Data & Analytics as a Source of Resilience with Boeing and Walmart Tech Execs | Technovation 699

Welcome to Tech Novation. I'm your host, Peter High. Our broadcast today comes from our most recent MetaStrategy Digital Symposium and features a conversation I had with Vinod Bidarkoppa and Susan Denise. Vinod is the Senior Vice President at Walmart and Chief Technology Officer of Sam's Club, a role he's had for roughly two and a half years. Prior to that, he served as a Chief Information Officer at UnitedHealth Group and at Tesco. Susan is the Chief Information Officer and SVP of IT and Data Analytics at Boeing, a role she's also had for roughly two and a half years. She's been a CIO multiple times over, including at Qantas. Immediately prior to her current role, she served on multiple boards, including that of Bayshore Healthcare in Canada. In this interview, we cover the topic of data and analytics as a source of resilience and growth. Susan explains the role data and analytics play in delivering predictability and stability as part of Boeing's overall strategy, how she trains her team and the skills needed to achieve that, and why data is integral to a company's culture and telling its story. Vinod describes how data and analytics impact the customer's experience at Sam's Club and how he's making Sam's Club into an AI-enabled digital enterprise, as well as how he's democratizing that data across his business. I hope you enjoy the conversation. And with that important topic, we're going to bring on two bright lights in this field to talk about data and analytics as a source of resilience and growth. Certainly a relevant topic, but boy, as I pull individuals or in groups, those items that are of greatest importance, and not so coincidentally where a paucity of great resources are, data analytics tends to be at the top of that list in both categories. And certainly something that any technology and digital executive needs to have actually or near the top of their strategic priorities. Susan, I'd like to begin with you. I know from our past conversation that Boeing's corporate strategy is about predictability and stability. You've noted, and I wonder if you could talk a little bit about the role that data and analytics play in delivering that.

Yeah, great summation there, but also appreciate some of those targeted examples as to where it's been particularly valuable. Susan, staying with you for just a moment longer, we've talked about building of talent. You know, data and analytics as being some of those areas of greatest and highest demand. Can you talk a bit about how you thought about building your data and analytics talent on your team? What are the methods you've used, combinations of training versus finding great talent from other companies? How have you thought about that, please?

That's really great. Thank you for sharing those anecdotes. But Vinod Bidarkoppa, I do want to bring you into the conversation. I mentioned you are the Chief Technology Officer of Sam's Club, and you're at least a 75 billion dollar division of Walmart. Sam's Club, different from Walmart, is a membership-based organization. And I wonder, especially in light of some of the variations of that model, how you use data analytics to impact customer experience, as we just saw one of those areas that our audience is anticipating, many of them anyway, putting the data and analytics to work. Talk a bit about the way in which you've done so within your organization, please.

Sure, thanks for the question, Peter. So, as you said, Sam's Club is an integral part of Walmart, and Walmart has three distinct divisions: the Walmart US, Sam's Club, and Walmart International. So, as part of, I mean, you'll see technology in today's world drives everything from what I call as from mundane to magical. And the reason for that is business is becoming the platform, the technology platform is becoming one and the same. And in order to do that, I mean, data is the central aspect and the central driver and the underlying theme. So Sam's Club is a warehouse model retailer, and what that means is really it's a membership-based model. And anybody who has to shop with Sam's Club has to be a member of Sam's Club. And the other aspect of this model is that we have a very finite SKU item model. So by virtue of being a membership-driven model and a finite SKU model, we have the distinct advantage of really getting detailed data about everything that we need to drive this business. So with the data as a fundamental ingredient of our business, we use data for almost, I mean, if you peel back the retail processes, it's planning, the buying, the moving, and the selling. Those are the four core processes of any retail business, and we have embedded data into every aspect of it. Whether it is experimenting with different business models, whether it's the A/B testing, the design thinking, and the iterative concepts of how we actually advance a different thinking and then bringing different models to bear, data is a central aspect of it. So let me give you a couple of examples. And so what we have done, Peter, is, you know, when you look at whether it's our omni-commerce site, so we've brought in the fraud and the security aspects. So we use data to look at what is our fraud level. So we've used the data to drive our fraud levels down by almost 35 percent there with our members. Right? So we're looking at how can we personalize and drive more convenience for our members. So by personalization, we mean can we add an item into the basket that the member would like to add at the point of interaction, whether it's in the club channel, whether it's the online channel, and so on. So by using this data in whether it's a member interaction channel or in the club channel, we're able to drive a higher revenue growth. Equally, on the associate side, we're able to use the data to look at how can we make our associates' lives easier in the club, whether it's in the supply chain. So we're driving a lot of, for example, we scan about 17 million images every day in the club, and these are compute, we use computer vision, and use these images to look at how can we automate tasks for our associates and really release, you know, make some of those mundane tasks for our associates to be done by machines, and then have the associates work with our members in the club and make their jobs more easier, as well as give our members a chance to interact with their associates. So I think the data and the aspect of insight and analytics is driving every aspect of our business, whether it's the merchants, the associates, and including our members. It's across the board.

Yeah, very interesting. And that last anecdote especially, even though, brings to mind something you mentioned to me when we recently caught up, that your vision is to make Sam's Club an AI-enabled digital enterprise. That was the phrase you used. I found that so compelling, and I hear echoes of that, an example of that, in fact, in what you've just described. I wonder if you can share a bit more of how you anticipate and even how you have already begun to bring that to life.

Well, absolutely. I mean, just double-click on a couple of those examples. So in the merchandising space, I mean, we have the need, and one of the things we look at is value to the member. So value to the member means, I mean, how do you drive better prices and make shopping easier for the member. So we looked at how our members shop with us, what are the prices of our items in different parts of a market, and where can we actually drive more price investment so it becomes more affordable to our members. So price investment is a good area where AI is enabled. The other example is markdown optimization or exits. This is where we probably have a little bit of extra inventory, we need to sort of exit them from our system so that we can bring in newer and fresher inventory. How do we do this without impacting our bottom line and the business model? So there's another place where we can drive more AI. A third example is we produce a lot of what I call as the fresh prepared food in our clubs. You produce too much of it, then there's wastage. You don't produce enough of it, then you have unhappy members in the club. So how do you find that happy balance between producing the exact amount of what you need so that you don't displease your members, but equally you're not driving a lot of waste in the system? And all of this is really enabled through data and building the AI models on top of that so that we can actually produce what is optimized for both for our members as well as for driving waste out. Would be a couple of different examples there. Fraud is another great example. I think in today's world, I think there are many threat actors who are looking, and we drive millions of transactions, whether it is the point of sale or e-commerce transactions, and we have to be always a step ahead of all of these threat actors out there. And the way we do that is we bring all of this data from a point of sale and e-commerce transaction, the credit card, the debit card, and members, and in real time, really run highly sophisticated fraud models in order to prevent some of these what I would call as bad threat actors impacting the shopping experience for our members. And we can't do any of this unless we have rich data, unless we have sophisticated AI models that help drive our business day in and day out.

Yeah, really interesting examples there. Thank you for those, Vinod. Susan, going back to you, I know from a recent conversation that you and I had, one of the things I really enjoyed was an insight you shared that a good data analytics player needs to be a good storyteller. That's in some ways one of the indications as to how talented somebody will be in this space. And I wanted to have you share some of your perspectives on that as to why you believe that to be so.

Yeah, really interesting. And I also wanted to ask you about the cultural element of this, Susan. I think so often as we think about data and analytics, we think about the solutions, we think about the algorithms, of course. But you know, culture really trumps all of this. If we don't have a culture that's accepting of this, that understands the kind of long-term implications, gets over the hurdle in some cases of the feeling of a loss of autonomy by combining say one unit's data with others in order to make better decisions across the entire organization, to name one of a variety of complexities to manage through this, then the value you will derive from this will be sub-optimized to say the least. I wonder if you could take a moment and talk a bit about some of the cultural elements that you think have led to successful adoption and use of, and ultimately value derived from data analytics at Boeing.

Yeah, very interesting. I appreciate you raising that and also the allusion to the data sovereignty and other complexities that we all need to manage through, especially in large global organizations like your own. Vinod, I know one of the ways in which you leverage data for daily, weekly decision making is through NPS. This is a metric that you keep and discuss as a leadership team with great frequency. NPS will certainly, that's Net Promoter Score, but I imagine that most who are in attendance have some familiarity with it and perhaps have been users of it. But I wonder if you could talk a little bit about how it has impacted the way in which you think about changes to the business or validation of what's working well and so forth.

Yeah, so one of the things that we're very proud of at Sam's Club is, of course, we are very data-driven. But it's not just saying we are data-driven, how do you really bring that to life? So the first thing that happens on a Monday morning is, as a leadership, executive leadership team, people look at what was the business the prior week. And the way we deconstruct that is really through understanding what was the NPS for each of the core critical member journeys, and in some cases, associate journeys. So once you know what the NPS for each of those member journeys is, you know, whether it's in the club or e-commerce, or whether it's returns, or digital joins, digital renews, you name it, right? So by understanding where we are, and when we set our target for each of those journeys, and if we see either an uptick or a downtick, we start double-clicking into it. And because there is data behind it, and then people are able to answer those questions in a very data-driven way: why did we see a downtick in an NPS with a certain customer journey? So that makes it a very rich conversation, and it's not somebody's opinion, it's actually, you know, what happened and why did it happen and how do you need to go fix it. So that's really the basis of how we run the company. And then the rest of the week, it's really the team kind of trying to figure out, hey, how do you fix this so when you come back in the next week, they really address that at the root. And our CEO calls it really grinding it to the powder, how do you really become remarkable at some of these journeys is really trying to focus and zoom in on those NPS. And we couldn't do that if we didn't have great data that enables running our business.

So just as an example, Vinod, in the last couple minutes that we have here, I also wanted to ask you about issues of model drift. I have a friend who's the Chief Digital Officer of a Fortune 100 company, and he talks about how the CTO, CDO, CIO needs to be sort of the CHRO, if you will, of algorithms, firing some, promoting others, suggesting workarounds or development plans, better put, for still others. And I know this is something you think about a lot as well, is how that evolves over time in order to make sure that model drift doesn't become an issue. Talk a bit about, if you wouldn't, in brief ways in which you've tackled that challenge.

Yes, so I think one of the things for us is really becoming this AI-enabled digital enterprise and really democratizing AI across everything. It's easier said than done. As what we saw in the pandemic is really the changing behaviors of our members, and as we implemented a lot of these models over a period of time, as the behavior started to change, obviously the data that was coming in was also changing. And the models that are built at a point in time, and because a lot of these are becoming black boxes for us, and we run our business based on the outcomes of these models, they're becoming very dangerous in some sense. So if you don't really monitor and really have those gating systems in place and see how, you know, you're monitoring what I call is the data drift which is causing the model drift, which ultimately impacts the outcome. And if you're making decisions on the basis of those model outcomes, and it's a dashboard for the pilot who's running in the cockpit, take the analogy of Boeing here, it can become very tricky and sometimes even dangerous. So really monitoring, what we have implemented a system within Sam's Club is to really look at what is the incoming data, what is the model, and use their model drift, and how are we enabling making sure that there's no model drift so that we are, one, conscious, and two, able to give that confidence to the business.

Yeah, that's a great overview indeed. Well, Susan Denise, Vinod Bidarkoppa, thank you both so much for the great conversation that you've simulated here, the great insights you've shared about the journeys each of you have been on and the big impact that you've led in your organizations. It's been a great conversation. Thank you again.