Contact Us menu-bars menu-close

Behind reporting dashboards

Fail Faster

Episode 435


26 minutes

Join us on the Fail Faster podcast as we sit down with Jag Bedi, Vice President – Data Science and Analytics at Purchasing Power.

In this insightful episode, we delve into the fascinating world of e-tail and how Purchasing Power empowers customers to improve their quality of life through affordable access to products and services. Jag shares the challenges and successes of implementing a data warehouse and the impact it has had on decision-making and customer engagement. Discover how data-driven strategies are shaping the future of this innovative company and get valuable insights into driving change within your organization. Don’t miss this inspiring conversation on leveraging data for growth!

Podcast transcript

Vandana: Hey Jag, welcome to the Fail Faster podcast. How are you today? 

Jag: Pretty good. Thanks for having me, Vandana. 

Vandana: Absolutely. I would love to have a little bit of a peek in your background Jag for our audiences if you want to touch upon, you know, anything that has been influential for you. You can start as far back as when you were a little kid. And then come to today, please. Sure. 

Jag: So I guess a little tidbit, I guess, IT was not my first choice. Well, I wouldn’t say that. I wouldn’t say it wasn’t my first choice. It wasn’t my parents first choice. Of course, being from a South Asian background, you know, it’s every parent’s dream to be have a doctor in the family. I tried that. My passion was always in computers. I love taking them apart, putting them back together again. And so I remember my dad’s favorite line was, hey, you’re never gonna make any money in computers. And so every time I see him today, it’s, I told you so. So fast forward to a while. 

My background is heavy development, Java, .NET. I was in those wars when that was happening. I got into databases, tuning, querying, all that great stuff. Fast forward to today, and where I work, we had an opportunity to advance our data ecosystem at purchasing power. And my CTO and I kind of got together and said, hey, would you be interested in kind of leading that initiative? And it’s, you know, for me, it was a nice change than just from heads down coding, you know, meeting a team of coders. So here we are. 

Vandana: Awesome. Well, congratulations. And yeah, it looks like you’ve been there nine years. 

Jag: Yep. It’s hard to believe. Yeah, nine years this last Friday, I think. Yeah. Well, great. Very good. 

Vandana: So can we get into some of the successes? Like what was whatever some of the early indicators that yes, this is what you were meant to do, and some contributions that you’ve been able to make at the organization? 

Jag: So I think, you know, we’ve, my boss, or the CTO, and I mean, we’ve, we’ve had some, I would call very colorful discussions, heated as well. I think we’ve always been in the camp that, you know, data drives pretty much everything. And so with purchasing power being owned by a private equity, you know, we were, were obligated to deliver results. And so, you know, it was, it was always on his mind and my mind that, hey, we need to bring purchasing power into a data mindset, right? 

And so I think we got the chance about two years ago, and, you know, we got approval to kind of move forward. And so, you know, he kind of asked me initially to take a look at it, go over it, and stuff like that. And then I think that’s one of the successes was at least acknowledging the fact that, hey, we, the Excel reports are not working inside of a team, you know, where everybody brings in their Excel reports, and everybody raises their hand that no, we should use this one, you know, so that was one of the key indicators that, hey, we need to have, you know, we kind of need to come together on this. 

Vandana: So what were some of the steps that you took to bring along the whole team? Because if you know, you have people who are just at they have, they have that habit, and they just have that pattern, they’re comfortable with the uncomfortable, right? So how to come along the ride to be more smart about data and create more funding? 

Jag: Well, I will tell you, it’s, it wasn’t easy. And I think we, we, we still have some of those challenges today, which we’ll probably, you know, get into, but it definitely was not easy, you know, because everybody loved their Excel and their SAS and nobody wanted to change. And so I think it’s, it was kind of, you know, saying, hey, this is change management, here it is. And this is what we’re going to do. And it’s going to be better. 

I don’t think it was up for a democratic vote of what I wanted to do. So which can backfire in certain situations, but in this one, it didn’t. So we actually presented it to the SLT first, got all the senior leaders on board. And then we just kind of moved down to the VPs and the senior directors, and just kind of, you know, moving down and presenting, here’s what, you know, we’re going to do not to the tune of here’s what I want to do. 

But here’s how it’s going to help nursing power, you know, kind of get into that next phase of growth, right. And data is going to help us get there. And so I think once we spent a good bit of amount of time, like almost about four months, kind of doing a roadshow, you know, about here’s, here’s how we’re going to help. And I think it was an easier conversation to have. So definitely spending that upfront time was what was definitely a good decision, you know, my opinion. 

Vandana: Awesome. And can you share some of the hiccups along the way, as you were going through this, you know, long term change management, what were some of the things that came about? 

Jag: Yep. So after we had the discussion, and everybody was, you know, gung ho, high fiving, that, hey, we’re gonna do this, I started to sit down with individual business units, you know, starting with marketing, supply chain, finance, the ecom teams, you know, I started to ask questions around the data warehouse, because to me, that was the first step aggregating all of our data into a single source of truth, where the entire organization would feed from that they’re reporting their, whatever that they needed to do, and hopefully move away from Excel. 

What I quickly found was that when I would ask some questions around, hey, what does your data look like? What do you need it to do? And what would you like to see in the warehouse? It was, I got a whole bunch of Excel sheets, hey, we have the shared folder of, you know, 1200 Excel sheets. I want to see all of that, you know, inside the warehouse, you know, talk to the next group, same type of answers. So one of those, you know, was quickly becoming that, hey, I want to see everything. 

You know, whatever I see today, I want to see everything. So I started to explain to them what a data swamp was, versus a data warehouse, right? A swamp would be pretty much, you see everything, nothing changes, you still get to see your reports. So I think having some of those tough conversations around, you’re not going to see this because you don’t need to see it, right? 

Yes, putting some of those controls in place, it was, you know, was definitely a kind of a back and forth among those team members. But I think, you know, once we started to kind of explain, hey, you know, this is going to help, but you have to aggregate your data, you don’t want to see what you see today, you know, creating multiple data sets for the same amount of reports that everybody else is creating. So you would have multiple reports with different data sets, we’re trying to do multiple reports with one data set. So all of you are on the same page. 

Vandana: And, and how is that going for you now? Like, I’m sure you’re far into this now. So do you go back to people like really think about the old days with the spreadsheets and everything? And how do they feel about it today? 

Jag: I try not to, it’s pretty traumatic. But, but yeah, you know, we do reminisce every now and then, you know, I am, you know, kind of proud to see like him. And when we have these, you know, big off site meetings, you know, the SLT is pulling, you know, reports from our existing, like the Tableau, you know, that that that we put up, and they’re, they’re thinking about a data mindset, right. And now it seems to be more in conversations along the lines of, no, this is right, because I pulled this directly from the warehouse, or I pulled this directly from, you know, Tableau. 

So these are the correct numbers, you know, so I don’t see too much of that now. But it’s, it’s kind of nice to see when the entire organization is getting behind an initiative, you know, that is going to drive into the into the next phase of growth. 

Vandana: Awesome, awesome. What a great, you know, achievement. And, and I’m sure that you people are trusting you more right now, you don’t have to convince them anymore. Yeah. 

Jag: Yeah, yeah, the I’m hoping that the convincing is kind of behind us. I do get now questions, you know, around how’s data gonna help me do this? Right. So and that’s a complete mind shift of where we were just two and a half years ago. So I mean, it’s a huge kudos to the entire organization kind of getting behind it. Yeah, you know, there’s always going to be bumps. 

And, you know, some, some folks that are just going to be adamant that this is not going to be the way but I think for the most part, the entire organization, you know, is behind us. I think it also helped that the SLT pushed in as well. I think that was a good hindsight decision to, you know, have a hierarchical selling at each particular, you know, level of the organization. 

Vandana: Awesome. Would you give us a little bit about purchasing power? Like what is your differentiator as a company? And how all this feeds into, you know, the bigger goal of the organization? 

Jag: Sure. So purchasing power is a retailer with products that we sell, just like Amazon does. But we are a closed loop retailer to, you know, clients that we contract with. So think of it as a employee purchase program for companies that we contract with to help customers who, who can’t, who are kind of, who have good jobs, you know, may not have great credit, but they have a steady job. And the cliche of your job is your credit kind of applies. 

But, you know, to them, we’re the lifeline to helping them get products that they would not normally be able to afford, you know, and in cases of like, you know, your washer dryer breaks down, you know, most of us can probably go out tomorrow, charge it, you know, get a wash, dry the next day, and then pay it off in 30 days. These set of customers can’t do that. 

And the differentiator that we have is, you know, with us, it shows up on our credit card, we pay it off in 30 days, we allow our customers to pay over time, over a 12 month or 18 month period with no interest, no fees, nothing but the price divided by, you know, the amount of months or their pay period. 

Vandana: Awesome, awesome. And so all of this stuff that you are doing with, with the reporting and all the data, you know, what can I say, compilation and proper hygiene of the data? What are some of the big goals that now as a company, you are looking forward to as as you get into the more growth phases of your organization? 

Jag: Yeah, so for us, setting up the data warehouse was obviously priority number one. With that, we set up a data engineering team to help us feed. You know, I think what we also did, you know, at the same time was also set up a data analytics team, and a data science team. So we didn’t wait, you know, for phases to say, hey, now we should go do analytics. Hey, now we should go, you know, do data science. 

The upfront effort and all of that was, it was pretty big. You had to set up an entire, you know, those three teams all at the same time. And then all you have them work all in parallel to a data warehouse that was barely four months old. Right. So I think that was huge. The you know, the next big steps, obviously, are machine learning, AI, all those buzzwords. 

We’re actually already doing, you know, machine learning off of the data, you know, that we have, we’re, we’re constantly finding ways to, you know, re-engage or engage with our customers who either are dormant or are new, repeat buyers, you know, all of that great stuff. We have machine learning models that help us create touchpoints with customers, engage with them in different ways. 

The next thing is obviously, how do we leverage some of these new, I guess, new technologies like chat GPT, you know, LLMs, and stuff like that. So we’re exploring that we’re, we’re always very cautious, you know, just jumping into something very new. But yeah, we, you know, we got a couple of POCs, you know, that we’re running to see how we can at least internalize some of those features. 

Vandana: And are you moving towards personalizations too for your customers? 

Jag: Yep, we have a, you know, whole suite around recommendations on how we, you know, which products are targeted to which customers and a whole personalization suite. We actually, you know, use a tool from SAP called SAP Commerce, which has, you know, held up very well to the times. It’s allowed us to customize it, you know, to our specific business need. 

Vandana: Awesome. What in terms of the trends that you’re noticing with with your customers, like, what are some of the things that you are now able to cater to? And the changes that because the customers are very dynamic, right? And what are some of the things that you’re able to do with them as a follower of data? 

Jag: Right. So, you know, definitely with our customers, they may, you know, even though they’re in a, you know, specific, you know, non-serviced group, you know, by some of the, you know, main retailers out there, just because of, you know, their financial situation, we provide that missing link. 

Like, for instance, I will, you know, give an example like travel. Most of us can book tickets and buy and pay off and there’s no issue. We’ve brought travel to, you know, our customer base, allowing them to pay over time for vacations or for attractions theme parks. So that was one, you know, net of product suites that we integrated with a third party to help bring that, you know, piece over, you know, to brushing power customers. We’re always looking for innovative ways and, you know, to kind of help our customers, you know, get the same type of product.

So they don’t like, Hey, yeah, we even have, you know, tires. Hopefully, we’ll start to add more and more product lines. So it’s not just a e-commerce website, we just buy products and move on. We’re actually helping our customers, you know, be also kind of help them toward financial wellness as well. 

Vandana: Awesome. And this is only products or also services? 

Jag: Products and services. We also offer educational products. Obviously, it’s it’s a it’s a need based, right? So it’s, you know, most of our customers tend to buy go to electronics, you know, they may have kids in school going to, you know, off to college or so laptops are a big seller of electronics categories. It’s a huge seller. 

Vandana: Awesome, awesome. So you’re basically empowering everybody out there to be able to elevate their lifestyle in a way, a lot of different ways. And then for you to be able to cater to new needs and everything, you need to have all the vendor relationships at the back end, right? So there’s like two this. So what are some of the things that you’re seeing in the in the partnerships that you’re bringing on board and the scaling of that and that dynamics? 

Jag: So, you know, definitely you brought the vendor relationships. I mean, those are those are obviously very key. You know, one thing is we are dropship. So we don’t own any of our inventory. So those vendor relationships become even more critical. And so, you know, all of our vendors, obviously, you know, the teams at Pershing Power work really, really hard to to kind of establish those relationships. 

You know, we’re seeing that, you know, they’re kind of they’re kind of looking toward us. I mean, they’re they’re saying, hey, you know, how much of the new iPhone 19, you know, would you like, you know, this Christmas? So we’re in a position, you know, with our customer base to be able to dictate, you know, some of those terms and say, hey, we want you to allocate X number of units for us, only us. And so we’re seeing the vendors, you know, cater to that because they know that the volume is going to be there for our customers. 

Vandana: Awesome. So with all of the work that you’ve done, and the teams that you’ve set up, I, I think, I love the point that you made that you were able to, together, like not incrementally, but you kind of revved up the engines together for all these three data analytics science team and the warehouse teams. Is that something that people are missing to leverage most from their data science initiatives? Or are there other things that people are also missing? 

Jag: Yeah, I think some of it is probably committing to it with kind of both feet in. And I think that’s what we did. You know, there, it was on the table that, hey, why don’t you set up the engineering teams first, get the data warehouse sourced, and then move to analytics and data science or whatever that order may be. And I remember having these conversations. And now, ago, you know, my boss and I would talk and I’m like, we’re already in it. So might as well just go and stand them up and see. 

And it was kind of interesting, because at the same time, you know, my boss was working on establishing an offshore center as well for the first time, you know, first empowered to be able to expand the capabilities of those teams and, and, you know, in an offshore center. So all of that was going on, you know, while we were trying to do, you know, the data science and the analytics piece, but, you know, at that, I think, I think communication, you know, was probably key, which we indicated every step of the way, every single project that we did on the data side, it was, you know, it was pretty much everyone at SLT, to all the way down to the individual contributors, everyone knew what we were doing, this was not a, you know, we’re gonna go do it off in silo, and then do a big, huge unveiling and say, hey, now we have a data warehouse, right. 

So we kind of brought everyone along for the ride, for better or for worse. Now, everybody was there, everybody kind of knew. And I think in larger organizations, that may not always be possible and completely understand that. But I think that the communication aspect is like, I think kind of gets lost. I think if you don’t put a specific scope around what we’re trying to do, we had a specific scope around what we wanted to achieve with the data warehouse. 

First, it was financial reporting, you know, we wanted to be able to report off of financial numbers and to get it from a single source, you have to provide to our board. Right. I mean, and it just kind of went from there, give me the next use case, give me the next use case. Those use cases continue to feed into the data warehouse. 

And then that data fed into the analytics world inside Tableau. And then from there, the data science team would now start to answer the questions of why something was, you know, kind of got into that whole prescriptive descriptive type of, you know, analytics. I mean, we always knew something was something happened after it happened. Right. 

Now we’re able to kind of say, hey, all right, we’re seeing some trends here. How do we get in front of it? So now we’re kind of into that predictive type of, you know, analytics, where we’re saying, because we’re behaving this way, what’s what’s going on? Can we get in front of it? You know, type of actions. 

Vandana: Awesome. And what are some of the things that you can tell us about the reporting structures, right? A big part of what you’re doing is output of it is the reporting, right? Now, there is talk about dynamic reporting dashboards, dynamic dashboards, and things like that. What are your thoughts on that? And some insights that people should think about before delving into the next shiny outputs there? 

Jag: Yeah, I think, you know, for us, you know, my personal thought is, you know, like, I don’t like to chase after the shiny new object, you know, like, it’s kind of like, you never buy a car the first year it’s released, because there’s there’s bugs already in the backlog that they have to fix. Right. And so I think, you know, that that’s one thing that I know, it’s tempting, you know, like, especially around generative AI, they made it so easy. And so and it’s not, it’s tempting. 

I mean, I’m tempted, like, but I think one of the key things would be to kind of stay the course, you know, if you will, I really believe in the crawl, walk, run theory. Even now, I’ll tell you today that we’re probably at the end of crawl. In the beginning of walk, I don’t believe ourselves to be in that run. Because, you know, to me running would be that you’re, you’ve got the ml ops piece talent, you’ve got autonomous decisions that the website is making on your behalf. 

I have models sitting there that are constantly evaluating our customers as they shop, right. So I’m sure, you know, the SLT might disagree with me a little bit, they might think we’re on the walk. But there’s a lot that can happen, you know, and for us, I think it’s we are getting there. You know, we are definitely getting there. 

Vandana: Awesome. Anything else, Jack, that I haven’t asked that you might want to share as a as a leader in this space with folks who are trying to achieve results like you? 

Jag: Yeah, you know, it’s kind of interesting, because, you know, I’ve talked to a few folks who, who are trying to start this journey. And it’s just like, we don’t know where to start, right? Like, how do we how do we get, you know, just to get a data warehouse, right? And I always go back to like, hey, we, you know, at Parts and Power, we have a strict, you know, set of initiatives we do every year to provide, you know, ROI to provide revenue or some sort of a return, you know, on what we’re doing. 

I always recommend, hey, you know, for me communication, like, you have to communicate up. And sometimes, you know, you have to be able to kind of stand your ground, you know, like, start off with a use case, you know, it’s like, hey, if I had this, if we had a data warehouse, I guarantee you, we won’t be seeing the problems, you know, that we’re seeing today, I would be able to drive, you know, with Tableau reporting with the dynamic reporting that that you mentioned, I mean, we don’t have real time reporting just yet, to me real time is probably less than every couple minutes. 

But we have 15 minutes that, that is pretty close to real time that you know, our cards pretty much update every 15 minutes, and everything’s coming out of the warehouse. So it’s interesting that so yeah, I mean, it’s levels, you have to sell it at the top, because if the top is not sold, no one else is going to buy it at the at the bottom. 

And we spent a good amount of time, you know, talking to the SLT members all the way from the CEO to, you know, chief operations, revenue officers, all the way down, you know, not just a technology officer, you know, because he sold already, or she sold already, they want to know, CEOs want to know how this is going to help me versus the chief of operations and like, how’s this going to help me, you know, so those conversations we did have, and it does take time. It does, it doesn’t happen overnight. 

Vandana: Yeah, it does not happen. Well, thank you so much, Jag. This was this was a very informative conversation. And I really appreciate you coming on. And I wish you all the best on whatever you’re doing next. Looks like there’s lots to happen still. And through the walk and the run phases, and I wish you get there very, very soon. 

Jag: Right. Thank you so much for your time. I enjoyed it completely. 

Vandana: Thank you. Same here. 

Get updates. Sign up for our newsletter.


Let's explore how we can create WOW for you!