Hi, everyone. Welcome to More Than Numbers live, data show focused on turning data teams from support functions into engines of growth. I'm Ollie Hughes. I'm one of the cofounders of count dot co. We're a Canvas based BI tool trying to move people away from the dashboard factories into their organization's problem solvers. This week, I'm really pleased to welcome Jimmy Mogrip from Podimo. He is the global, director of insights and analytics at Podimo, and this was gonna be amazing opportunity for us to hear how, a data team, a data leader actually focuses solely on driving insights which drive their organization forward and doesn't deal with any dashboards or at least we're gonna find out if he does or not. Jimmy, thank you so much for being part of the show. Welcome. Thank you so much, Ollie. Pleasure to be here. Say that Ponomo in itself is actually a subscription based podcast service for audiobooks and podcasts. You have over, I believe, over a million active users. And, so this is probably your first or many podcasts you do? Tell me. No. Actually, this is this is the first time I'm actually recording in our very own small podcast, dude. We have a bigger one as well. So it is a little bit daunting for me, but, I'm I'm up for the challenge. You said he looked apart. I gotta say I love the fact you've got, like, a a potty mode kind of a a microphone muffler as well. It's absolutely amazing to see. So you definitely look very much in the right place at the right time. So, we're gonna dive in. We're gonna work with you on, like, what your team does, how your team operates. I think it's one of the, you have, in my view, one of the best roles in data that you're doing the thing that everyone coming into data wants to do, which is to use data to drive insight, which moves the business forwards and solving problems. And I think you have that, in spades. We're gonna dig into that more as part of the show. But before we get there, my opening question to you is tell me, the data tool that you fell in love with, the data tool that you first realized the data was the thing. Tell me, have you got a story we can get into to help to help us understand your background a bit? Yeah. That that is a very fun question because, what I should mention is I never planned to go into data science or analytics or anything from the get go. I have studied behavioral sciences, understanding why people behave the way they do, and then I added a master in communication to that as well, which was what I started with, at one point in time as an actual communicator, at a bank, a small bank in my hometown of Helsingborg. And what actually happened, this was more the startup vibe. So one day, my manager at that point, we were every one of the more senior people. They were doing multiple things at the time. So, yeah, he was, like, salesperson, communicator, and head of IT in one go, basically. He just taps me on my shoulder once. You don't have anything to do today. I have lots of things to do, but you're my manager, and I'm a junior. So I can't say nothing. No. I'm fine. What do you need me to do? Oh, we have purchased this tool. It's called iMarketer. It's an analytics tool, from a British company called Experian, and we're having a full day training of it. And I think, you should probably be in it. That was okay. It was from nine AM to, five PM. And what actually happened as I went into that meeting, I went into that meeting as a person who wanted to be a communicator. I left it as a person who was dead short that he wanted to work with analytics. Because how it actually operated was it was like a GUI of based upon an SQL base. So as a bank, you had transactions. You had accounts, and then you had users. And the way that you can drag and then see, I was sitting there, like, this is, like, real data of behaviors that you can actually use, and it was so, so cool. It was so intuitive, in a way it would build up. And I just said, like, this is what I wanna do. All those years at uni for nothing. Now I'm gonna do something else. So I've got to say I had that opportunity as well. So I mean, it feels like you've needed everything to go really well because you did behavioral science and communication. You now work in data for a communication based audio podcasting company looking at user behavior. So, I mean, like, it feels like you've got the perfect triumph from all the things you love. I I do think so. I do think so. It has, yeah, Sometimes things the stars are just aligned in the best possible way. I don't know what I can say. I've been I've I had a lot of luck. I one of the things I love, we're gonna dive into more exactly how you what you do and how you love it in a bit more detail. But before we dive into exactly what you do and what your team does, can we just orient people to what what it's like to work at Podimo, a bit more about the kind of data stack? Just give us that context of the organization maybe. Yeah. So from the data stack perspective, Podimo is a very Google heavy company, which is something I prefer as well. So, like, all the bay all the ground laying data is obviously stored in the BigQuery, environment. On top of that, we have the Looker, to build the self-service part of, of data. I think self-service is more than just dashboards, in its way. We have DBT, and then what my team is mostly using is, Python or R, to, like, do the more static visualizations and running through the data and all of those kind of things. So Right. But it is a very, very Google heavy company, Polymer. There you go. That's really good. And so, I mean, in fact, using r there you go. You're really adding some stats right there. There you go. Like, someone who preferentially chooses r. Yeah. You you get the yeah. For in my team, like, it's not like we say, do you have to use Python or you have to use R. It's basically about finding the right person if one wants to use R and one wants to use Python. It's no big deal. But most people are using Python. We have a few things running in r because r obviously has more statistical features in one kind or if you have really been into the academic version of statistics, r is the get go. And you usually stick around with what you enjoy with, so it makes sense in that way. That's cool. That's That's definitely, gonna appeal to some people who are stuck using Tableau when they prefer to use something else. I can tell you. I can see a lot of people out there jealous of the fact you let them pick their tool. I, we're gonna get into more about your team particularly, but just also just tell us how the data organization is structured, like how the you're one of a few different teams running data basically at Podimo. So tell us how that works, and we can then go into more your world. Yeah. So data data at Podimo is based on three different legs or three different divisions. We have, So the entire data team is twenty twenty plus people, spread evenly across the the three branches that we have. First branch is AI, who are mainly working as, like, implementing AI solutions into our app, like recommendations, algorithms, and all of those kind of things. Then we have a BI team, which is responsible for, like, the whole tooling phase, what tools we should use, from the BIA angle, how the data is being processed, making sure that we have the correct dashboards, in place. So for and they're also driving the self-service, out out of the company, making sure that Yeah. Yeah. Whenever you have the specific number you get, you can get it quite instantly. And then the third part is obviously my team, insights and analytics. And we are obviously focused solely on looking at the data and trying to derive as many insights and recommendations, as possible we can do. And we support the entire company from product, content, growth, and the local markets by doing deep dives, making sure that we have for each area the right type of, like, metrics that are both influenceable, but also that, they are connected to our overall company goals as such. And then we cross pollinate each other within these, various disciplines like growth content marketing and making sure that we actually make, like, more company wide recommendations as well about what we should do as a company. And this this is this is the this is what I'm I'm linking to you on this episode. Right? How what does it look like when you're you're basically living and dying by the ability to drive insights to the organization rather than just providing the data and providing the the kind of the the operational reporting? And this is, it's a it's an amazing thing. I'm I'm not sure there's it's not a typical thing to see this split out. Most people who are doing BI, maybe do analytics as well. There's a separate function doing this. Tell us how that before we dig into that nuts and bolts, tell us how the team was formed. Was it a thing that you that that became obvious was needed separately? Was it a spark of an idea of yours or the founders to, like, say, no. We need this as part of the way we do data to make it a standalone thing just focused on value. I I think, actually, it started already when when I was hired because, my my manager here, who is our chief data officer, is very much a person who, like, listens to people, what they think. So already in the recruitment process, I was quite clear that my profile was more in the more deep deep side research of things and, giving recommendations. So, obviously, there, there was a match from the beginning. And then at that time, it was a little bit more, how you should say, loose. It was a startup. So, yeah, we we can get you in. You have a great profile. You have some good experience for that we are interested in. So let's let's see what we can do here and see if it's a match. And it turned out to be a match. So I think I I always was granted that that type of freedom to really see what I wanted to do and how my role should behave. And then we just started adding people, and then people really saw, like, oh, this this is this is really bringing value to the company. Now we are in a in a different position, like, where is feedback. Like, oh, we need more recommendations. It's not like we need more dashboards. That obviously is a thing that but it's not being asked by my team as more. Okay. We need more insights. We need more actual deep dive from it, which I think is very, very cool, and I'm feeling extremely happy to be in this position that I have at. Yeah. I mean, it's it's great the organization knew this was the value of data from the start that it's it like, that inside generation is the kind of is what's gonna move business forward. Like, reporting I mean, as you as you know, we talk about operational clarity a lot. Getting a consistent clear view of performance is essential. So there is a need for BI, hundred percent, but it's then how you take that and drive insights from it. So, I mean, there's so much more to talk to you about here. One of I just wanted to just the simplest thing is, like, how do you go about doing it? What's an example or a means to actually deliver insights that actually change the game? Like, what is that workflow? How does that there's a lot a lot of insurance obvious to some people what they even looks like as a workflow from end to end. No. I I think I think, like, a lot a lot of the things, a lot actually, I I I can't really quantify it, but a lot of project that we are doing and the things that we're looking into is not an actual request from someone. It's something that we have seen somewhere or something we, as a team, think is missing that this is probably something we should look more deeper into, and that's how it starts. And then we'll see. Sometimes, you you have to be a little bit wary to not go down the full explorative rabbit hole, which you can where you don't have any insights. Yeah. Yeah. Sure. Like us steering, what we should look into. But then, obviously, we have, like, the company goals. So we have, like, company themes, which are obviously phrased in the more, like, business manner we should grow with this percentage or whatever, then it becomes like, okay. If we take that actual strat strategic decision made by the leadership team, what are the data points that we can probably dig out, and what are the suggestions we can come up with to actually support that? Yeah. Yeah. Which, yeah, is, is, also and we also have, like, an ongoing discussion with the lead team always where we, like, both share what we're working on, also share our insights to making sure that they actually happen and become action act actions in the company as well. And like the CEO, we we meet him once a once a month with a larger data team to discuss, our planning and just keeping an eye on we're actually doing the right type of stuff as well. But with that being said, there is always the explorative freedom in this on our type of call. And I guess that freedom is the is the fun. Right? I mean, we all as I say, I always say that we all get into data all like you, follow the data by accident because we love the fact we can solve problems. And the fact that you can do that is obviously the main benefit. I'm assuming suddenly you've got quite a lot of freedom to do to find things that then you can put forward as suggestions. Yeah. In your problem. Totally say that. Yes. And then but then what so, yeah, go on. Yeah. No. Go go ahead. Go ahead. I was just gonna ask, like, that's the benefit, the the the freedom that you have, particularly not been given requests. But what's the negative side of this? What what where do you have to make sure that you're delivering? What's the what, I mean, what else do you have to make sure that is that make it fully rosy? There obviously are things which make the job difficult for you. Sorry. We can probably cut that up. Now but, like, obvious obviously, a challenge in all this is there there are things and we should, first of all, mention that we do get requests as well. Can you look into what happened to this so we can do be be do this better the next time? Like Yeah. Sure. You're doing an experiment. How can we actually then improve on the next iteration? What are the suggested steps you you can see in the data? Maybe not just from the results from the experiments, but also from an additional deep dive, etcetera, to make it look even better. So that is obviously the direct task. On the more open ended questions, that we have with the team, I think that that the main challenge, as I was alluding to earlier, is this do not go into the explorative rabbit hole. Sometimes you have to say, okay. This was fun digging into this data and going through it, but there is nothing actionable we can make of it. Or it can be that the actions that we see do not have the impact that we needed to have or something like that. You you you will fall into that. And, obviously, we need to prove our worth based on the recommendations we give. So that is obviously a trade off, like spending spending enough but not too much time on each task is is a is a real challenge, but I think that goes for all analytics in general. That's true. Yeah. Exactly. The the the temp as as we discussed many times in this podcast, the temptation just answer the question the business asks means that you at least know they they giving them some way, but you're also capping your limit the better value you can bring because if they aren't asking good questions, you're not gonna give them good value. The flip side is if you're given flexibly go get value, you've gotta go deliver it, and you live and die by the the the, the recommendations you make. I guess I I guess this is a thing where you've developed trust over time. You mentioned you meet the CEO every every month. You obviously work closely with different parts of the each department. I'm assuming that there's you're plugged in basically very closely. I guess, is that something which has grown over time, and how have you delivered trust, or increased it? That that that's a that's a very good question, actually. Obviously, I I meet, like, with senior senior stakeholders from the lead team on a weekly basis, to always make sure that I am on top of things and also that we can provide things, in in the product org, for instance, as well as other parts of the company. The peep the more senior people in my team, they also function as embed. So if you think about, like, a product squad, as such, there is, like, one product manager. There is one lead designer. There is also one of my senior product analysts working collaborative, like setting the road map and stuff like that. So they always feel like feel like they're close to, what we're doing. We always provide as much visibility as we can about what we're working with. And either you get a thumbs up and a thumbs down, which I think is great. But I think it's important in these cases to, like, share your ideas, what you're working on, and the initial results as early, as possible. That is how I would say the best way of actually gaining trust. Totally. So you're you're giving them a chart. You're you're baking it to You're making it a clarity problem solve rather than just providing things without any real context. That actually reminds me a lot. We actually did our interview with Connor from Clio, which if you're watching, you can go back to count dot co slash m t n to find this, where he talks in a very similar way of embedded analysts into product squads as a way of maximizing the value and blurring the lines between the the data around the actual the actual business context and making that a wonderful way to work. So check that out as well if you're watching to get more on that kind of model and how that can look. I I think that makes total sense though. Like, the way to just leaning in, being shoulder to shoulder with the business, helping them, thinking, having complete over understanding of what's really going on. So you're in informing your hypotheses, informing your thinking as you're in the guts of an analysis is just the best way to do it. Yeah. And you mentioned as well. I I like, this is the idea we want to look into. And sometimes already in the ideation phase, like, people can say, okay. That's probably not the best thing. We we probably believe we should look into this, more or something else, or you should try to fudge it a little bit. Okay. Makes sense. So there's always the ongoing discussion. Totally. I mean, it's it's back to the idea that if you, you it's a to basically get the business to think through the best questions, you wanna have the data people and the business contacts together at the start of the process as much as you possibly can. So you just haven't got that disconnect which leads to inefficiency, leads to missed work, and missed, like, suboptimal time spent on the wrong problems, ultimately. I think that's just a great just a generally great principle in general to be as close as you can to the decision. So the data and context are not separate silos, but actually are one thing. One thing I wanna pick up with you, you mentioned right at the start about how you your team also helps the business understand the metrics. You're not just about finding ideas. You're also saying, like, you're you're shaping the metrics that matter and mapping that. You you go into a bit more detail about what that means and how you do that and how that then feeds back into the the AI and into the the the sort of the the BI team as well, how you work with them? So we are when we are setting up metrics, when we when we do that, we obviously have a company goal, whatever that that might be. Like, for instance, we need to grow the paying subscriber base. And then we need to focus together with, like, a product manager to understand, okay, the part of the product we're currently working on in this squad, how can we actually influence that? And that becomes a little bit of a data science y, exercise to actually understand which is the best metric for actually influencing the overall business objectives, that we have. And then, obviously, you can think of it as a metric tree. Like, you have one NorthStar metric, and then you have supporting metrics. But if they move in the right direction, we can always say that we are having a positive impact on the overall business goals, whatever they they might be at that point in time, which I think is important also when you run experiments, etcetera, because then you can actually also combine that specific increase of a North Star metric to actual business value. So you can actually do that translation, as well. Yeah. You'll see. As you can imagine, we are big fans of MetaTrace account. So it's you're speaking to the converted there. And I I think also, like, it's just the you say just the allows that simplicity of of connection to outcome, which is just really powerful. It's also just a really amazing problem solving mechanism. Right? A metric tree as a way to break a low level is also just a great way to explain what you're talking about. That MECE, kind of mutually exclusive, collective, exhaustive way of breaking a problem down is just another great mechanism for this. Yeah. You should always when I have a complex problem, like, in Sweden, we have a saying that you shouldn't eat the elephant all at once. You should eat it in small pieces, and that that's basically what it is about, like, understanding the small not so balls that you can actually influence. That's absolutely right. I mean, the so back to that, which I think is a really important talk about skill set. I'm assuming that that kind of breaking things down is a kind of fundamental way that you and your team work. Just that little critical thinking, that logical problem solving, breaking the problem down. To to people who are listening who are like, I wanna do this. This is I wanna do more of this. This is why I wanted to get into Dave from the first place. I'm still stuck doing dashboarding, and I wanna help build my team maybe more into that direction and help coach to get there. Like, what are the fun what are the skills that that differ for your team that maybe the BI team or elsewhere that you've been? Like, what what are the skills that you're hiring for maybe that, like, really stand out as important? No. But but but I think, like, when I'm in, recruiting processes, I think I pay a lot of weight onto the more soft skills. Like, because if you think about, first of all, it's obviously about understanding the problem. Like, if someone gives you a business problem, how can you actually translate that into an analytical question that you can then work with the data that we have available to actually solve. So be able to show that you have that mindset of actually translating that into actual analytical question is key one. But then there is also the whole thing about, like, communication and, like, being able to, like, prove prove a point with, what everyone talks about these days is data storytelling. Like, being able to actually being produce a whole story with data is very, very powerful if you wanna move, something and make sure how everything fits together. So I think those soft skills are way more important than being, like, very, very good at the specific Python package or whatever. And I think that's also why I don't care if my team is using R or Python or something else, because what actually matters is the outcome, and the outcome is more shape of your soft skill in my team than they're actually hard data coding skills. With that being said, you always the basic understanding of statistic and mathematical methods, but the soft skills are more important than hard skills in my team, I'll be saying. I I think and that's just only gonna be a very, very, resilient way to hire anyway in the world the way the world is going where the ability to write SQL or Python is getting abstracted away. So statistics and then the ability to communicate is always gonna be the skill which is gonna matter more and more and more over the next few years and definitely in a decade. That makes that makes total sense. We're running out of time, but this is but this has been this has been an amazing I wanna get to a few other questions we had. So the first thing what I wanted to ask you is just what advice do you have to people getting to data or early in their career that that you think would have helped you as you got started? Yeah. It it's actually an advice I got a few years in my career, where someone told me, Jimmy, remember that a great data scientist never says, I found it. He always say, that's odd. Because what it actually means is that sometimes you get this result, and it's the best result you could wish for. It shows so much impact wherever it is. But it's so much better to go back, take on your thinking hat, and say, okay. Does this result really, really make sense? Do not jump into conclusions. Try to just dig a little bit deeper in the data or take a step back to actually see if the result that you're given actually actually makes sense. Maybe there was a problem with the sampling or anything. I think early on in the career, it's very, very easy to striving to get to those moments, in a way. I love that. I I think that's just playing back what I see you said it's the idea that you don't want anything to be surprising. The art of storytelling is the art of making something seem so blatantly obvious rather than being surprised where there's ambiguity. Like, it may be powerful, but actually, you kinda want the person you were receiving information to have to make it seem like the most obvious thing in the world, and then they can take an action because it's the obvious thing to do. So you've gotta have that back work done, that storytelling told. That's cool. I love that. As you know, on us on our, our on our podcast, we get a community, our community of data leaders. If you go to Cal dot co slash MTN, you can find out how to be part of our community, a part of this kind of community of data leaders trying to change, value from data. I've got a question from someone. Jimmy, I've done this in advance. I mean, about an hour ago, to be fair. From someone I want to know from you that when we're speaking to them, they want to know this. They said consumer datasets are fascinating in how they show human behavior. For example, the famous Spotify controversial campaign where they literally put really odd human behavior up on the walls and advertising campaign. How tempting is it just to explore? It sounds like it's quite tempting in what you said earlier, that you can go down the rabbit hole too much. Yeah. Fact thankfully, there is no way of, like, actually, in our data sets, pinning down who the individual is, which is, fantastic. Obviously, I I I would say, like, you will always, when you work with, like, behavioral data, find edge cases. You will always find interesting behavior or that a few or a couple of people have in a way that can come across to you in a, like, unsupervised clustering method or whatever. And but somehow, I think, like, it's important to also look at it and this and see, like, is this just something that is going on that you can disregard, or is this actually a behavior of users who have figured something out that they are after? So, I think you need to, like, look at it from both angles because it is tempting, but it can also give you some ideas, to do it. But, also, do not go into the explorer to rabbit hole. I've gotta say, and I used to work at Tesco, which obviously in itself, own way, has an amazing data set of human behavior. People buying certain things at certain times, certain demographics. It was very easy to find really cool things like there's a real spike of people buying bananas on a every third Thursday of the month from this particular shop. But the question I guess there's a there's a kind of a algorithmic. We make sure we have bananas, make sure the algorithm could pick up that spike repeatedly. But in reality, what you're gonna do with that information, it's interesting, but, you've gotta make sure you can operationally deliver against that insight effectively. And I'm not sure if telling the store manager that was gonna be much to change behavior in the operation. So I I get the I get that. It's got there's a balance to what you can do to make something weird and what's what's an oddity. Yeah. It's also what makes it fun. It's a they're always fun datasets to work with humans because humans are fascinating, much more interesting than widgets of a factory line. So it's a cool place to be. I'm I said before, I'm very jealous of the role you have. I think you've got an amazing role, and it's a role which I think everyone in data wants should be doing more of. It's the role we all wanted want to do. It's just how can you help make sure you're delivering the operational clarity, and then you can start a problem solve in the right areas. I think it's a really exciting job. So thanks for talking us through it. Thanks for being able to share about your world, share the pros and the cons, talk through the the skills that matter. It's been a really great time. Yeah. No. Thanks for having me, Wally. It was a pleasure.