be great if you could please give us your take on the state of retailing in the US today.
Um,
Patrick Kelly: yeah, absolutely. In, in you're spot on the retail landscape. Shoppers needs and, and preferences seem to be moving at a faster velocity and, and are more dynamic than ever. Um, A lot of disruption happened over two, two and a half years ago, um, as the pandemic was, was introduced to, to the us and that was a, you know, a very point in time shift in consumer behavior and having [00:10:00] a finger on the pulse of, um, of your consumers.
And again, not just your, you know, your sales and unit volume is meaningfully important to being able to, to, to understand that and then be able to react to provide, um, the best experience for, um, for shoppers. And, you know, while, while Covid, of course hasn't, hasn't gone, um, you know, gone away, the relevant, you know, macro trends of today are around inflation.
Um, you know, in, in September we, we, we have a consumer research, um, offering as well that helps us really keep a finger on the pulse of some of the kind of subjective, um, in, in the why questions that our, our shoppers are able to answer directly for us. And we found in, in September that. Still, nearly 50% of consumers were feeling uncomfortable with their finances.
And, um, while that's, that, that kind of level is, is is starting to kind of be consistent, um, nearly, nearly, um, 65% of households still report that they're, they're meaningfully looking for [00:11:00] sales and deals and coupons. Um, and, and over half are cutting back on non-essentials. And so those macro level trends are, are incredibly important as to how you think about how you build the right experience, all the way from, you know, again, merchandising to operations to price and promotions.
And then while those macro trends are incredibly important, there's, there's layered trends on top of that. Uh, as we move into, um, you know, this, the holiday season we're, we're seeing that, um, there's a number. About a, a fourth of, uh, our shoppers are actually gonna be cutting back a little bit on their, on their Thanksgiving purchases.
And so we need to meaningfully understand that as well as we move into the winter. You know, 25% of of shoppers are planning to stock up on natural remedies for, for cold and flu. Um, so that's some of just like the subjective. Kind of wrapping around some of these macro trends. And then as we, you know, as we've, we've talked about leveraging our data and our science to then take those and then make them [00:12:00] relevant for, for each, for each, um, consumer.
So, you know, the, the, the price sensitive, uh, shopper that is really deal seeking, ensuring that that's at the top of their, their, um, e-com journey when they're online, or that when they come into stores, they can be confident that across every single category and across every single need state, there's a really competitive offering for them.
Um, while there's others. You know, as we again, move into the excitement of the holiday season, still want to experiment and, um, and, and bring joy through food as families to come together. And so we wanna be able to continue to inspire our shoppers as well, um, re regardless of, of, of who they are. And I mentioned this before, but that goes all the way from supply chain and ensuring products are on shelves to offering the right promotions and creating the, you know, the most relevant and personalized experiences.
Um, so that, you know, shopping for, for groceries is, is inspiring and, and delightful. Um, you know, regardless of where you are and what your top concerns and needs are.
Steve Hamm: Yeah. Yeah. Now, you mentioned before that your [00:13:00] data represents about half the households in the United States, about 60 million households. What types of data do you have and how do you gather them?
Patrick Kelly: Yeah, so you, you mentioned again, the 60 million households representing about half of the United States. That's a, that's, that's the foundation of our first party data asset, and that's an earned asset from shoppers in exchange for value. For over 20 years, consumers have been opting in to the Kroger loyalty program, which has been, which has allowed us to capture purchase data on, uh, 96% of sales that go through Kroger stores and.
And in exchange, consumers earn fuel points, receive discounts, um, are, are provided a person, you know, personalized offers and a personalized experience and an overall enhanced and raw relevant experience. That value exchange is, is key to the relationship that Kroger has with the shoppers and again, to, to that extremely valuable first party data asset.
And, and while that asset represents [00:14:00] thousands of stores and online, millions of shoppers and, and billions of transactions, um, you know, I, I'll, I'll actually, I'll use an example that I think will, will highlight, um, some of how we actually build upon that data. And, um, the, the example I'll use is really through our, our personalization sciences that really drive the curation of the shopper's online experience.
Um, so just to start with the scale of it, right? In total we're mining 10 petabytes of data, and that's a, that's a magnitude that's really hard to comprehend. Um, to put it into, to different terms, it would take two and a half years of nonstop binge watching of 4K movies to get to one petabyte of data.
You could take 4,000 digital photos on your smartphone every day for the rest of your life and still be short of one petabyte of data. So that scale is absolutely critical for our data scientists to apply the machine learning methods on the data signals that inspire discovery and really drive conversion.
And [00:15:00] our personalization is, is rooted in the rich observed behavior, right? So we talk about that, that really massive and granular data asset, well, we build things on top of that, you know, basket composition, purchase modality, uh, product metadata and so on. So, From, from taking that fundamental asset, we're actually able to create descriptive analytics and derive needs on top of it.
Things like key customer segmentations, such as, um, price sensitivity, your affinity for convenience, uh, the importance of quality, uh, your affinity for innovation. You know how likely you are to, to try new flavors and build or, or buy new products. And this is all, this all helps, you know, highly. This is all very highly predictive.
And so when we, we actually start to think about how we activate on that. We've got hundred dimensional customer embeddings, hundred dimensional product vectors that creates 60 trillion combinations of customer and product per week. And then finally, when we apply the predictive and prescriptive machine learning on that, [00:16:00] you know, things like deep learning, BA and techniques and other machine learning models that produces the highly valuable and highly relevant personalization that, that creates the inspiring shopper experience.
So there's, there's massive amounts of data and science, um, and enrichment that go on behind the curtain so that when you, when you go on to kroger.com or you, or you pull up the Kroger app, you're, you're, you're actually provided with a really simple experience, which shows the things. We know you most want at the top of your list and offers you the most relevant deals, the most, most relevant promotions, uh, the most relevant coupons, and most importantly the most relevant products.
Because the endless aisle, while it's nice that, that, that means that, you know, you can have a, you know, a massive inventory online, um, you don't wanna encumber your, your shoppers with the realities of that. And LaSalle, you don't want people to, to scroll and make it a challenging experience to find the things that they want.
Um, so that's, that's a little bit, uh, you know, of a personalization example to bring to life, again, that [00:17:00] first party data asset, how we enrich it, and then how we apply science to it to create, you know, an inspiring and an irrelevant and simplified experience for, for our shoppers.
Steve Hamm: Yeah. Yeah. It's interesting. You know, I, I'm not gonna mention our local brands, but I recently wanted to make an apple pie, and I had the New York Times recipe for it, and I went to a local supermarket and they had like an entire row of different salty snacks, but they didn't have two or three of the key ingredients I needed for for an apple pie.
And I felt like, wait a second, maybe these guys aren't doing the right data analytics on this, um, this assign here. But, you know, I I, I can see how critical it is to be able to have the right things available and, uh, I I, I discovered that firsthand. You know, I wanted to ask you, you mentioned that Kroger is the biggest seller of sushi and flowers in the United States.
Did those things kind of emerge out of data insights, or was that kind of predating this kind of real, really, um, you know, heavy due to data [00:18:00] crunching that you guys are doing now?
Patrick Kelly: Yeah, it's a, it's a little bit of a combination of two. I mean, the floral department and offering sushi. It certainly, you know, it, it, it predates some of the like, advanced analytics, but I think it actually goes back to, um, you know, a, a, the kind of critical principles that we're talking about, which is, Understanding your consumers, what their needs and wants are, and then meeting them where they are.
And, um, you know, what we, what we found, and part of the reason that we have, uh, floral departments and so many of our stores, is that that's an opportunity that, that occasion going to the grocery store is one where, uh, those two needs, um, are actually joined, you know, are actually joined. Not to mention that, um, I don't know if you've ever been to a florist within a Kroger store, but it's a wonderful experience.
They've got great flowers, and it goes even back to, uh, the principles around, around freshness. You know, Kroger stands by, it's fresh for everyone, and it, it recognizes that 70% of of shoppers actually make their decision on, on where to shop based on, um, you know, the freshness of, of the, of the [00:19:00] food. And so, again, when you, when you start to expand on that, when you really meaningfully understand your shoppers, it presents you with an opportunity to grow your business, um, in meeting the, the needs of, of the shoppers.
And in this case, you know, A lot of people wanna be able to, to not have to go to another location in order to buy flowers or, or to buy their sushi, um, and your, your local grocery shop. Um, or, you know, or, or going online is, is a great way to, to capture that need for, for shoppers.
Steve Hamm: Well, it's tremendously powerful. I know that I go shopping every week for the family, and almost every week I bring flowers for my wife. And I gotta tell you this, this is about love. You know, how often do you go out shopping and bring home a little, a little bit of love to your loved ones? It, and so it's, it's, it's really powerful and magical.
So I
Patrick Kelly: Well, I'm, I'm taking, I'm taking notes Steve, and I hope everyone in the audience is as well. That, that right there, that's a pro tip.
Steve Hamm: You can make me, uh, you can do a study of me, you know,
Patrick Kelly: Yeah, [00:20:00] that sounds.
Steve Hamm: and you'll get, you'll get some good tips. Yeah. You know, you, you, we talked about the, the 60 million, uh, households and all the data that you guys have. Uh, I've heard you talk on, you know, some of the videos and stuff I've seen about the awesome responsibility that managing all this data entails.
So how does 84 51 address these trust and transparency issues?
Patrick Kelly: Yeah, it is. Um, it's a, it's a great question. It is an awesome responsibility for sure. Um, a lot of that responsibility goes back to, um, what I mentioned before about the exchange of value that loyalty is earned and it's, it's, it's earned repeatedly. Um, that value exchange is about, um, consumers continuing to, to opt in and leverage that loyalty program because they see the value coming on the other end.
And on and on top of that, you know, back to, to the purpose and the mission of both, of both Kroger to feed the human spirit in 84 51, which is centered around making people's lives easier. [00:21:00] Um, You know, building trust and transparency is, is key to, to that mission as well. And that responsibility fits squarely in our, our, um, you know, our philosophies as an organization.
That's both, again, with that relationship directly with the customer, but also in terms of how we leverage that data and insights with all of that whole ecosystem, all of our clients and partners. And the under the understanding of the consumer is matched with being an advocate for the consumer, so that as we engage consumer packages goods companies, it is so that we can jointly better strategically understand shoppers to improve their experience.
Um, you know, as, as we engage, uh, our media clients and agencies and publishers, it's to ensure. Wherever a, a customer is that they're being, they're being served relevant, um, information to them and they're, they're delighted, you know, when they go online and they see that their favorite products are, are on promotion and that, that's served up to them.
Um, so it, it really is pretty fundamental. And, and, [00:22:00] you know, the, the, the landscape of, you know, data and, and security and privacy is one that is ever shifting and, and very dynamic and particularly now. And so, you know, we hold ourselves to a really, really high standard, not just in terms of the regulatory, but we wanna, we wanna set our own standards that, that meet and go above those standards.
And again, truly be an advocate for customers across both the, the experience as well as the responsibility that we, that we have managing that first party data asset.
Steve Hamm: Yeah. Yeah. I want to talk about you a little bit, you and your role. You recently became SVP in charge of product and design. So what exactly does that mean? What do you, what, what is your role? And also, you know, you're, you're new to it. What's your strategy for taking the company to the next level?
Patrick Kelly: Yeah. Excuse me. Yeah. In, in my role, uh, leading product and design, uh, of course, you know, I'm responsible for managing our, our product and, and design and agile teams as well as for that enterprise product strategy. And a lot of what we've talked about today, I think highlights, [00:23:00] uh, you know, the complexities of the full ecosystem of, of grocery retail, as well as the complexity of, of all of, of that full ecosystem and all the stakeholders and all of the partners and all of the clients.
And so, you know, a big focus for me, Is not only continuing to enhance, um, the experience that we deliver to our clients and partners to make it easier for them to, you know, create better experiences, you know, invest in better products, you know, ensure that their, their r and d that their merchandising decisions are on the right path to, to growing their business with, with, with shoppers, which is mutually beneficial.
Um, but it's also ensuring that as we build all this complexity that we're not also making it more complicated. Um, it's easy. It's easy for a lot of these different areas of, of our business across insights and merchandising and media and loyalty marketing, um, for, to, to run in, in different directions. So one of the biggest challenges is continuing to also pull that back into a connected experience, you know, a, [00:24:00] a promise that we have.
Um, And Kroger has to, its shoppers is about a seamless experience, and we want that to even extend to the ecosystem that that brings the, the grocery retail experience to life. We wanna, we wanna make it more connected. Uh, we want the sophistication of sophistication of our sciences to actually create simpler experiences.
So things like going from insights to activation, um, become easier and easier over time. And there's, there's massive opportunities for us to do that. Again, whether it is, you know, providing the right data at the right time in the right context for operators, um, and, and individuals in the store to make sure that the stores in, in shape and has products on, on the shelves all the way to, you know, working with, again, brands and CPGs so that when they do get that.
Of, of, of insight that it's aligned to their strategies and it's actionable, um, so that they don't have to then take on the burden themselves to figure out how to, to not just make sense of it, but how they turn that into strategic action that ultimately impacts, uh, [00:25:00] the consumer. So, you know, to kind of wrap all of that up, and if I were to oversimplify it, it's, it's, you know, it's furthering, uh, a lot of the, the capabilities and science that we have, but doing so in a way that's, that's even more connected, um, than we are today.
Across, again, all of that, that complexity, uh, across the full e.
Steve Hamm: Yeah. Yeah. Now product, you know, that's the capabilities, the analytical capabilities or the availability of data design kind of suggests how it's delivered. And I would think you have a whole range of clients. I mean, we talk about users. I guess I would, I would, I would say, and every, everywhere from data scientists to business analysts and, you know, you're, you're creating a lot of new services.
I would think that UI design must be critical for you. And, and, and how do you, how do you manage that? And how do you manage to have the whole, you know, to deliver the services? In a way it's consumable to, you know, everybody from these real expert people to other people who really aren't technologists.[00:26:00]
Patrick Kelly: Yeah, you're hit, you're hitting the challenge right on the head, Steve, which is, um, there, there is a whole spectrum of, of users from non-technical users to to, to business users, to analysts, to really, you know, an increasing trend and, and data engineering and, and data science and machine learning across our own clients and partners, not just within 84 51.
Um, and when we think. Design, experience, design. Um, it, it, it even goes beyond the, the ui. Um, and just, you know, the kind of the buttons and colors on the page, so to speak. It's much, much more deep than that. It's actually about thinking about that experience holistically and designing for the experience and the workflows and the systems and how they all fit together.
Things like, you know, having a single source of truth is incredibly important because you don't want, um, and simply can't afford to have, you know, different answers showing up differently across these different platforms or across these different user types. So design is critical. As I think you're [00:27:00] alluding to, because it's not just about, Hey, we wanna make this report look pretty.
It's actually about thinking and really empathizing with that full spectrum of users, ensuring that the workflows that we build, um, the experiences that we build are, are, have, have those end users in mind. Um, and so that, that is, that is a real, a real challenge. You know, we're, we're also, you know, we're fortunate with the talent that we have, that we have, you know, we have great data scientists internally, we have great, um, you know, um, consultants and, and account managers that we get to leverage as, as both users, as well as, um, you know, people within 84 51 that have direct relationships with our clients.
Um, so, you know, I, I, I'd say that's a journey that's, that's definitely a journey That's a key focus, as I mentioned before, is balancing the, the sophistication and advancing of capabilities with not losing sight of ensuring that that sophistication doesn't make the world more complex. It needs to actually simplify the experience and, and drive more and more connection and design plays a, a key role in all of that.
And actually how it comes to life.
Steve Hamm: it seems like data is just [00:28:00] absolutely critical to the success of the company.
Patrick Kelly: Yeah, it is. you know, and, and at the end of the day, um, you know, Kroger is a, is a fantastic operator of, um, you know, of, of a grocery business. And increasingly, you know, Data and, and science and, and analytics and all these capabilities are, are a necessary part, um, of, of the business.
And they have been, um, Kroger has a history of innovating in terms of, you know, being, being quick to, um, to, to jump on new trends that benefit consumers. You know, um, a funny anecdote is that, you know, we talk about e-com and some of the, the kind of acceleration of, of e-com. Uh, due to the pandemic and, and other and other shopping preferences.
Um, Kroger today is actually a top 10 e-com, uh, retailer in the United States, which, which is, is wild. And when we go all the way back to the late 18 hundreds, when Barney Kroger founded Kroger, he actually had, uh, a delivery service himself, uh, where [00:29:00] he would get on a horse and he would deliver groceries to his local shoppers.
So, um, while e-com is, you know, is new and, and sexy and fresh and growing, it actually goes all the way back to the late 18 hundreds, um, when, when Kroger was founded.
Steve Hamm: So he was the original DoorDash, I guess.
Patrick Kelly: that's right, that's right. Uh, you know, talk, talk about being early to the. Uh, the, the other thing I would say to that Steve too, is that when, when, you know, when I think about analytics, there's to, to maybe oversimplify it, there's really three legs to the stool when it comes to data, um, science and, and business.
And you, you can't be successful, uh, without all three legs. You know, it, it, we focus a lot, um, on data and this and, and science and the sophistication of it. Um, and if you don't have good data, you know, you're, you're, you're, you're outta, you're outta luck, right? That's the foundation of it. But if you, you don't, you don't apply the right sciences on top of it.
Um, you're really under underutilizing and missing the mark with a, um, with how you activate on that data, that [00:30:00] data asset. But there's a third leg of the stool that, that will fall, that will fall over without it, which is the business context and the business acumen. Um, and, and that's one of the kind of pendulums that you see swing in the market as.
As companies and industries really start to take on data science and, and enter the world of, of prioritizing that as a capability, it's really meaningfully challenging as to how you bring those things together, and you cannot successfully deliver value on that meaningful investment without the proper business context.
And you talked about it earlier, whether your end user is an in-store associate that is, uh, checking products for outta stock and how you're augmenting that experience to make it effective and efficient. Or if the end experience is, you know, the real time, um, you know, personalization and, and serving of products and promotions online.
It's critical that you not just have the data and the science, but also that business context. We can't lose the importance of that. The world is not going to going, going to just, uh, [00:31:00] move to being, you know, run by data scientists who don't have that domain and, and business acumen. Um, and that's, that's one of the, you know, that's, that's one of kind of all the, the, the ever growing challenges of how you bring all three legs of that stool together.
Steve Hamm: Yeah. Great. So I wanna talk about Cloud for a minute. When and why did 84 51 begin using the Snowflake Retail data Cloud? And what are the main benefits you're getting from it?
Patrick Kelly: Yeah. Um, you know, we're, we're really, um, fortunate and happy to have the relationship that we do with Snowflake and, and really where it started for us is, Um, uh, there, there's a couple things I wanna mention. One is just kind of the cloud and, and cloud capabilities, generally speaking. You know, when we talk about our data asset, um, and all of the benefits of, of leveraging such a massive data asset and, and what that provides to 84 51 and, and Kroger and all of our clients and partners, it also provides a massive op operational challenge, um, because it, it historically is [00:32:00] required for us to be our own operators of data centers and purchasers of hardware and how do we provision and maintain and, um, and grow.
And so when you move to the cloud, it simplifies that, right? It, it takes that, that that competency, which is not necessarily a core of your business. And it puts it in the hands of, of companies like Snowflake that are, that are best in the industry, um, at offering, you know, the, the, the data capabilities, um, that are necessary in order to, to make use of the scale and size and value of our data.
And so that's fundamental and, and. Where the relationship started with, with Snowflake and, and our use of the Snowflake Retail Data cloud. Um, there's, there's two key things that really I think kicked off our relationship. One was, um, as you can imagine with the size of our data, there's an incredible computational intensity, um, depending on what it is that we're doing with our data, right?
So you can think of things that are really simple, like summing sales over time. Um, that that's, you know, in today's world, pretty op [00:33:00] operationally simple and, and easy. However, when you're looking across 60 million households and you need to do a dis, a distinct count, um, or, you know, you're, you're putting forward a, a assortment methodology that takes into consideration each of those.
It's not just, you know, summing sales, but you're actually taking into consideration all of the interactions that all your different shoppers have with all of your products. You can, you get to a point where, result in folding over trillions of rows of data. And when we started, um, to kind of test, uh, and, and do some proof of concepts around cloud providers that could handle that in a way that was, that was performant, um, snowflake really shined.
And, and that was, that was probably the beginning of our relationship, was around the ability for Snowflake to, to not just manage and handle this, this volume of data, but to also meet, meet our needs in terms of some of the unique computational intensity that we have. And that's really a kind of an internal facing, really necessary capability that [00:34:00] we have in order to, to surface up some of our things, uh, or some of the insights that we have.
And then the second was around the capability of, of data collaboration.
Steve Hamm: right,
Patrick Kelly: I, I think data collaboration is, um, you know, it's a really interesting concept because it's, it's so much of the discussion as we think about how we interact and how we're evolving our interactions with our clients and partners.
And frankly, historically, it's been a friction point and collaborating and sharing data, um, It's something you wanna be able to take for granted, right? You don't want the focus of your discussion with your clients and partners to be on, you know, systems and technology, and how do we do this and how do we do that?
You want to be focused on the strategic initiatives, the insights, the machine learning that you're building and working with a company like Snowflake that takes, takes the load the complexity of, uh, historical data, uh, complexity of data collaboration, and enables you to take it for granted. Um, it's a, it's a massive, massive capability, uh, for that and enhancement for, for us.
And so those [00:35:00] are, I would say, are like the two things that really started our relationship, um, a couple years ago with Snowflake. And we continue to build on those capabilities and, and we feel like, you know, we're getting the best of, um, industry leading capabilities through Snowflake, so we can focus on the things that we do best, which is building the science, um, you know, building this, those strategic insights, partnering with our clients, um, and ensuring that, that, you know, the insights to activation and ultimately the, the, uh, the result that the ends Chopper sees is maximized.
Steve Hamm: yeah, yeah. Hey, I wanna drill down a little bit on sharing and collaboration. When and why did you develop the collaborative cloud, and what role does Snowflake's technology play in it?
Patrick Kelly: So we actually launched the collaborative cloud, um, in the market last year. And, but I'm gonna go back, um, in time a little bit prior to that, the, the history and the legacy of, you know, being an insights provider within the grocery retail space was predicated on, um, You know, web-based applications and [00:36:00] syndicated reports and, and inevitably, um, that requires the provider, you know, 84, 51 to have a curated view of how we present insights.
Um, and then the end result of that are aggregated insights that come out of these web-based, uh, reporting applications and that application. Really meaningful value and continues to have really meaningful value, uh, in the market. It enables a lot of, uh, non-technical users to be able to get fast, relevant insights, um, that are really, that are really pertinent for how it is that they wanna continue to manage and, and, um, action in their business.
However, there's, there's an emerging trend over the last several years to compliment that, which is we're seeing the industry really invest in, in data, in technical capabilities and, you know, advanced analytics and data science. And so we had an opportunity a couple years ago to kind of take a step back and say, you know, how is, how is it that we can actually reimagine our experience with, with the industry, with our clients and [00:37:00] partners?
And what we realized is that we, we really have to go beyond the web-based application, the kind of syndicated reports we have to, we. Build and offer capabilities that enable collaboration on data science and on that data asset. Both meeting clients where they are, but also catalyzing their journey, um, and their investment in data science.
Cause that's a meaningful investment. But for all these, you know, all of our clients and partners to become. Science and data organizations is no small task. You're, you're asking, you know, consumer packaged good companies where the core value proposition of their business is to manufacture, um, and sell goods to also be a leading technology organization so that, you know, that entails technology decisions, you know, capabilities, data science, you know, data scientists and the skill sets, um, that they need.
And so by partnering with 84 51, again, we're able to not only support them and catalyze that journey, but we're able to ensure that their, the return that they see on that, on that investment is [00:38:00] maximized. So what the Collaborative Cloud is,, is an experience that's specifically built for those, those new users in the industry.
Um, and it enable. Our clients to get the best of 84 51, not just, you know, access to our data, but also, you know, the building blocks of our science. So, you know, as you know, as it's, it's aptly named. So that can be a collaborative exercise. Um, and, and ultimately it, again, meeting the needs of where our clients are now and where they're going, um, and data sharing and data collaboration is a, is a meaningful part of that.
Steve Hamm: Yeah. Yeah. Um, is,
Patrick Kelly: Hey, hey, Steve. Steve can, so Steve, actually, I'd like to share an example, um, to, to kind of bring some of that, that value to life. Um,
Steve Hamm: of course. Go ahead. Just sh go.
Patrick Kelly: Yeah, so when we, we actually were, uh, piloting the collaborative cloud prior to launch. Uh, we were working with one of, one of our clients and, uh, had agreed to, to a pilot, um, right at the beginning of, of 2020.
Um, [00:39:00] and we'd scoped out, you know, initial use case and analysis, uh, around market basket analysis, that that tied to kind of some of the way that their marketing teams both kind of viewed, viewed data and insights as well as activated on it. But as you can imagine, at the beginning of 2020, uh, the whole industry was sidelined by the pandemic.
And what we realized through that pilot is, um, that, that, that in March of 2020, as the pandemic hit everyone in that organization, in that, in that cpg, from the CEO executive team on down, was asking the questions of what is happening? What are our shoppers doing as a result of this pandemic? And the realization was that, well, through the collaborative cloud, our clients actually had access to.
The 84 51 data asset and it really removed the black box so that their most important strategic questions they could take, they could take ownership of how they wanted to answer those. And so that executive team turned to their analytics organization and they had [00:40:00] access to 60 million households, half of the US' behavior, and the data is clean, it's trusted, and it's ready, and it's last week's data available on Monday.
It was a transformational opportunity for us as we realized the true power and value of data collaboration because they shifted priorities away from the market pass analysis directly onto what are the impacts of, of, of, of the pandemic as things were going out of stock, um, that were, you know, core products within their portfolio.
What were the shoppers doing without those products on shelves? Were they moving to other products within that brand portfolio? Were they moving to competitors? Were they switching down on size? And those, those, those shopper insights, Impacted everything, all the way up to the manufacturing of that organization.
And I bring that, I bring that example to life because it really demonstrates the power of, of data collaboration. It, it significantly reduces the lead time to go from strategic need to insight to action. And while it was the [00:41:00] pandemic, you know, in, in 2020 and it's continued today, you know, it's inflation today, it's the holiday seasons, whatever the strategic initiative is.
Our clients and partners are able to take advantage of leveraging the collaborative cloud to go after their, the things that are most strategically important for them. Um, and, and, and leverage both 84 51 s data assets as well as our science in order to make sure that that, that, again, that time to value is drastically shrank.
And that the, the, the total scale and the representation of that data asset gives them the confidence to make strategic decisions.
Steve Hamm: Okay. Gotcha, gotcha. What role does snowflake's technology play in the collaborative cloud?
Patrick Kelly: The collab. Yeah, so the collaborative Cloud, as I mentioned before, and you know, of course, is in its name, is really all about collaboration. And where Snowflake really shines is in making collaboration and making, you know, things like data sharing. Um, just, just simple and easy, being able to take it for granted.
Um, [00:42:00] of course, you know, on, on this podcast, and, you know, even within our teams, we, we like to talk a lot about, you know, the data and, um, all of the kind of capabilities and experiences around it. But the reality of it is, is that when we engage with our clients, we wanna be able to take those things for granted.
And with Snowflake, we're able to do, we're, we're able to do just that. We don't actually have to think about the complexities of data sharing and collaboration. We can take that for granted and focus on the strategic initiatives. plays a, a fundamental role in, in ensuring that, you know, our, our ability to, to to share data, but really more meaningfully is collaborate on, on data.
We want to, we wanna build science, um, jointly with our clients. We wanna meet them where they are and, and accelerate their journey. And Snowflake is a, is a pivotal capability in order to do that seamlessly. Um, because coming out of, of the collaborative cloud, we also wanna ensure that, that that insight, um, meets its end user.
While the data scientist or the data engineer might be the actual user for the collaborative cloud, they're likely not the business stakeholder that needs that insight to make a key, key decision. So, so [00:43:00] Snowflake makes it so not only can we, you know, collaborate on data effectively, but then that data can easily meet its end destination and meet its end user so that they can make a better decision for, for their business and ultimately for their shoppers.
Steve Hamm: Yeah. You've talked about how important data sharing and collaboration are to 84, 51 and your clients whenever you, whenever you share data, whenever you start collaborating on data, there are issues of ownership and access and things like that.
So how do you use clean room technology to kind of merge the data without compromising privacy of the end customers or the ownership boundaries of the data?
Patrick Kelly: Yeah, Steve, you're, you're spot on. Um, in that, you know, the, the, the ideas of ownership and data governance are, you know, are, are fundamental. Um, and historically, frankly, they've been, they've been deal breakers to meaningfully collaborating, um, and on, on data and on on science creation. So, [00:44:00] you know, the introduction of, of clean room and clean room capabilities. A new, a new tool, uh, in the tool belt. And while I think clean room right now is a, is a, you know, is a very, um, it's getting a lot of attention and it's definitely a buzzword. You know, our perspective isn't necessarily that it's a silver bullet or it's a panacea, but rather it, it's, it's another capability that unlocks, um, new use cases and, and new opportunities.
And so specifically, you know, when we talk about data collaboration in the collaborative cloud, you know, there's, there's not a lot of, you know, governance concerns with maybe, uh, a client bringing in product attribution, um, into the collaborative cloud and, and enhancing their view of products and what that can do for analysis.
But when you start to talk about other first party data assets, that, that, that quickly becomes a deal breaker. And, um, the clean room is actually, is an unlock for that because you can bring first party data assets together in a way that is, is highly secure and highly governed. And then have some of that collaboration happen within the [00:45:00] clean room.
And, and, and again, Per, you know, really unlock use cases that just simply weren't available before. Um, how we're leveraging that today is, you know, we're, I think we're really ear early and I think the industry's really early in realizing the full potential of, of the clean room. Um, there's a lot of value and, and really simple analyses like, you know, just seeing the overlap of, of.
Um, you know, some first party data assets to do some, some basic kind of descriptive, um, analyses that, that shed meaningful strategic light on. Like, oh, you know, what's the sales performance of, of, you know, loyal, uh, uh, Kroger shoppers, and how does that overlap with maybe the, the, um, you know, the CRM of, of a given client?
Where I'm excited for it to, to continue to go, to go and, and really excited with the developments that Snowflake continues to put out in its investment in the collaborative cloud is to actually be able to do science, um, in the clean room. Um, because then you can actually apply machine learning directly into that clean room and, and you can, um, you can really, [00:46:00] really amplify the value that you're able to get out of that collaboration.
Um, you know, as, as you mentioned in your question as I started with, you know, security and governments are first and foremost for us, for us, I mean, that's, that's, those are standards that we're just not willing to compromise on. Um, in the clean room. Again, it just expands the, the surface area by which we're able to collaborate effectively with our clients.
Steve Hamm: Yeah. Yeah. You mentioned being able to do data science within the clean room. Well, let's, let's talk about future capabilities here for a minute. When you look forward over the next year or so, what are the key data analytics technologies that you see emerging as key for the retailing industry and, and key for you guys?
Patrick Kelly: Yeah, there's, um, I think some of the, the, the kind of future capabilities that, um, that I think are really relevant are some of, some of our, our advancement of existing capabilities. I think clean room's gonna play a meaningful role in that. Um, again, a lot of this actually has to do with not just advancing the capabilities in themselves, but the advancement of what that capability can [00:47:00] unlock.
Clean Room is a great example, right? We're able, you know, we're able to build great machine learning algorithms today, you know, leveraging languages like Python. Um, but we build that on our own first party data asset as, as a capability of the clean room. You know, continues to progress and, and Snowflake continues to put enhancements and features into that.
It's gonna enable us to actually move some of that machine learning into the clean room and fundamentally enhance what we're able to get out of it. So I, I think, you know, advancements of, of kind of current capabilities on, on a, you know, a linear or, or an exponential path are gonna present a ton of value opportunity on top of that.
Um, you know, there, there's also the spectrum of, uh, of connectivity. And I, I think this is, this is one of the most important things that we, we all need to focus on, um, as we advance our analytical capabilities, which is the connected experience, um, across analyses so that, um, everyone from, you know, a business user to a data scientist, to a data engineer, to an analyst can all perform the [00:48:00] necessary analyses that they can't, that, that, that they have in their own pockets, but they're not actually, um, you know, the experiences themselves are not in pockets.
So one way that that could come to life is. When you present or perform rather a, you know, a data science machine learning algorithm, that that's not just done in, in, you know, in some provisioned cloud, um, environment. Um, but instead that's very much connected to the interfaces that the business that the business teams use.
And there's a seamless connection there. Similarly, as a business user may be defined to the segmentation or, you know, puts a perspective, um, onto the data in terms of the, the importance of certain product attributes that, that seamlessly can flow into the data science environments. That, that an that, that your, um, your scientists are using.
Um, cuz today, That while all those capa, a lot of those capabilities may exist, they exist in pockets. So there needs to be this, this meaningfully connected experience. And I think, you know, the near term innovation is simply gonna reduce the friction to have connected experiences [00:49:00] across, um, our clients and partners as well as within 84 51.
So you, you can really make sure that you wanna have a single source of truth. And two, that the collaboration across this wide spectrum of users is really, really intensified through, you know, easy and seamless connection. And, and, and in some ways it almost kind of feels like magic, right? A lot of that complexity needs to be behind the curtain in terms of data movement, data sharing, data collaboration, as well as science collaboration.
Um, so I, I think that's a lot of, kind of the near term of, of, of where we're going. Um, and it's, it's a lot of the advancement even that we've seen to date in working with, uh, with, with Snowflake.
Steve Hamm: Okay, so that's the near term picture. I'm gonna ask you to put on your visionary cap for a minute. Looking out five years or even more, how do you see data analytics transforming business or even society?
Patrick Kelly: Yeah. Um, I love this question. I always think it's really fun to, you know, to think, to think five years out. Um, you know, some of the things that that come to mind for me, I mean, is, is AI and like the [00:50:00] true, you know, kind of the true sense of the word. I know that, that, that sometimes the word AI gets, gets a little bit overused and co mean a lot of different things to different people.
But as we look at some of the capabilities that are being actively developed today, um, you know, companies like Open AI and capabilities like stable diffusion, um, and if the, if you know, if you're in the audience and that doesn't mean anything to you, and in some ways I'm actually jealous because you get to, you get to go and onto Google and search those things and just be like wowed by the capabilities that are, you know, have even come out in recent months.
But, but what we're seeing today, That AI is actually able to create art. It's able to, to create, to create copy. You know, g uh, g PT three is able to write, is able to write like a human leveraging ai. And so as we think about, we're just on the edges of, of what's, what's possible and what's feasible from ai.
And oftentimes the actual capability kind of outruns, um, the, the kind of valuable application. And we're starting to see that the, the meaningful application of that capability catch up, um, [00:51:00] to where, where the science is. So when I think about five years from now, bringing that capability over into, um, into business more fundamentally and into grocery retail, that's everything from, you know, a ton of, of predictive and prescriptive decision making.
I really believe in the power of augmented decision making where it's not necessarily about, you know, taking the human outta the picture, but leveraging machines and science and AI for what it does best. And then leveraging humans for what they do best in terms of understanding, you know, the broader context, um, having some creativity, understanding strategy, and how you bring those two things together.
I think we are, you know, despite all of the impressive innovations and advancement we've made to date, you know, I think we truly are in the early days of, of unlocking the power of, of ai. Um, all of that, you know, coupled with what, with what makes it feasible, which again, is, comes back to computational, um, intensity and the cost.
And those, you know, based on Moore's law, those are things that are simply just going to, [00:52:00] you know, those friction points are just gonna continue to, to decrease at an expe, uh, exponential pace. And I'm, I'm, I'm really excited for it.
Steve Hamm: yeah. Yeah. It's interesting you mentioned G PT three and uh, the other one is Dolly two. That's the, that's the art. That's the art one. These are based on foundation models and which has really been the explosion in the past couple of years. You know, using deep learning, using neural networks, things like that to really, uh, take on gigantic corpus or corpuses of of data, different kinds of information, and, and you've got it.
I mean, you know, the Kroger universe of data is that kind of gigantic kind of multidimensional thing that can really benefit from this kind of technology. So that's gonna be really cool to see what you, what you guys do with that.
Patrick Kelly: Yeah, Absolut.
Steve Hamm: Yeah, yeah. You know, we typically end the podcast on a lighter, more personal note, and I understand from our previous conversations that you actually love to go grocery shopping, especially with your [00:53:00] family, even with the three kids in tow.
So tell us what's so great about that. Why do you love that? You know, what's the, it sounds chaotic to me, but
Patrick Kelly: Yeah, no, I appreciate, I appreciate the, the question. I think, you know, in some ways, maybe I'm just inflicted by, uh, by my profession, cuz you know, personally it's hard for me to walk into a grocery store without just, you know, seeing and appreciating the, the millions of decisions that need to be made in order to curate that experience.
So I just, I have a great appreciation for, you know, what products were selected, you know, where are they in store, what are the price points, what are the promotions? Um, so there's just an intellectual curiosity that I have when I, when I grocery shop. But I think, you know, more importantly, you mentioned, you know, I, I do love to grocery shop.
I also love to do it, you know, to, to bring my kids along. I have, I have three boys and, um, that are all, you know, at the toddler stage and just food and. And grocery is so centered to family connection. I mean, for, for us on any given day, the one time that we are guaranteed to be get together as a family is, is around the dinner [00:54:00] table.
Um, and, and so like that, I don't take that for granted. It's not, it's not lost on me. So when I get to go to a grocery store and, and shop the aisles and bring my kids and my wife along, um, you know, seeing what my kids get, get excited by what inspires them. And as well as, you know, oftentimes it's because we're anticipating, you know, cooking a meal together and I get to do my own discovery and exploration and, and find new foods and flavors.
So I, I just, I, I love that experience. Um, and to your point, it, it certainly is chaotic, but like, that's the, that's the, there's beauty in that chaos, uh, cuz that's the day to day, um, you know, experiences that you get to have with your, your, your children. And I'm, I'm super happy that, you know, food is a, is a definitely.
Is a binding experience and a, and a loving experience for our whole family. Um, so yeah. And ho with the holidays coming up, you know, there's gonna be a lot of, there's gonna be a lot of grocery shopping and there's gonna be a lot of food, um, that brings us all together.
Steve Hamm: yeah. You know, I grew up in the fifties, you know, shocking. But back [00:55:00] then it was all white bread and peanut butter. So I, I'm glad that things have gotten much more, the, the, the variety have gotten much more, more, uh, broad for, for the kids these days.
Patrick Kelly: me too. I'm not sure I'd have that same perspective if that was, uh, just, you know, being, being offered white bread and peanut butter. There's just, there's so much fun and innovation, uh, to explore in a grocery.
Steve Hamm: Yeah. Yeah. Well this has been a wonderful conversation and you know, it seems like, you know, week after week I'm talking to, to partners or, or clients of, of Snowflake that are dealing with just huge amount of data and it just is absolutely mind blowing. And I just to think about Kroger and the, the assets that Kroger has and the, and the tools that 84 51 has is just, is really very impressive.
And when I think about kind of the, I think about your collaborative cloud and that incredible amount of data from, you know, first person data from, from [00:56:00] Kroger, it's almost like. It's almost like a, a public utility or something for the CPG industry because there's just so much of it. It's so rich, it's so multidimensional.
So it's, it's, um, really fascinating to learn about this and I'm, I'm really glad we got a chance to talk about it. And I think, I think our podcast, uh, listeners are really gonna enjoy this one.
Patrick Kelly: Yeah. It's been a pleasure, Steve. Thank you. I appreciate, uh, I appreciate the invite.