Speaker 1: Hey there product lovers. Welcome to the Product Love Podcast, hosted by Eric Boduch co- founder and chief evangelist of Pendo and super fan of all things product. Product Love is the place for real insights into the world of crafting products. As Eric interviews founders, product leaders, venture capitalists, authors and more. So let's dive in now with today's product love podcast.

Eric: Welcome lovers of product. Today, I am here with Lars the co- founder and CEO at Dreamdata. Lars, why don't you kick off this podcast by giving us a little overview of your background?

Lars: Hey Eric, thanks for having me. Yeah, for sure I'll do that. So my name is Lars. I'm the co- founder and CEO of company called Dreamdata. And I come from a background in product management. Most recently had a company called Trustpilot. That's a review product in the eCommerce space. Before that I worked at other software companies. And at the very beginning of my career started out in UX in consulting and what you back then would call a web agency.

Eric: So tell me how working in user experience gave you that background for product. How did having that experience help you be a leader of product to Trustpilot and now a CEO?

Lars: Yeah, I think that back then at least it was, and it still is a typical path into product management that you can come to that UX angle. I would say our flavor of UX was... We were in an agency setting and we basically were the people. We had of course, engineers working on the product. We had sales people in, like we were an agency. So you had sales people engaging with the customer, but nobody was really taking that, I would say product management perspective, actually like thinking about the whole product. So we thought about the user experience, but we were also the people who would think about the whole product and think about would it actually work in the context of our customer. So I think it's not unusual that you take that path. If you are on the design flavor of UX, it's less normal but if you are on the business or concept side of UX, I think it makes sense you end up in something like product management.

Eric: So now running product to Trustpilot, take me through that. Take me through how you got into product at Trustpilot and what it was like to inaudible product there.

Lars: Yeah. So I did come from a product management job at that point. So I had done a bit of product management before that, but maybe without, I would say not having a very good definition myself of what product management actually was. But for me, Trustpilot was the first place where I got to work with a major product and feel what it meant to be responsible for a product. So coming into Trustpilot, I would say a lot of things that I think are right, that would come out of like ideas about how to lean products or talk a lot about value and how important it was to do small things that created value, think about value. But I think what was maybe missing, was a healthy understanding of how do you actually measure value, how do you know that you as the product manager is contributing value. And I think another thing was that when I joined Trustpilot, that was around series B, it was a fairly small team. So there wasn't a lot of product management there when I started, but as we grew the team there of course you end up having to deal with, how do you do product management across multiple teams, how do you manage something which from a customer perspective is just one product. How do you manage that, how do you get a bunch of people together who call themselves product managers and get them to manage it in a meaningful way together so that it takes a common direction. It all makes sense together, but for each of these people it also has to make sense. So I think there was a lot of that journey that happens in a product organization that grows. That was my experience with that. Yeah.

Eric: Yeah. Talk to me about what else you learned there.

Lars: So there was lots of painful learning. When I started, I think the great thing about Trustpilot was, there was a very, very awesome engineering team in place when I started. So it meant that there was a lot of things going on in the engineering side that were taking the company in a very modern direction of this is eight, nine years ago. So it was maybe not common to release all the time, but already at that state, that was a conscious decision by the engineering management that they wanted to be able to release very often to microservices, like carve up the product into small parts. So there was an engineering team that was wanting to do a lot of the things that you need to do, to do modern product manage management. So they were going through this process of slicing up the product, making it actually manageable for a product and to get some part of the product work. But I think that that was maybe the first painful experience was going through months, if not like almost a year of restructuring the technical side of the product, just trying to find small wins that we could implement while preparing this product for actually scaling the product team really, because to scale the product we need to scale the product team, but the technical infrastructure wasn't there yet. So there was a lot of re- architecting of the product first then once that was in place then came all the discussions about like, how do you actually divide up the product? How do you create meaningful units of product teams that can... What is a meaningful unit, what is a meaningful responsibility? So there was a lot of learning. I think one of the pivotal things that we did was we worked with Silicon Valley Product Group and got their help in how do you do modern product management? I think they were very key in our approach to this.

Eric: What was a meaningful unit there for you, what did it look like?

Lars: I think what we ended up... I think it varies across... This isn't true at all stages, but where we ended was, it's a unit of product where you can have ideally the full stack of the product. So you are able to deliver product to solving a specific use case or a specific scenario or for a specific persona. So you can work with a slice of the product that is all the way from concept down to the data layer. You can release it with very little friction from other teams and you know how to measure business value inside of that unit. So I think that was key thing, you need to find slices of product that in itself can create business value so that you can measure that. And I would say at Trustpilot, we never got to Nirvana, but my Nirvana is a place where every team has a clear, this is back to the discussion about value, is like where every team can see the business impact of the product. So ideally it's like it's revenue, it's retention. It's maybe it's in some cases it can be cost savings, but going back to the fundamental business objectives of the company, how is that team contributing to that and how do they measure that. That's the perfect world in my view. So that's a meaningful crosstalk.

Eric: You mentioned one other thing I wanted to dig into before we move on to some other stuff. And that was this re- architecture, because there's always a challenge as a company in general, when you're rebuilding or re-architecting a product to get to scale, then it's often taking away a ton of capability to innovate. Right?

Lars: Yeah.

Eric: So you're often perceived as you're not innovating even though you really are innovating a ton on the backend for the things that matter to customers scalability, performance, all that kind of stuff. And that's always a struggle to make that decision. Was it a hard decision to make that or was it a choice like you had to do it?

Lars: It was insanely painful. And I think it's something at all cost, if in any way you can avoid that stage, if you can avoid building up a huge volume of tech debt that you need to sort out, that's great. And I think after that experience, we did a handshake between me and head of engineering that this would never happen again, that we'd never get to a point where we would have to dedicate maybe 80% of resources to re- architecting and just killing tech debt. And I think the main thing that you do to avoid that is basically having an agreement that you continuously pay off your tech debt. So you don't have this huge buildup that all of a sudden means you have to stop. We had things got to very bad, what caused us to start saying," Hey, we need to do something about this," was we had the experience. We were dividing things up between backend and front end, and we were building a feature. The whole feature was built and now we just needed to put that thing into the user interface that would actually activate the feature. And that went through, I think a full month of attempts to release this thing, just because of the messiness of our code base. So at that point we said," Okay, this is not sustainable. We're going to be around for a long time. So we need to sort this out." Then we went through this whole process of carving up the product so that it never happened again.

Eric: And now you've gone on and co- founded Dreamdata. So talk to me about making the leap to starting a company.

Lars: Yeah. I think many founders, you start from something you experience. You have an experience, our experience was with Trustpilot and you face a problem and you look for solutions and you realize that there aren't really any great solutions for this problem. And then in your context, maybe you like, if it's a tech problem or a product problem, you start solving it. Maybe you're solving that context. You realize," Okay, hey, this was very painful. Other people have this problem. Maybe this is a product." That's what happened to us. Trustpilot was like a two- sided marketplace. You had people writing reviews and you had companies buying the attention of reviewers essentially, and buying other benefits of interacting with reviewers and reviews. And it meant that we had a huge inflow of business from this consumer facing part of the product. If you're a Tripadvisor you have the same, lots of people are seeing Tripadvisor because they are looking at hotels and restaurants et cetera. And then they end up interacting with Tripadvisor. That causes companies to interact with Tripadvisor and then those companies then buy products from Tripadvisor. This was exactly the same model. So we had this huge component of a business that was originating from a consumer facing say marketplace. And then we had because a marketing team that was spending a serious amount of money on traditional marketing efforts, like B2B marketing efforts. And we had a very active sales team that was doing outbound sales, but when we sat together in the leadership team and we were trying to figure out like," Okay, so this month we made X million dollars in new business where else did that actually originate." There was just no consensus around it. And when people added up their contribution, it inevitably ended up at like two to three times the actual amount of money we made that month, which is close. Can't be true.

Eric: That always happens. Right.

Lars: It happens a lot. And I think for us, it was not so much a question about finding the truth, but at least saying," Hey, if we are going to run this business in a rational way, we need to understand what works and what doesn't work and what are the ratios like how much of this should we do?" If we do more dimension, like B2B dimension, what's the result, how should we then act in sales, what about like, we have steady increase in people coming in from this marketplace. How does that affect the other things, because there was no consensus about how things actually worked it was very hard to talk about. And it sounds like our take on that was it's a data problem. So the data was actually there, but it's a data integration problem. And then on top of it, I'd say the most common way of describing it is like attribution, you want to know your revenue where does it originate. So that's what we built there. And then I think we felt that when we did it at Trustpilot, we had so much data and it was quite standardized. We were using segment. com for tracking. We had Salesforce CRM system and we had a hotspot as our magnet automation platform. It was super standardized, we were buying ads from Google. It's like this isn't special. It's very especially in B2B SaaS it's like 90% of people look like this. Of course sometimes they use a different CRM system. Sometimes they use a different magnet automation platform, but it's like the number of inaudible isn't that big. And it was weird that it was such a challenge to solve this because the data was there, so why wasn't anybody just building an easy sort of integrating this data, giving us the result. Why did we have to go through the pain of building data, like ETL data pipelines, modeling the data, and then we could do the attribution analysis, which was what we wanted to do. So that's how that product idea came about. And then basically the step is what it looks like is like you go out, you test your idea, you see like was it just us or are there other people, is it actually true that there are other people who have this problem and are willing to pay for it and gradually get sucked into it and then all of a sudden you're a company.

Eric: Yeah. So talk to me about you. The personal journey for yourself, from running product into being a CEO. What surprised you, what did you expect and where were your biggest personal challenges?

Lars: I think, because at Trustpilot, when we made this transition to being very value oriented and very revenue oriented, I was actually in leadership position. So I was never forced to do it. So when I started this company, of course when I was a consultant I did a bit of selling, but I'd actually never sold a software product. I'd never been out there pitching a product to customers. And I think that was a big challenge for me. We were pushing our product managers at Trustpilot to do it and say," Hey, if you want funding for your project here that you think is so great, go out and sell it, show us that you can actually bring in some money with this product before we start investing heavily in it. And we want some customers on it before we fully build it." And that was a big challenge for me, just like selling the product. And then I think the next thing is when you're the CEO, you're typically put in charge of fundraising. So you are the guy who's going out trying to get money for your company to, invest in product, invest and go to market, et cetera. And that whole game of venture capital was just, that was new to me. I'd never done it before, I wasn't close to the process in Trustpilot. So there is a lot of learning in that for me.

Eric: So what advice would you have to other heads of product thinking about leaving great companies to start something. What would you tell them?

Lars: I think there are of course practical stuff you need to think about, but I think most people can work it out. You need to think about money, et cetera, like your personal life. But if you are okay with that, then I think I would definitely advise people to try it. I don't see how you can become anything, but either a successful founder or a very, very good product manager, if you do this right. Even if you say you start a company and you fail, at whatever stage you fail, there is an insane amount of learning in this that is extremely relevant for product management. So some of the best people I hired for product, they had done startups at various levels of success. So I would definitely do it, but of course accept the risk of failure, but know that what you learn from that I think is probably you can't get those learnings any other way because yeah. I don't know. That's how I think about it.

Eric: Yeah, no. I can understand that. You're definitely going to learn things no matter where you struggle or where you succeed, that you wouldn't learn if you weren't exposed to things like sales and fundraising. Now Dreamdata, data obviously is very important. Right? It's your name. Talk to me about the importance of having a data oriented mindset at your company. Talk to me about the importance of data in general for companies. And then obviously you're dealing with a ton of data at your company.

Lars: Yeah, I think there are many sort of truisms around this, but we definitely subscribe to things like," Hey, don't bring your opinions, bring data." And we very much... That's how we ran product at Trustpilot. That's how we try to run our business. Of course, when you're early stage like at Trustpilot, if we wanted to do any change to the running part of the product, especially the consumer facing part, we would have insane amounts of data and you could do true AB tests and experiments and you would get very, very solid answers. When you're a small startup like we are, we have hundreds of leads whereas a company like Trustpilot have maybe tens of thousands of leads in a month. Right? Of course, it becomes different the way that the level to which can be data driven, but you can definitely commit yourself to using the data you have, working with it, taking it seriously, never making decisions without looking at the data you have. But sometimes we will be looking at something that is like,"Oh, 12 converted into two," because you know that, that might just be that month or that week, but you still have to look at it. Then you might have to apply a bit more of quality of analysis to what happened. Like what was that deal that didn't convert, was there a reason for it? So I think that for me, there's no way around being data driven and because the product we built and sell, it's very much, it's on the same premise, right? It's like this is our product targets marketers primarily and helps them to be data driven in their approach to the market. And our take there is the same. You want to take all the data you have, you want to put it somewhere, you want to apply analysis to it and you want... If you can be experimental and do true AB tests, of course you should do that. If you can't do that, then at least count what you have, look at what you have and make decisions based on that. And even if it might be that you don't have enough to do true statistical analysis or if you can't build beautiful machine learning models, you can still work with data.

Eric: So let's go back to Dreamdata a little bit and revenue attribution. Talk to me a little bit about the challenges of revenue attribution, because that's a big challenge for product managers, especially if they have either P& L responsibility or for some online responsibility or just see their products succeed, how do you help people with revenue applications?

Lars: So we at Dreamdata, we don't do anything specific for product, which is maybe weird because at Trustpilot we definitely saw the same challenge because, if you're in a beautiful world, your team has a part of the product and it's directly tied to sales. You built this new beautiful feature and has a price. But at Trustpilot that wasn't the case. We had six teams and they all had different parts of the product that was sold at one price. So how would they know if their part of the product was actually valuable? So with it pretty much the same thing that we do in Dreamdata around attribution is that we would have all the tracking data for the usage of the product. So we'd know for each account what are they actually doing in the product. And then on the other side, we had the business data of those accounts. So like what was the ACV of that account, had they renewed, were they a customer for several years, et cetera. And with that, we will build dashboards to help our product managers understand what was the value of their part of the product. There's a simplistic way of doing attribution. But I think if you're in that situation in your product, that's a reasonable way of doing it. There are many cases where you don't need to do that. Like if you are a large PTC company and you're building the shopping cart, it's probably pretty easy to see the impact of the change that you do to the user interface. Or if you introduce a new algorithm for proposing products, like add on products, you'll be able to see the impact. If you're a large company, you can see it very fast. I think again it varies, if you have B2B often there's like lower volumes of data. So then you probably need to work with less and be less sophisticated.

Eric: Now talk to me specifically about the marketing side of the house and attribution for marketing people. I'm interested.

Lars: Yeah. So you can say in many ways, all marketers that are doing digital marketing, they work with attribution all the time. If you're looking at Google analytics, you're looking at maybe your spend, got your traffic coming in and you're looking at conversions, there's some level of attribution that are like a funnel. I would say any funnel is an implied idea of attribution. So marketers in general work with that, but where they are struggling is with B2B specific. In B2B the big struggle is actually to connect what I just described there is like behaviors on a website or responding to a campaign, connecting that with the actual business result, like generated pipeline, close to one pipeline, renewal also, et cetera. That's very hard because those systems such as, like they create separate data silos and they don't... If you're a great company, of course you have built a solution for this. You build a super nice data warehouse. You have the pipeline built for all this stuff. But in reality, most companies, they stop there. It's too complicated. They never get to build it. And then marketers are lost. They look at revenue over here in sales, they celebrate revenue and marketers are building stuff over here and they disconnect and they can't see what... One thing is, I think maybe marketers feel underappreciated. That's not good. But I would say also they are flying a bit in the dark. They're lacking instruments to tell them okay, of all the things I'm doing that I feel are good, which of those things are actually good. And if I have to stack, which ones are better than other ones and also how do they interact? So I think marketers very much need attribution in that sense.

Eric: Yeah. No, absolutely. Having a background in product and marketing, you deal with this issue of multi- touch and B2B too, right?

Lars: Yeah.

Eric: Which of these touch points were actually really valuable and how do you calculate that? And then you're dealing with also companies where you have multiple contacts. Which ones of those that were required from what mechanisms actually helped move the deal along, so to speak. And how do you calculate that. There's a lot of tough problems there.

Lars: For our customers that we're targeting solely B2B companies and primarily SaaS companies and sometimes some other things as a service, but those are our target customers and they all have what you just described. They're selling the multiple stakeholders in an account. The deal might take 60 days just in the sales pipeline. But before that it might be several attempts to sell. You can have years of history before that. And how do you make sense of all that? And how do you even get it somewhere where you can start to make sense of it. It's super tough.

Eric: Yeah. Yeah, absolutely.

Lars: So we got plenty of work.

Eric: So what's the ideal customer from Dreamdata, what do they look like, what problems are they trying to solve, how do help them?

Lars: I think, of course the customers we love the most are those customers who are looking to solve this problem, who are, I would say ideally they already have a healthy collaboration. They go to market team between sales marketing, maybe product. So it's less of like" Hey, I need to prove that I'm doing well." It's more of like, everybody agrees that we're or a data driven company, but to get this data somewhere where we can action it, it's just going to take us so long. We won't be able to build our own product. We're going to have to pull our best engineers out to do this. So," Hey, we're going to buy this product instead." And then we can start making those decisions based on data. I think that's a very, very good customer for us. It's someone who is very data driven, but also understands that building a solution for this problem is going to take too much time and say they would have to sacrifice a lot of time with the best engineers to solve it. That, and then of course, it's someone who is spending a lot of money in the go to market. So we like, if you are people who are growing aggressively, that fits very well because then the investment you're doing in to go to market is so big. You need to figure out in small changes or optimizations can help a lot. Also companies that are searching for that channels, or they know that channels, they know how to generate revenue, but they're forced to continuously increase their growth because they're building up their revenue base, they need to grow, grow, grow. So they are constantly searching for new channels or new tactics that can continue their growth. And how are they going to work out whenever they test something, how are they going to know if it works or they don't have this type of product, that's a very good customer. And then I'd say we do, because the data foundation itself is very exciting. If the company also wants to use that data for other things, that's exciting for us. If you want to use a reverse ETL tool to put some of this data back into your CRM system or put it into email automation platform, that's also very exciting, it's like both because our product becomes more valuable to the customer. And of course it also becomes more for us, as it becomes more valuable, our customers will stay along, stay with us longer.

Eric: Absolutely. So talk to me about the biggest challenges now. What's your big challenge on Dreamdata side, beyond the," Hey, we need to close and sell more customers," but what's the biggest challenge that you need to deliver from a product standpoint or probably you need to solve for customers?

Lars: Yeah. I think it's a bit... I think because we are sitting in this attribution space, that's an interesting place to be, but we feel that the product we have underneath it is very powerful. So it's like, how do we make that step from say, say that we are solving B2B revenue attribution to say, maybe we are the foundation for B2B revenue automation, or how do we make that transition? I think we are making it customer by customer because as we're selling to a customer, they like the attribution part of the product. We give them access to the data part of the product. We show them the value of that. But now we're like, how do we concision maybe our go to market and should we? I think it's not a given that we should, but that is a big challenge I feel. Beyond fundraising and selling and all the other things that are part of running a business apart from a product perspective. I think that, that's a big challenge for us.

Eric: Well, as we're wrapping up here, I thought I'd ask you a couple of questions Lars. So I've been asking you questions as we've gone, but let's talk about you personally, as opposed to your professional background, what's your favorite product?

Lars: So yeah, I mentioned reverse ETL as something we like to work with companies that do that. So my favorite product at the moment is a product called high touch, which is a reverse ETL tool. I love it. I mean, do you know the product?

Eric: I do not.

Lars: No. Okay. But reverse ETL, it's basically like you have your data warehouse with all your beautiful data models, like what we build and then you use a reverse ETL tool to extract data from that and push it back to say Salesforce, HubSpot or any other system where that data originated. And that's super powerful. We use it for a bunch of automations on our own data for ourselves. And it's just... You built very, very useful solutions, very, very fast. And you wouldn't want to... so high touch is my favorite product at the moment.

Eric: So what makes you like high touch in particular?

Lars: I think for me, it is the amount of business value that you get from a very, a little effort there. We use our own product. So we have... All our customers and prospects are in our own product. And we have all the data around how they use the product, how they interact with the website, et cetera. With high touch for instance, I can take that data and I can give it back to my sales people and say," Hey look, this account that you engaged with two weeks ago actually is now engaging with the product. Maybe you should contact them." And it's just something like if I wanted to do that without high touch, I couldn't. I'd have to pull out people to build complicated stuff with high touch, I can do a bit of sequel, click a few buttons and then it's running.

Eric: Cool.

Lars: So it's like the amount of business value I get from a very little effort.

Eric: Well, kudos to high touch. So one final question for you Lars, three words to describe yourself.

Lars: Yeah. So I think doubting is maybe a good word or inquisitive. I think as a CEO, some CEOs have a very high degree of conviction. I have actually two co- founders who have very high conviction. So as a CEO, I'm probably a bit more on the doubting side or I'm the person who asks the questions and not always sure that we are going in the right direction. I think data driven or data obsessed is maybe also a good way of describing myself because I do feel that, that I am, and we are as a company, like insanely committed to data. And then honesty, I think I try to be honest. I try to say things the way I see them, even if it's not always favorable to myself, I try to put things the way I see them even if it might mean that other people are not, they completely think that, that would be the most clever thing to say in that context.

Eric: Awesome. Well, thanks Lars. This has been a good time.

Lars: Yeah, thanks. I really enjoyed it.