Evan Wimpey: Hello, and welcome to the Mining Your Own Business podcast. I’m your host, Evan Wimpey. Today, super excited to introduce our guest Sravan Vadigepalli, who is a senior director of products and engineering at Lowe’s the home improvement store. Sravan, thanks so much for joining us on the podcast today.
How are you?
Sravan Vadigepalli: Thanks. You and glad to be here. I’m super excited to be here and would love to talk more about data, data science, data analytics, and anything related to data. So looking forward for the conversation.
Evan Wimpey: Perfect. My kind of guest. Me as well. Always happy to chat data. Sravan, maybe to get started, can you just give folks an idea about your background, where you came from, how you ended up in the role that you’re in now at Lowe’s?
Sravan Vadigepalli: So right now I am a senior director at Lowe’s. I lead both products in engineering side of the house. And before Lowe’s I was at Best Buy and Target. So if you can see my profile, it is completely retail in, in data and analytics space. , I did my master’s in information systems. , so I do have the computer science background.
And I did my bachelor’s in electronics engineering. So mostly, like, how do we solve problems? What problems do we want to deal with? And that’s my. The area of expertise is, let’s think about a problem and find an optimal solution on how we solve the problem. So that’s why I like to be in the data side of the house.
When I think about Best Buy, it was consumer electronics focused. And Target was more of a discount retailer even though Target had like five core categories, be it your clothing and grocery and also some electronics, right. Then, when I came to Lowe’s it was Home Improvement. It is a little different, even though it’s a retailer, it’s a little different.
So because of the, the customers we deal with, or the type of products we have, but mostly what I like to do is, doesn’t matter which type of retail you are in you have a similar set of data problems, you want to solve for. So that’s what that keeps me going within this field.
Evan Wimpey: Oh, fantastic.
Yeah. And the retail background I’m trying to marry up, you know, as a, as a consumer, as a shopper, it feels very different to go into a Lowe’s than a lot of retail outlets. Is the data roughly the same there? Are there, are there new challenges that sort of you’ve seen once you’ve gotten to the, the home improvement space, maybe at scale or just professional services that, that are there?
Sravan Vadigepalli: I think the main, main difference that you would see, if you think about let’s say Best Buy or Target or Walmart for that matter is the type of customers are more demographic based. That when you think about home improvement retailer. You basically have three main core segments of customers, right.
You have your pro customers who are the ones, the plumbers, the traders, they are the ones who are working for this big enterprises. And then you have your DIY customers. And we all know the rise of DIY during COVID era. Like everyone was a DIY customer, right. And then we have DIFM customer, which is, hey, I’m not necessarily really good at doing this stuff, can I hire a pro to help me with this project I’m dealing with.
So I think that’s the main differentiator between, let’s say, Targets of the world, Walmarts of the world too. Home improvement like Lowe’s is it’s more customer segment based in, in, in terms of in terms of the complexity. And the data is also complex in, in, in, in the sense of, now we are talking about different levels of product complexity.
For example, your pro customer might be buying stuff in bulk, right. The pro customer might be, Buying lumber versus your DIY customer might not be buying in bulk, they might be making more trips to the store. And your seasonality with the DIY customer could be, hey, it’s spring is coming, it’s sunny outside, let’s go do some shopping.
Versus for pro, it doesn’t really matter, I mean, they’re going to do the shopping no matter what. , the, the seasonality is going to be. So, those are some of the nuances with, with the data and how the data is situated, right. One thing I always say is, for data folks like us, you need to get into the weeds of the business.
We need to understand the data. Who the customer is what do they make purchases with us, what is their frequency? These are like the basic statistics that we all need to know being in the data group. Then you can think about what products are we going to build for these customers. , we have a saying internally and I might be butchering this a little bit, but the intent is we either serve the customer.
Or we serve those who serve the customer. So, when we think about the roles within the corporate office, we don’t necessarily interact with the customer directly, right. Our store associates are doing that, being the front line with the customers. But we are supporting our store associates. We are supporting our corporate business functions with the data insight.
So, that’s how we navigate internally. So, for you to do an effective job at it, you need to know the basics of the business. They’re different. It’s all retail. But it is specialized retail. That’s how I see it.
Evan Wimpey: Sravan, I think that’s a great attitude to have about it. I think it feels like there’s almost risk that you walk into a role like this and think I’ve been at retail places.
I know retail, I know what to do, but I think that let’s focus on the customer, let’s get to understand the types of people that are buying the types of things, the process they’re going through. I think that’s a really healthy thing. And it’s, I would.
Sravan Vadigepalli: Yeah, I’m not to cut you off, but I would give you an example when you are in retail, the common pretty sure we all heard about RFM, Recency Frequency Monitoring, and when we define new customers, repeat customers, we generally have a similar definition, and then you slice and dice the data by different demographics.
For example, if you’re looking at. Hey, I want to know how many new customers came in. That definition is generally going to stay the same for all of your customers, but you just look at the data differently for each of the demographics. If, if you are working for Target, Walmart, or Best Buy. When it comes to Lowe’s, let’s say if you bring the same attitude, you are going to be completely missed and wrong.
Because if you say that my repeat customer who is coming back in six months from now, you That’s not going to be true for your pro customer. I mean, we want them coming in like every week. So thinking about the data, if you don’t have the business context, if you don’t know who your customer is, You cannot necessarily build the tools that are going to solve those problems.
Evan Wimpey: Yes, that’s a great example that really hits home. And you think about like the gap and the seasonality and the DIY it, this hits home as well. I was, I was trying to sneak a peak at Lowe’s website, trying to find out when the mulch is going to go on sale. I gotta do my spring, spring mulching.
Sravan Vadigepalli: You can buy mulch online by the way, and people do, so you can go to lowes.com. You can look for brown bag watches and you can buy them today.
Evan Wimpey: All right. That’s where I’m going. As soon as this episode’s over. Perfect. I want to point out something, your title, your senior director in engineering and products.
And, you know, I get a chance to take a look in several organizations that have data teams. There’s a lot of data Experimental analytics effort. There’s a lot of project-based work. I think a lot of folks would love to be in the position where they’re developing data products or analytic products. Can you talk about what that process is like and sort of how that’s grown?
Sravan Vadigepalli: I think Lowe’s is very unique in terms of home improvement retailer. And you would be surprised by the level of maturity that we have here within the organization. At the core of it, we are a product-based organization. Right? And we are a product organization with centralized data and analytical capabilities.
What that means is we have data engineering, we have software engineering, we have machine learning, data science, and AI, all sit in, all sitting within the umbrella of tech. And we serve all the different business functions accordingly. The beauty of this is product team is still going to be the front facing for the business.
What that means is, let’s say we are we are building a scorecard to measure a given business performance. What happens in this case is, for us to build a scorecard, there are different players in it, right. You got your engineering. They’re the ones who build the pipelines. They’re the ones who make sure the data is right.
They’re the ones who make sure the data quality is right. And then you have UX who is going to build the visual layer of your scorecard. Then you have data science team. , with, with the whole ai, we want to have automated insights within the scorecard. You have data science. If we don’t, if you’re not operating in a product world, what happens is that all of these individual teams will work in silos and all of these individual teams will go to the business in terms of, Hey, what are we trying to build?
And the business would say, don’t talk to me anymore because there are only 50 people who reached out to me on the same question. So in a product organization, what happens is that all of these teams work cohesively, but product is going to be the front facing for the business. So in my team, I have lead product managers, directors of product management.
They are the ones who sit with the business. And they are the ones who are primarily responsible for and asking about the problem, what are we trying to solve for? On the back end, they work with DE, our UX partners, our data scientists, and we build one holistic product for the business. So, I mean, to put it lightly it’s like they have one throat to choke.
End of the day, if something goes wrong, it’s the product that is responsible for it. It’s a product responsibility to deliver the needs of the business. So that’s how we are structured. I think the really cool piece, and I’m not sure, Evan, if you have seen this with other companies, is we also thought about product is fantastic, right.
But product takes longer time to develop. Meaning if you are building a AI based scorecard with automated insights, it might take us anywhere from a few weeks to a few months, right. But in the meantime, business still has, like, ongoing questions they need to answer. So what we said is, We don’t want to distract the product from what product is doing.
So why don’t we create another layer, which is our insights arm. So we have a, what we call as rapid insights. So these are the folks working with the business. If there’s a question on, hey, I saw sales dip last week, x, y, z. Can you help me answer this? That’s where that question goes to, right. So your product is still working on the product.
That he had insight function, like churning out those really needed analysis. So they are the ones who are keeping the lights on. So the, as a product manager at Lowe’s, what you really learn is you learn to navigate all of these different functions internally and also lead them and bring a single point of view for the business end of the day.
Evan Wimpey: Yeah, that sounds great. I think at less mature organizations, I think it’s more common to have sort of that analytic insight functionality. And the qualm is, well, we can’t develop products. We’re too busy answering the, the questions that are always come in. So we’re either going to just stop serving the business or we’re never going to be able to develop the products.
But I think that, that sounds like a great solution where you’ve got a product focus team. That’s not Having to handle the one-off ad hoc request all the time. You’ve got the rapid response
Sravan Vadigepalli: and think about the cross pollination of the talent right if you are somebody who has been in Rapid insight space you’re doing bunch of analytics But you are at a point where you know what I want to I want to change in terms of what I’m doing Can I go into product?
Yes, that is possible. Or the other way around, if you’re, if you’re being in product, you feel like it’s kind of slow or it’s taking a long time to see the outcomes, you want something tangible like every day, then you can get into analytics and you can think about insights, right. So I think that’s the beauty of it.
And we do have quite a good number of folks like making, making those shifts within the company. So I think that’s the beauty that I like and I’d love to see more of that.
Evan Wimpey: Yeah, that’s, that’s great. Talent management perspective. You’ve got, you’re able to cater to folks and their career growth. I want to ask about, you know, in the sort of the, the rapid insight it’s take things as they come in the product world.
How, how does something like a scorecard, an AI related scorecard that delivers insights. How does that get prior? Presumably there’s competing product ideas. How do you, how do you go about prioritizing and knowing what to focus on?
Sravan Vadigepalli: I think one thing I would, everyone appreciates at Lowe’s is we are, we are really radical about prioritization because we always have competing ideas and we’ll always have competing things that we want to deliver.
But how do we make sure we are focusing on the right things? It starts with. Number one, the yearly planning is what we call within Lowe’s, is we think about what is that business trying to do the next year. And business is going to put a document around, these are the top things I want to achieve, right. Let’s say marketing, as an example. Marketing would say, I want to drive traffic, I want to do X, Y, Z. And the first question that comes is, how is that rolling up into. The top five initiatives of the company as a whole, right. So that it happens at a senior leadership level. So they, they think about what are the big rocks for my business?
And how are those big rocks going to roll up to the company big rocks? That’s number one step. Then from there, we do have variety of frameworks. The, the, the, the two top ones is we use Rice Framework. I don’t know if you heard about Rice Framework before. It’s essentially every idea, what we do is, what is the reach of the idea?
Like how many customers are going to be benefited if we go and build this initiative? And the I would be impact. What is the impact that we are talking about, like, in terms of dollars? Like, how is this going to help the company generate the money? And C is confidence. Like, what is our level of confidence in terms of getting this idea?
Is this like a moonshot, or is this like a low hanging fruit? Or are we, like, where is the confidence level, right. And the last piece where we come into play is the effort. Like, how long is it going to take for the tech to deliver it? If we, if we think about essentially the, the whole framework provides you a clear structure.
And your reach, the business usually comes up with the reach, right. They are coming up with the business idea. So they’re going to tell you that, Here’s the reach we would expect to see from this idea or from this intranet. And then finance is going to come in and say, okay, here’s your impact. And here’s the confidence that we feel that you can achieve it.
And then tech would come in and say, here’s the effort. Essentially, that’s the framework we use. And let’s take scorecard as an example. And if we provide this framework, let’s say we want to build a scorecard for the digital business, the question then become, okay, how many users are going to access this scorecard?
If digital team is thousand, I’m making this number up, thousand people. Let’s say we expect at least 60 percent of them, 600 of them, use the scorecard every day. That’s, that’s one. That’s in reach. And the question then becomes, what is the impact? I mean, looking at the scorecard is not going to do anything for you.
Like, what are the actions the business is going to take out of it? And do we have them listed out? I’m going to do XYZ. Hey, once I know how many customers are going from my homepage to checkout, and once I know where is the leakage, I can do XYZ, right. And the finance is going to come in and say, okay, here is potentially what your impact is going to look like.
And comes the confidence piece of it. Like, what is your confidence level to deliver this? And finally, the effort, like I will come in to play from the data side of the house and the tech side of the house and say, this is going to take us three months, or let’s say, if I’m going to say, this is going to take us two years, then the question would become, do we really want to focus our efforts on it?
Because your effort is essentially going to dilute the impact, right. So that’s the beauty of the framework. So that’s how we use it today. And it is a framework. It is a lively conversation, to say the least. It is needed to have this conversation so that we can drive the right outcomes. We also use OKRs, objectives, and key results.
So everything that we build should have the objectives, and we need to have quantifiable key results. So I think, Lois, you’ll be amazed by the rigor that we put in to get there. It’s a planning process that we do, but we also believe that once we log the process we can confidently say that we can deliver XYZ.
Evan Wimpey: Thank you. That’s great. And I’ve never heard of the RICE framework before. I like that. I think implicitly a lot of teams try to do that, but I really like the way that’s, that’s encoded and captured there in a space like data engineering analytics products, there are a lot of people outside Lowe’s that are, are trying to sell you different products is that sort of your role in estimating the effort sort of this is an internal effort, or there’s potentially already products out there that can service this need.
Sravan Vadigepalli: I try to not open my LinkedIn messages.
Folks trying to sell always right. And I think the RICE framework is going to definitely help us with that. But the second piece is we always go back to. So when Marvin says that we either serve customers or serve those who serve the customers, right. That’s the, that’s the team of like how we operate.
We always need to think about for a given problem statement, is it a better idea for us to develop something internally or do we want to go external? I don’t think there is a hey, if you hit like one, two, three, we’re going to go internal or we’re going to go external. But most of the cases, what we look for is, we understand that we are a home improvement retailer first, and we are using tech to accelerate how we serve our customers.
That’s the main goal for us. We’re not necessarily a tech company. We are a home improvement retailer, but we have a great tech presence within home improvement, right. And we all know that. So, what we do is, for example, if somebody tells us, are you going to build your own LLM? Well, maybe, but that’s not the primary outcomes that we are trying to achieve.
Our number one goal is how do we help our customers? And building LLMs internally may not be the suitable solution today. Maybe in the future, but not today, right. So, that’s how we, we approach it. What is the strategic alignment for that specific problem statement? , we, we do a lot of hybrid. As well, where we bring the expertise and work with the internal teams and find a solution.
So we do that as well. So it’s not necessarily we either go by or bill it’s more of a hybrid approach of what makes sense for that given the problem.
Evan Wimpey: Sure. Yeah. And I feel like the hybrid approach often is, you know, you mentioned your career opportunities for the folks on your team. If they’re able to learn and grow and then end up owning or being able to, to develop a refined tools internally, after the fact they’re, you just walked us through a pretty robust process for prioritization. Let’s just, let’s just scrap all that for a second. And if instead they said, Oh, we don’t have time to do planning this year, Sravan, you get to just decide what it is that you want to work on. What it, what should we work on in the, the engineering and analytics product space and everybody’s aligned.
Evan Wimpey: Everybody’s thumbs up. Yep. Let’s do it. We’re aligned with Sravan’s vision.
Sravan Vadigepalli: In my utopian world where the money is not a problem, the resources are not a problem. I think the way let’s start with analytics as an example, what we don’t have today, Evan is we don’t have a tool that is going to explain the second order and third order consequences of a given action.
Think about any scorecard or any tool out there, be it Tableau, Dermo, or any other visualization tool. They all tell you in a really good way of what happened. Some of them will get you to why something happened, for example, if the sales were down last week the tool could help you. But, well, you’re Your traffic is down, so maybe that is impacting your sales, but you don’t see a tool in the market today that is going to get you to the level two, level three of the analysis that is being done by analysts today, right.
That’s the whole area where you’re, you’re using your analyst’s brainpower to dig into those like second order and third order consequences of a given event. In an ideal world, I would like to build a product. Which is going to help the teams to get to that nth level of information. I don’t think we have one today, but that’s where we should be.
And think about how much productivity it is going to give, both for the business and also for your analyst team. If analysts could focus less on trying to join multiple different data sets, but if we feed a system with, All the required data, and if the system spits out, hey, these are the potential correlations between your 20 or 30 different metrics.
I think that’s a really cool thing to work off of, right. You still need human in the loop, but you’re, you’re giving the human with so many tools that you’re making their life much easier and productive.
Evan Wimpey: I’m with you. I vote for that. And I think especially when you get into the, the idea of causal modeling, it becomes very difficult.
Sravan Vadigepalli: So if there’s a product that can help point to possible causal factors. Right? I mean, think about in any given problem statement, causality is such a big factor because essentially we are talking about counterfactuals. What could have happened if we didn’t do X, Y, Z? And for humans, we can only think in so many, there it is, but we don’t, we can’t get beyond like level one or level three.
So I think that’s where we can use the power of AI or machine learning in general to, to enable something like that.
Evan Wimpey: Very cool. Very, very exciting. A lot of exciting work that you guys are doing there. Thanks Sravan for walking us through your role in the team there at Lowe’s. It’s good to have someone representing a very analytically mature organization to hear sort of some, some best practices and hear what’s possible.
Sravan, if you’re interested, you can reach Sravan or you can find him on LinkedIn. He also writes on Substack and Medium. We’ll have links to all those in the show notes. Our guest today, Sravan Vadigepalli. Thanks so much for coming on the show, Sravan.
Sravan Vadigepalli: Thank you for having me, Evan. Great to have this conversation.
Explore the dynamic world of retail data analytics with Sravan Vadigepalli, Senior Director of Products and Engineering at Lowe’s. In this episode you’ll learn about the unique analytics challenges and opportunities in the home improvement space and how Lowe’s approaches product development.
Sravan Vadigepalli | Guest
Evan Wimpey | Host