Episode No:85

From Trustpilot to Dream Data: A CTO's Journey in b2b SaaS

Ole Dallerup

CTO & Co-founder, Dreamdata

Listen & Subscribe


Ep#85: From Trustpilot to Dream
Data: A CTO’s Journey in b2b SaaS Ft. Ole Dallerup
Ep#85: From Trustpilot to Dream Data: A CTO’s Journey in b2b SaaS Ft. Ole Dallerup
  • Ep#85: From Trustpilot to Dream Data: A CTO’s Journey in b2b SaaS Ft. Ole Dallerup

Episode Summary

In this engaging episode of the Hyperengage podcast, your hosts Adil Saleh and Taylor Kenerson sit down with Ole Dallerup, the CTO and Co-founder of Dreamdata. Ole has a fascinating journey, starting as a software engineer at Trustpilot and ultimately rising to the position of CTO within just eight years. Now, as the co-founder of a B2B revenue attribution platform, he shares insights on the challenges of building an MVP, the importance of early customer feedback, the quest for product-market fit, and the delicate art of balancing short-term delivery with long-term technical vision in his role as a CTO. Tune in for this insightful conversation and much more!
  1. Prioritize shared vision and team dynamics over just the initial idea when joining an early-stage startup.
  2. Validate your Minimum Viable Product (MVP) by creating prototypes and securing early customer commitment before investing in a more extensive solution.
  3. Strike a balance between having a clear product vision and roadmap while remaining flexible to iterate rapidly based on valuable customer feedback.
  4. As a startup CTO, carefully manage the trade-off between immediate feature delivery and technical debt versus long-term architecture and infrastructure planning.
  5. Leverage data and analytics to enhance the customer journey, from the moment they sign up, through onboarding, and into retention.
Key Takeaways Time
Ole joined Trustpilot early on because of his connection with the founder and other people there. Startups have a different energy and type of people compared to big enterprises. 1:16
To figure out what features users find valuable, start by building a basic prototype and getting feedback from potential customers. See if they would actually pay for it before building it out fully. 2:38
As CTO, Ole takes responsibility for driving revenue through product and technology, not just building the tech. He is focused on building valuable and sellable products. 5:00
It’s important to balance delivering features now while also building infrastructure for the future. You have to be able to scale eventually but can’t over-engineer too early. 10:15
Always focus on shipping the core value of a feature first, even if the technology behind it is a bit shaky at first. Prove value fast and get feedback. 18:17
Dream Data intentionally didn’t build much AI initially and focused more on getting data pipelines ready. This allows them to now easily leverage new advanced AI models like DALL-E to provide value. 29:20

Empower Your GTM Strategy

Monthly expert advice and top GTM insights in your inbox.


Adil Saleh 0:03 Hey, greetings everybody, this is Adil from having The Hyperengage podcast I have my co host Taylor Kenerson a very generic, slightly different product and the guests we have, we are now running in more seat heels because these technical conversations go way beyond than any of the conversations that we're going to be talking about tech stack, you know, associated with our guests, their product ID, non technical teams, how they're scaling operations, where they're sitting at as a business in a b2b space. So thank you very much, Ali, for joining. Ali is the CTO and co founder at bring data, it is a b2b revenue attribution platform more for the revenue teams more for the data centric teams that are basically investing so much in pure revenue, data, like revenue ops, data, integrations, all of these we'll learn more about, what what all these things are Dream, Dream data, thank you very much all the one more time for for taking this on your schedule. Ole Dallerup 1:04 It's a pleasure. Thanks for having me. Looking forward to the talk. Whoo, Taylor Kenerson 1:09 let's dive into it, Ollie, before we bring it to the current moment, how you started is really unique. So I would love to take it back a little bit. And what inspired you to join the team at Trustpilot, after you advise Apogee now, a Google company? I mean, you started, as I believe it was employee number 10. Or you were one of the very few that were a part of the company. So what even inspired you to get involved in a startup at such an early stage? Ole Dallerup 1:44 Yeah, I mean, I think it was humans. So at this time, the story was that one of my friends, very good friend I studied with, worked with, he joined Trustpilot. And he knew I was looking for something else. And he said, Hey, I think you should come over here and talk with the founder. That's Peter. And I think this is interesting for you. And I was like, Ah, no, I think I need to join these kinds of enterprises and make some real money. And then I went over and had a sandwich with PETA, and went back, and resigned and started. And I said to my girlfriend at that time, but wife now that hey, I think I know what I'm going to do. Right? In my life. I'm gonna do startups. Yeah, so I don't know, I think that it's people, it's humans, there's very different types of people in startups than what we see in enterprise. Adil Saleh 2:45 Yes, absolutely. You know, that. I mean, that makes it, that's, that's gonna make it so real. When you're so connected with people that you've, you've known long enough. And, you know, it's not just about idea. It's not just about the vision, it's about people and your connection with those with those people. And we get to hear a lot of these stories from a startup that the strong bond with the founding teams, or they have collapsed, when when it comes to relationship and everything got like me at a point when they shut down, there's so many taps on that just got shut down, just off of this reason that they will not so much, you know, look like? So it depends a lot, what kind of people you choose to work with? Adil Saleh 3:30 absolutely. So I mean, I'm so interested to know, only your background at, let's say, Trustpilot. It's as a VP and then you know, VP of engineering your turned out as a CTO, how do you see it as different? Of course, CTO is just like, not a business owner. But you are still taking all the high level decisions on the technical side, that's fine. But how do you see it different than building your own product co founding a product, such as spring data? How did that transition? Take into things? Ole Dallerup 4:02 So I was very early employee at Trustpilot. So I had some of the co founder pains as well. But I think the main difference is that when we missed sales targets, across pilot, I didn't sleep bad. Now that is kind of hurting a little bit more and kind of something that has to kind of hit as well. I think also that was the journey I took to Trustpilot. I went from being the technical maybe leader and more moving into actually being the business leader and taking responsibility and maybe my expertise is technology and product. But But I think I see it more or less. I'm a technology leader that is responsible for driving revenue with with technology and product So I do kind of take revenue responsibility. Yeah. And at least if not everyone at first pilot would maybe recognise that, but at least I'm pretty firm that they want to dream, they will recognise that. Adil Saleh 5:15 Hmm, yes, yes, because you spend like seven long years there. And as you mentioned that you were since making all the high level technical decision, but you're closer to driving sales, because in a b2b space in this day and age, it's your product, you your product has to be bigger than your marketing. Ole Dallerup 5:34 Yes, exactly. And it's not like I'm driving sales, I don't think that's kind of what I do either here or before, but, but it's more that I take responsibility of making revenue. So for example, we talk with customers, and prospects all the time in tech and product to understand them. So make sure those valuable products we measure and understand that what a feature is used, which features can be charged for which features are valuable, at kind of multiple level, talk with sales. Of course, I will also help kind of close a deal and have the part of the sales conversation. But it's not so much that part I'm talking about, it's more kind of taking the responsibility of making sure that our product is valuable. And sellable. Adil Saleh 6:28 Cool, really interesting. And talking about like, in the early days back in 2018, when you started with green data, I'm sure you know, a lot of these founders, like many of them, it's newer, been been their CTO as Trustpilot, for seven years now, that kind of thing. But a lot of these co founders, they are so much overwhelmed with so many things that they don't know. So what kind of challenges you had, first off getting the foot in the door, when you started back in 2018. In terms of getting a course on the marketing side, there was a different challenge, what kind of marketing you need to tap in, that you already might know, but how to position the product, how to finalise the product features, like what was your process, like you made a minimum viable product handed over to some of the folks that you know that you know, and trust whether your past relationship? How did that or workflow and the beginning started with a lot of these startups would get a huge help? Ole Dallerup 7:18 Yeah, so I think it's the hard part is, is figuring out what is valuable. And so it's very easy to come up with a lot of features. That sounds awesome. And it's very easy to convince yourself that these are amazing. Some of them might be more likely most of them are not. This is a guy called Mark Kagan, that kind of learners that Marty Kagan is the guy who also look Google and eBay, these kinds of things, amongst others. But so it's figuring that out. And so then how do you figure that out? Well, you it's a little bit like sales, actually. But but the first part is you build a crappy prototype. Like this could be presentation. This could be mock ups, this could be something half assed that works. And the bit that depends a little bit on the situation. But but it's usually not something that works completely. And it's definitely not something that is in a good state when we talk about technology and software engineers. And then you go knocking on people's doors. Hey, do you want to talk a little bit with me? Do you have this type of problem? Try to talk to me about, like, how you're solving this problem today? Would you like to solve this problem? How big a pain is it? Hey, I have this prototype. I would like to show you if you had that. Would that help you? Right? And probably some of the first prototypes, you showed that people say, Yeah, not really. Right. And then hopefully, at some point, you have a prototype where people start saying, yes, that seems like something. And then you start saying, Well, if I built this, would you pay for it? And that's the real important question. Right? And if they say yes, then you would kind of want to commit the so what if we do could be a non legal binding letter, but an intent to buy? For some amount, let's say $10,000. Right? And then if there, but that's not how I'm ended? Well, that's an indication that just tried to be friendly. That's of course good, but not what you're looking for. Right? You're looking for those people who are actually willing to commit. And to do that, and when you start finding a few customers, that has some of the same problems so that you can solve without having to build too much. That's where you start having something you can go out and build. And I think that's what yeah, sorry. Taylor Kenerson 10:12 That's a really interesting point that you brought up. I feel like so many founders, especially in the early days, they're looking for that hack or that tip. And it seems like through every single conversation, so many just start off the same way, it's iterating. It's putting something together in its most basic form, and just getting it out there. And the more reps you get through those conversations, and the more feedback, then you begin to refine what you have. And it's usually never the, the initial thing you go out to, you know, the people you're getting feedback is never really the thing that you actually wind up building. So that's a really interesting and important gift for all founders out there looking to create their own thing to recognise that there is really no easy way. And it is this nitty gritty, messy, middle and interesting, you know, journey and initiation. Ole Dallerup 11:01 Exactly, exactly. And then the other side of things is the most of shitake. And kind of things that some investors would care about, is that you build, like one thing is you start building something you can sell now, right? That's what I'm talking about. And that's important to start getting feedback and start building up revenue and customers. But you need to also have a long term vision, where are you moving, how you're gonna do that. And that's kind of where thread like strategy technology, and also the market plays in. And this is why it's important that it's actually the CTO that does this. And doesn't have to be the only person involved, of course, but the CTO has to be super central to this. Because we have to understand what what are we trying to achieve for who posts short term like very Microsoft term, and then long term, right, even though we don't know where we're heading long term, and in the early years, this changes, daily, almost right. But later, you have to build that up. That's your kind of job to make sure you build the technology to be ready at the right time. And particularly as early stage we are, we can't do too much. So we can over engineer, I think many, most engineers at least have a period in their life where they end up over engineering. That's bad in this stage. It is, Adil Saleh 12:36 it is because as you mentioned that having a clear roadmap, you may have pivots on weekly basis monthly, but that's fine. But you need to set down a direction of your product where you're heading, as soon as you get your first initial customers, you need to make sure Okay, these are our customers that we need to build these kinds of features, most probably on a very high level, this is the direction we're gonna be taking. And a lot of these only in this year, a lot of pivots, big platforms have made pivots due to the AI. That is, I would say, once in decades, kind of, you know, transition that you know, tech space has, in my lifetime, I'm 28, nine years, it is the biggest transition that we faced. But again, as a CTO, building a product, having a domain knowledge, you can definitely have a ballpark high level roadmap of your products. You know, ever since you get like the customers mutual customers, what do you think about this? Ole Dallerup 13:30 So definitely, so you have to think ahead, I think and try to solve the right to I mean, it's the hard part, I find that most new kinds of people who move into this, what I see, they often find very hard as this, you have to deliver now, while you build up infrastructure for the future. And you have to balance this out all the time. And CTO you responsible for both, and you are accountable for both at the same time. But no one will praise you for building infrastructure today for the future. Right. But you have to do that. Right. And if you get I maybe at least many of the listeners will probably have experienced that, that hey, we are working in a start up and that's a lot of technical depth. Well, that's a sign that disbalance wasn't correct, right. It's not like we all have technical depth. So so like that happens. But you have to balance this out constantly. And that's art. That's it's painful. But you have to do this. Absolutely. Adil Saleh 14:52 I mean, I cannot tell you how bad I wanted to hear this and I want to because we also a b2b tech stealth mode startup. I have CTO there was talking to me last night, there's something that we can, we may overlook right now, but it's going to be a bigger problem years later. So we need to, we need to either ship it now versus we need to just wait on Build it a scalable model on the data side data engineering side pipeline, data pipeline, we also a data analytics platform, going to launch in a few weeks. So I mean, I absolutely agree with this. Because, you know, I'm not an engineer, but I've seen him, you know, banging his head around, around these kinds of problems that are, you know, creating some complications in the beginning that man, right. Ole Dallerup 15:40 On the other side, you have to also kind of pull back as an engineer, most of us like to have this clean architecture and, but you have to move fast here. So like, speed is everything. And I, I actually like, like Facebook's old, move fast and break things. So at this stage, this is super, super important. Like, it's okay to break things. And it's very, very, so one of the things I actually had to do this at Trustpilot, a few times where we were shipping fast, and probably breaking things. And then sales got upset. Right? And I had to stand a little bit in the middle here and say no, like, because we can't beat up engineers for that. Because then they can't move fast. Right? Of course, we have to respect the customers and the sales process, and so on. So we have to find the groundwork for them to work as well. But we can't beat up engineers for this. And this happens in more startups than I like, is that and then they get scared, right? And then the test, the over engineer, the over test, the suggest to hire cute MOQ ways instead of actually hiring software engineers, and then you end up not delivering value. Hmm. Yes, it's also such a Taylor Kenerson 17:06 delicate balance to have, you know, as a start up, and definitely, you know, launching something for the first time that you have to ship fast, but then it's also the mentality of you want it to be just good enough. But then sometimes you get caught up in that just good enough is perfect, and then you're too deep in the trenches, and you're trying to pull back. So what would be your advice, I'll lead to, you know, people going off on their own journey, starting a company on how to balance that how to manage those priorities, and those you know, different tasks that you know, need to get done. But when is the question Ole Dallerup 17:48 this part, and so, it but you need to be a little bit into the dirt, you need to be be very clear on where's the value lies. So what is the valuable piece of this feature we're delivering. And then you need to have a focus within the team of software engineers, product managers, designers, that you always ship the value first, like, always get the value out and be okay that there's a little bit of shaky technology behind it. But get the value out. This enables us to show that we move fast to customers, this is exceptionally important early stage. But it also enables us to prove that it's valuable. Very often we have these ideas, and we have some proof, right? We need to prove it valuable. What if it's not. And that happens, and this happens definitely for me, we ship something, it's not valuable. Well, if we wasted as little time as possible, killing it is not as painful. But if we did all the work to make it a perfect feature, and it's just them killing it is very painful, right? And we can't do that. So we need to get that value out and get that feedback. Real Estate, we also have that lock, that scale is actually not a problem, right? Because we don't have scale. We have few customers. Maybe no customers, right? So scale is a problem for later. But we can still architecture are set up so that we can maybe decouple certain services and so so that, okay, we have something over here. We know that doesn't scale, but we'll handhold it for the first few customers until we prove it's valuable. And then we'll we address this problem. Sometimes we will be outspoken about it. So we'll tell the entire team saying hey, we can get three customers on this. When we hit three, we have to go back and then we can't sell this anymore. Right so I've done that many times. So that's one way. Yes. Sometimes you just know you can trust pilot, I have to rule that if this is not a bigger problem, then I can just not sleep for a couple of days, then that's kind of. Okay. So we did that for a while. Adil Saleh 20:21 Yeah, I mean, we were doing it for 30 customers for 30 beta users. So then we'll have to do some very on the back end and just manage on the data pipelines, all of that this is what I got from the CTO, and which I think will help you out to validate that it's, it's the right decision right now, Ole Dallerup 20:41 that's a good way of often doing it. Right, you get some real data in, you get to see if you are on the right track, you get some feedback from customers, while 30 is I assume that a large amount and you can actually manage it by hand, maybe there's some processes, and engineer every morning gets up and press a button, or whatever. To sync it up. That's fine, right? Adil Saleh 21:08 Yeah, absolutely. So now we have like, pretty much up on time, like five, six more minutes. If you have a hard sub tools, let us know, we can squeeze another five minutes since we are going to be talking about dream data, your post signup journey, I see you have a premium plan, I'm sure you're driving your tech stack more heavily towards, you know, converting these free to, you know, the pro pen and then teams plan, or sorry, business plan and enterprise. So how does the tech stick look like a dream data? While you're someone that has made kind of decision on incorporating the technologies for the post signup, and you know, how you're ensuring the success of these accounts, these paid accounts, what they are converted to pay the con, like, how does the onboarding look like just give us a brief on that a bit on the technology. So once you mentioned technology or audience and even ourselves will know, okay, this is how they leveraging, let's say, this technology to ensure these kinds of use cases around customer success or postings, journey, all of that. Ole Dallerup 22:08 So so we are very much about data. So the green data is, is a contribution platform, but when General go to market analytics platform for b2b companies. So we will we gather a lot of data about the customers journey. And so we use that actually a lot ourselves to onboard customers to do analytics to target customers, and so on. So that when we talk about onboarding, we get people in as a self service sign up. And then we start communicating to them in various ways. So we, we send out emails to educate them about the product, we try to motivate them to onboard themselves. So that's kind of the, you could say the product. And we'll continue doing that, like in various ways, depending on how they act within the product. While we do that, we also kind of sent the signups to sales, and sales are able to look at them and classify them. Today, we do that mostly manually, but I think it's very clear in the near future, we'll kind of use the large language models to look up different information about these companies, we have also a lot of investment data from our own product, come in, and then kind of use that to make sure they do that well or maybe even automate that part of the process. And then sales might start reaching out to them. They might also wait until the customer says hey, I'd like to upgrade or the until they onboard a little bit more. And then sales start having a conversation with them. The inner, I think relatively straightforward, normal process. So that's kind of how we do all the inbounds. Adil Saleh 24:06 So, when you when you talk about you know economic events like you know, renewals and all of those those kind of data metrics, and they are so much associated with the product usage of customer. So how you're ensuring as a data company, data centric company that you are team, your customer facing team mean the sales support whatever is absolutely on top of the product uses product activities of the customers to ensure that they are more likely to retain and more that span they are going to churn out is that the process? Ole Dallerup 24:42 So our sales team use our own platform a lot. So inside our own product, you can log in and see an entire view of the customer journey up to the stage they are so they have an understanding of that giant kind of activity data from website tracking and and data, and so on. So So, so kind of sales use that customer success, can also use that and Mitel to kind of look at the data down, they're usually a little bit more interested into certain feature usages of various kinds. And so we have similar tools for that, that's fought back customer success can go in and see, Hey, okay, this customer, which features are they using? how active are they? Who was active from the account, and so on. So they make sure to reach out to the right people. And also understand how they're using the product. Yeah. Adil Saleh 25:38 Cool. Cool. So you mentioned customers that what kind of platform are you using? You can absolutely name it, the Ole Dallerup 25:46 dream that is built on on pre cruise, all the data is in BigQuery. And so we will have all the data always available in BigQuery. The dashboard here was building what's called Lucas studio now. Adil Saleh 25:59 Okay, yes. Ole Dallerup 26:01 But it could be any BI tool out there. So that's where we have it. It's, yeah, I mean, because the product kind of delivers this product delivers this out of the box. Within, then it's pretty easy to map it out in a local studio. Adil Saleh 26:22 Yeah, you are your own customer. So and then you on top. On top, you're building custom objects using luchar. And all these integrations. That's cool. So now, one last thing, like looking at your customers, how many customers do you have on a recurring like annual recurring basis? I'm not talking with customer, just users, how many users do you have, annually, that you guys serve? What kind of let's say, data metrics are data points that you're serving per month, around these kinds of things. So I'm just trying to see how Ole Dallerup 26:57 we are b2b. So so we have around 1000 customers now, and it will vary how many users are in there? Per customers, some have hundreds in permanence, and some have a few. That kind of varies a little bit on size of business. And so the volume kind of we go through, I think it the volume is mostly kind of data volumes, and how often we process? Yes, so we process many kind of million tracking events. So we have a tracking component. It's probably a few million every day. So that's kind of data, we get some data from UCM, and so on. So scale today, and those kinds of things has become easy. This is some of the things that were technology changed a lot. So when I started, like the startup world, then building what we built here would have been heart, I would say, almost impossible. If you are not a huge, huge corporation like Google or something. Today, this is easily accessible and scalable, and something you can build, at least with the right engineering mindset and scale is they are Adil Saleh 28:34 saying this, you're saying this because of January were? Ole Dallerup 28:38 No, no, no, no. So this is more about scaling data volumes. So like when I started, I mean, now I'm old, maybe. But when I started here, like out to scaling of services, load balancing was available, but not a press on a button today, you just like scale to any load of CPUs with price on a button, right? That's also what is happening, which is AI today. Right? So and that was actually the bet we did at Green data. So we we have not impervious not build too much AI stuff into the product. We did that because we wanted to more focus on getting data streamline and getting data ready for to your on data Adil Saleh 29:34 trend. You're looking towards your own data machine learned or trained. So that yeah, Ole Dallerup 29:41 yes, we didn't do that in purpose, because we saw the hard part. And what we saw people struggled with was actually collecting the data and getting it into a form where they could run machine learning on top of it. So we focused on that initially, not saying that we could like we wouldn't necessarily continuing only doing that, but we started focusing on that. Because we believe that there will be some new models coming out. Not necessarily I predicted the either the time or the, like the how it looked. But we predicted at least advanced models would come that would solve many of these problems better than we could most likely do. And that's also what's happening now. So the last language models we see both from Google and Microsoft, K DBT. Open AI. Amazing, right? It's models that most of our startups would not never be able to build, even if we technology, we couldn't kind of afford building it. Now they're accessible. So we can crew it directly from BigQuery. It's amazing, right? So that means we can do very advanced stuff now. And because we have the data ready, we can do. So for example, a thing we just did the other day, in a couple of hours was we have titles pulled out from our customers, car cm. And then we did a model that could kind of build this into roles and Synology. Right, something that was potentially hard before, or at least something that took some time. And today it is very easy to do right as an hour. And so some problems are better pushed to the future, even though you couldn't do it today. And you are eager to do them today. Sometimes it's better to push them to the future because technology will solve it for you. And some problems you have to fix today. Adil Saleh 31:49 Absolutely. Absolutely. So while it was real nice talking to you today, and we learned a lot like since that's why I love so much having people with technical background more often. And you know, we get to learn a lot different things and really appreciate your time here today. Ole Dallerup 32:07 Most welcome. Thanks for having me. Taylor Kenerson 32:09 Thank you Ally Adil Saleh 32:10 Love that. Have a good rest of the day.

Keep Listening VIEW ALL EPISODES >>>