Building a $4B data platform: Inside dbt Labs’ unconventional path | Tristan Handy (Co-founder and CEO, ex-RJMetrics, Squarespace)

Brett: All right, let's do it. Well, thanks so much for joining.

Tristan: Thank you for having me.

Brett: Whenever you think about the moment the company was started. what was like the story, the two years before that, what was going on?

Tristan: I had been working in data for a very long time. I, you know, my, my core skill set over the course of my career has been quantitative. Uh, I had practiced that in a bunch of different ways. Like I had been at Deloitte Consulting and did large IT projects and data stuff as a part of that.

And, went to the, startup side of house and, uh, ran, data team at Squarespace and very nascent data team. and then I, I joined an analytics tech vendor called RJMetrics. And, I was an early user at Squarespace of this company, RJMetrics. And, um, we were supposed to be what Looker became. Uh, the thing that RJ was never, and I don't think would have ever been able to. Uh, anticipate was the cloud, the rise of the cloud for data specifically. And, uh, so I remember experiencing this in, in 2014 or 2015, I got to use a product called Amazon Redshift for the first time.

And as a, as a long time data practitioner, I was just. Very immediately within writing a couple of queries, I was like, Oh, this is going to change everything about my profession. 

Brett: What did you see in that moment?

Tristan: the performance was like two to three orders of magnitude faster than the analytical products that I had had to use previously. I mean, I don't know how geeky we want to get on this particular topic, but, before Amazon Redshift, in order to get access to that class of performance, You had to be a large enterprise because the only products that, uh, were sold with that type of performance were called data warehouse appliances.

Netiza, Vertica, Teradata, et cetera. And, and you bought them like literally on a box and you installed it in a server closet and the box cost a hundred grand or whatever. And so all of a sudden you could like swipe a credit card and pay 160 a month for this class of technology. And it was completely game changing.

At Squarespace, I had been running the analytics functions such as it was on like a, an old MySQL database you worked with what you had, but it was very clear that this was like a different world. and so the, the cloud actually came for RJmetrics and, you know, it, it disrupted our entire business and it was what created the opening for, Looker to come and kind of steal the exponential curve that we had been sitting on top of and, and they did great and, but I was.

I was excited to just get out there and use the stuff. My fondest desire was not to go and build another technology company. It was actually to like real companies, real, real customers use, the new cloud based analytics tech that like had already been built.

And so I, I didn't actually start a technology company. I started a consultancy. 

Brett: Maybe you could share more about, you're at RJMetrics. You're seeing the sea change. Why decide to sort of start a consultancy of all the things that you could do.

Tristan: So I had worked in three VC backed companies at that point, you know, so, so when everything goes well for a VC backed company, It's great. And it, is like the kind of thing that, books get written about and episodes of the acquired podcast get, made about, but, it often doesn't go that well.

and you know, the venture capitalists talk about this, you know, we're, we're shooting for the power law returns and, and all of that stuff, but, I was involved in one success Squarespace, and then I was involved in two, like, not big successes. And the challenging thing in that environment that nobody really likes to talk about is that it is, it's like an unbelievably high level of personal stress when you're a leader at, at an organization that is not winning.

it brings stress into your home life. you fundamentally feel like you're not in control of the process because you have investors that have expectations and ta ta ta ta ta. I wanted to get back to my roots, which is like doing quantitative work, not thinking about, Oh, how do we hit our next quarter, revenue goals?

like I'm fascinated by this technology and I want to figure out how to help customers, deploy this in the best way possible. And I wanted to own a hundred percent of the business so that I didn't have anybody that could prevent me from focusing on exactly what I was curious about.

Brett: and so was that an easy decision to go and start this consultancy? Or was it one that you spent a lot of time wrestling with? 

Tristan: I spent a bunch of time wrestling with it. Yeah. I mean, I'd never started anything. I'd worked for three other founders and I had always had this idea in my head that founders were a different type of person and I was not that type of person. So I ended up talking to those three folks and, they tried to help demystify it for me and, all encouraged me to do it. And I think that was, it was helpful. I, I honestly don't know that I would have had the confidence to do it myself if I hadn't seen the process before. And if I hadn't had the people that I had done this with say, like, I think, I think you can, you can do it, it was still a gut check.

And the thing that I used to kind of de risk this for myself is that, I mean, one, like we didn't take on any real financial risk, like, I wrote a 10, 000 personal check, I deposited it in a bank account and then we like went and got some clients. So like the worst case, I could lose some months of my life and I could do lose 10, 000.

It's just like, not the end of the world. But like, I still was afraid of the idea of like, I would have to go tell all my friends and family that like, I tried and failed at a thing. And I live on the East coast and the culture on the East coast is not quite as entrepreneur friendly as it is on the West coast.

So, you know, that felt stressful, but I, I like sat down and I wrote out a list of values where I felt like if I showed up at work every morning and I lived these values that whether or not I was commercially successful over the years, I would feel good about the experience if I, if I did that.

Brett: what were some of the values that you wrote down?

Tristan: things that you would kind of expect, like, um, transparency always wins is one of the most important ones that we have. But, but that's, think, not that controversial in the startup ecosystem. I think most startups really have a hard time living it and we really live it, but it's not that controversial of a value.

One of the ones though, that I think is a little bit unique to us is we are humble. I think a lot of startup founders start companies, Specifically because they're not humble. and, and that's okay. I'm not being critical of that, but I think that if you start from a place of like deep humility about, yourself and your own, fallibility and limited capacities as a, as a human, you actually build a different organization and that can have its own strength too.

Brett: why has it served you? if you sort of take a crude brush to great founders, right, you just kind of lump them together. I would say high humility, low ego is probably underrepresented.

the very act of thinking you're going to start a company that can be generational, you need to have some level of arrogance to do something so insane. but what I love about what you're saying is, the reality is there's lots of ways to build wonderful companies.

Tristan: I don't think that you can have a company value of we are humble and go found Amazon. just wouldn't, wouldn't work. yet the company that we have built, the reason that we've been successful is that we've been hyper attached to our users. We've built, a thing that everybody said was a bad idea to build.

But it was because we were our users we then turned around and open sourced it and no one had ever built a meaningful business in this particular category, which we haven't really talked about yet. But the thing that we do is called data transformation. it's, we build kind of developer tooling, that allows you to do data work inside your warehouse.

Nobody had ever really done it this way and, it turned out the thing that we built so resonated with users it kicked off an organic growth curve of users telling other users. And in data that basically never, and like enterprise software broadly, organic growth almost never happens.

there's too many points of friction, like the buying processes and, salespeople and all of these different things, inject friction. But we grew consistently at 10 percent month over month for about six years running. And it turns out you, you grow like very, very large user community when that can happen.

And that all comes from our, we are humble value because users will buy from you if you, offer a great product at great prices, but they won't always Advocate on your behalf, unless they really like you and they, they like what you stand for. And so this humility, I think is what allowed us to build a real community.

Brett: Now we kind of sped up a little bit. Let's go back to you put 10, 000 in the consulting business. You saw sort of the sea change that was happening with Redshift and new databases. what happened next? 

Tristan: Very fortunate to, have turned out to be my first co founder. Drew, uh, was somebody that I had worked with at RJMetrics for three years. He was a part timer, he, he, uh, worked, to pay his way through Drexel while he was, finishing up his degree there. So Drew and I had worked together for three years at that point.

you know, as some college kids do, he like missed all the on campus recruiting deadlines. And so, right around when I was starting, he, he pulls me aside and he's like, Hey, I hear you're starting a company. Like, can I help? And I was pretty resistant to that, but, uh, ultimately, decided to say yes.

I wasn't resistant because of Drew. I'm Drew's freaking incredible. And it has been my partner in this thing over the, over the many, many years. But it was just the, the idea that I had another mouth to feed, another salary to pay, and, and I had not yet closed my first client. And so I was like, hope I can figure this out.

but the first six months were really like making sure that we found enough clients and were able to deliver work for them in a consistent way such that we could pay ourselves 

Brett: Did you find them and what did you pitch them on you helping them with?

Tristan: so this is 2016 and this is, there's two, market trends happening in VC that really helped us out. One of them was direct to consumer and the other one was SaaS. I had worked, I'd been an exec at, three SaaS businesses to that point. And I had worked at my last company at RJMetrics with, with a bunch of direct to consumer e commerce companies. So I like knew both of those business models, like quite well and the data behind them. we were very explicitly focused on, series, A, B, C, companies that, needed to understand their, their data. And we knew the playbook, we knew the technologies that we had to use.

and we can deliver like really an unbelievable amount of value, very quickly and with not a lot of consulting dollars, so the, so the play was like, You could go hire a full time data person, but they're not going to have the experience that we have. Or you could hire us. And I think our annualized rate was, I don't know, something like 60K a year.

So it was like kind of half of an entry level data analyst. we were just like, look, we will produce more value for you than if you hire internally.

Brett: And were a lot of those customers folks that you had met through your work at RJMetrics 

Tristan: my first two were, but after that, it was marketing and community. Most people who start consulting companies don't have a background in marketing, but like the thing that I did as my day job for seven years in the startup space was, was digital marketing and 2016 was a very good time period for content marketing specifically.

And so. I built a brand for us by writing blog posts that really resonated with a bunch of people. And I had a newsletter that was read by, I don't know, 6, 000 people. And, all the time that most consultants that are trying to build businesses spend like going and meeting potential customers and pitching them on like, Hey, don't you want to work together?

I spent writing and then I would, you know, people would come to us.

Brett: For the first two, what, what did the work actually look like? So they're like, yeah, we, our data's a mess. It's really important. We figure this out. I, we're going to hire somebody and then they go with you and then you show up and what do you do?

Tristan: the first two weeks where like, Hey, you're going to need three or four pieces of technology. let's help you understand what the options are and let's make the right choices, but like, there's only a couple of choices to make, like, great. Let's go procure a couple of pieces of software and plug them all together.

And we would never just do that. We would always like do a proof of concept, like, Hey, here's a piece of value at the end of the day. there was usually some thing and it was often either the CEO or VP of marketing that would be our initial customer.

so it was more often the VP of marketing at direct to consumer companies, and it was more often the CEO at SaaS companies. They would come to us with like some kind of core problem. Like, uh, if it was marketing, it was like, I need to understand my channel mix and like, if it's working and where I should invest money.

And if it was, the CEO, it was like, I need to understand my, churn rate, my cohort analysis. And maybe it would even be to like, get together an investor packet so that I can go out and raise money or what, but it was like, understand the core, user dynamics of the business. But like once you landed that core thing, there were like infinity other things to do. So you'd like, we would build credibility, at the outset and then we'd be trusted partners. So some of those early customers we worked with for multiple years.

Brett: Over the first handful of customers, how similarly shaped were the customers and, when you went company to company, did you start to see we're doing very similar things? Or was it still a little bit, a little bit different over here, a little bit different over there? Like what was the shape of work and how repeatable was it even in the consulting days?

Tristan: Very repeatable. I don't know that this was like that clever. It wasn't like a, some genius strategy. It felt kind of obvious. I know e commerce and SaaS, let's focus on that. And I knew how many e commerce and SaaS companies there were because I'd actually done market analysis at RJMetrics to like figure that there's like a lot of e commerce and SaaS companies out there.

We were very tightly scoped from the very beginning and it meant that the work was very repeatable and in a cons it's very strange to me. I think this is a thing that many, people who build consulting companies, the thing that they get wrong is that they accept whatever work comes their way.

I said no to anything that wasn't in those lanes. and that meant that we not only did we get better doing that work, but we also developed like internal IP. Around how to do that work better. we were also able to hire, people recently out of undergrad who had good quant skills, and train them to do this work because they were able to develop pattern recognition.

So like, if you, just drop a consultant in and say like, go solve these random data problems with this startup executive, they have to have 15 years of experience. But if you say like, Hey, this is exactly the set of problems you're going to solve. And here's other instances where we've solved this before.

You can have somebody brand new in their career come in and do that.

Brett: So where did you go once you started to have your first handful of consulting clients?

Tristan: There were only a small handful of data technology companies that were specifically built for the cloud. We talked about Looker before, but there were a couple other. Analytics, BI tools, there were a couple of data pipelines, and because most companies were not using the cloud at that, point, certainly the enterprise had not gone to the cloud yet.

In the process of building a consulting company, it was very clear to me based on work that I'd done earlier in my career that there was a missing category. in data people talk about this thing called ETL. extract, transform and load. And it's the ability to, take data out of all these different systems and pull it together into one place.

And in the process, reshape it so that it's useful for analysis. And the cloud really changed the way that this work was done. And there were a bunch of companies that I would, well, there were two companies that I would consider to be the E and the L, the extract and load. So they would go to Salesforce or go to Stripe or go to Zendesk or whatever and pull all that data into your warehouse. no one had taken seriously the T part of it, the transformation part of it. In order to do good consulting work, I felt strongly that like we needed a tool here. The question became like, okay, what's the, what's like a quick and dirty way that we can build a product here that is immediately useful to us, and doesn't, doesn't require us to spend a tremendous amount of engineering time. And so we can talk about the like initial design philosophy around the product, if it's interesting, but, a really DBT started out an internal productivity tool to support the consulting business. 

Brett: And you didn't sort of have a glimmer of maybe this could be its own thing. It was not a part of building it?

Tristan: No, not really. it was so different than anything that existed before. So like data tech prior to 2016, always thought of data analysts as kind of low level, non technical people. You can see that in, Excel or Tableau or other products that like existed before that. Yeah, you can do powerful things with these products, but they're not fundamentally treating you as software engineers.

They're, they're trying to like, make it simple for you, drag and drop or whatever. For a variety of reasons, my, career had like brought me very close to working with software engineers so that I had developed some software engineering chops, modest though they are. And I understood all the benefits of like modularity and source control and deployment pipelines and all this stuff.

And, I wanted all those advantages in the tool that I was going to use to do my work. And so we kind of modeled the original product off of Terraform, but like, I actually talked to a couple of VC people in my network in the early days and they were like, I mean, that's. Interesting. But I think that you might be like an end of one user who wants a thing like that.

And I was like, okay, that that's actually totally fine. that wasn't the goal. It was a product that had to create a category in order to be successful because nobody had a thing like that before. And so nobody had shaped a career or a skillset in order to, to do this.

So data people had to learn how to use the command line, had to learn how to use Git, had to learn how to like write modular code. In order to get value out of this product, but it just so turns out that like, at the same time in a very adjacent space, data science, everybody, like every magazine was proclaiming data science, the sexiest job of the 21st century.

And so all these people who had been sitting in data analyst career titles, they all of a sudden thought like, Oh, I want to develop more technical skills. And so that perfectly collided with like DBT's entry into the market because a lot of people saw DBT as a way to build up their, tech skills.

Brett: But again, you didn't build that with that tailwind in mind. It was just to solve this pain point in terms of the work you were trying to deliver to your customers?

Tristan: Correct.

Brett: What did the first version look like? And did you already have team members who were technical enough to build it or who built the thing?

Tristan: Drew did. So, recently out of undergrad, was learning, how to do the analytics work to like keep some consulting clients happy so that he could pay his own salary too. Half of his time and probably nights and weekends, he was hacking to, to build the initial versions of this thing.

And, we had played with, in the run up to starting the company we had played with prototype versions of it. I wanted to start a business having an understanding of like how I was going to help clients on day one, and so was trying to get my tool stack all ready and all this stuff.

So we had a version that had probably, two weeks of engineering time put into it on day one. and fortunately, like when you build command line tools, you can get a lot of value out of them. You don't have to build a user interface. You don't have to build login. so you can get a lot of value out of a small amount of effort.

Brett: when you use the tool for your first couple of customers, what did that look like?

Tristan: Honestly, the customers didn't even have to know this tool was being used behind the scenes. And that was a really critical part of it because if we'd tried to convince people to use this piece of technology that like nobody had ever used before, probably wouldn't have gone that well. But, and it's also kind of esoteric.

if you're not actually a data practitioner, like you have no idea what the hell we're talking about. And so we kind of made it transparent to them. Over time, as those companies grew and hired data practitioners, and as we started to work with bigger companies, people started to actually get exposed to the product. The first customer that their internal team actually started using DBT actively was a company called Casper. I don't know if you like sleep on a Casper mattress, but I only sleep on Casper mattresses. Their data team hired us to solve scaling of their data warehouse back in 2016 and, we were like, well, we, gotta like implement this thing. The way that you're doing this stuff right now is bad and we've got to move it over here and they were like, okay, give us a demo. And we gave them a 15 minute demo. And over the next week, we like refactored all of their existing pipelines and brought them over to DBT.

And they were elated. 

Brett: That first internal tool, what did it tangibly allow you to do for your customers?

Tristan: We could deliver work at like, I mean, this is like an unscientific number, but I would say at least 10 times faster than we would have been able to do otherwise. It was like the magic silver bullet that made the entire consulting model work.

Brett: And the actual product allowed you to do what? The early version of it?

Tristan: Take all of the data ingested from Stripe and Salesforce and wherever else and create tables that allowed you to analyze marketing channels and customer attention and all of this other stuff. But like when that data originally comes over from Salesforce or Stripe or whatever, take Stripe for an example.

Stripe sends you a table with, Payment events. It's like one big event stream of like, you charged this person this much money, this person canceled, this person did this, et cetera. And you somehow have to get out of that. what's my churn rate? And it turns out there's like multiple steps in the process to get there.

Brett: So you started to expose more of the product to the end customer as you worked, moved up market and have more sophisticated customers, what then happened?, 

Tristan: The craziest thing happened in like late 2016, we maybe had been at it for five months. We had this Slack group I mean, it was maybe a dozen people or something, that were early adopters and we all just kind of interacted together as peers, trying to figure out like DBT really allowed you to work in a new way as a data practitioner.

And so we were all trying to figure out like, how do we do this stuff better? And then a random person showed up in the Slack group and said, I heard about DBT and we've been evaluating it for our internal use. And that person's name was Steve Ritter, and he was the VP of data science at SeatGeek. And we were just like, what?

random person off the internet finds out about DBT and is like taking this thing seriously? And his big complaint was your test coverage and your python repository is garbage. Like, do you treat this thing like a real piece of software or. should I, you know, do do I not trust it? And we were like, okay, Steve, if you and SeatGeek are going to use DBT, we like, we will add unit tests.

Cool. and so we did. And so SeatGeek has been using DBT for like eight years now. but Those kinds of things happen more and more often. And so the Slack community became from like this group of people who all kind of knew each other and was all hanging out in there. It became a self aware group of early users where like a new person would join and in the general channel, they would say, hi, I'm X person from X company, and I heard about DBT.

Like I'm here to learn, help me with these things, and, the community like really came together and, supported those new people. And it felt like we were all kind of figuring it out together.

One of the weirdest experiences was people in that community that were like, that ran data teams, they would host meetups for DBT, to try to convince other people to use DBT because they needed more people to hire.

We didn't have a marketing channel where we, we were like, Oh, let's market the use of DBT. It was like, people would invite us to meetups about the product that we were building. It's very strange.

Brett: While you were consulting when the SeatGeek head of data science came in. You started to let people use this little tool that you've built it in a non consultative engagement at that point or no?

Tristan: When we first built the like initial version of DBT, we stuck it in an open repo on GitHub and we put an Apache 2.0 license on it and so it was just out there available. And so as I started writing more marketing stuff, as, as early users started telling their friends. People would just find it.

Brett: Did you choose to open source it? I would think it's an internal tool, what made you want to do that?

Tristan: One of our values is kind of esoteric and it's, it's borrowed from, some folks at, uh, union square ventures. It is, we contribute to the knowledge loop and, the knowledge loop is just this idea that like the only thing really that separates us from caveman a hundred thousand years ago or whatever, is this iterative cycle of like learning new stuff and then passing it on and then the next people build on that stuff and they add, the next.

So, so it's like all cultural knowledge that we, that we pass on that, that, you know, separates us. And, you know, we didn't know that a ton of people were going to use DBT, but we felt like we were investing in this thing and it was a good idea and maybe some other people out there would, would find value in it.

Brett: So you have this ecosystem sort of forming around this internal tool that there are events that you're not putting on, but you're getting invited to about the thing that you've built. What were you thinking at that point?

Tristan: this will make finding new consulting clients way easier. We did take seriously the idea that like, huh, this is neat and, and after the Casper engagement, we, Drew and I took the train home from New York and, uh, we live in Philly and, we installed anonymous event tracking inside of DBT for the first time ever.

so that. We could just get some idea of not who the people were, but how many of them there were using the product. you know, we're data people. We want to know that kind of stuff. But, but like literally these kinds of cool things would happen. And then we would. Go back to thinking about our consulting business.

Cause that's how we paid our salaries. it was really like a year later when we, we pulled up the chart and we're like, there are now a hundred companies using DBT and we've only worked on a consulting basis with like 20 of them. And. we threw that time series in a Google sheet and we, we fit a line of S fit to it.

it was very consistent, 10 percent month over month growth. When I was at Squarespace, the period of time that I was there for, we very consistently grew at 10 percent month over month. And so I like understood the value of compounding very clearly. and I just continued to check that line and, it just continued.

 The next honestly, from like 2017 to 2019 it was a very similar story. Like we added more people, we added more clients, then the community continued to grow at 10 percent every fricking month.

Brett: What was informing the product roadmap for DBT? 

Tristan: I think we finally hired a full time engineer to work on DBT in like 2018. 

Brett: So like two full years. After building it,

Tristan: yeah. So my, my co founder Connor. Joined in Jan of 17, six months after we got started. And so Drew and Connor were both software engineers and they spent about half their time doing consulting work and half doing software engineering, but they kind of like pay their own salaries.

it was weird though, to hire our first full time software engineer where like we were paying him, a reasonably good salary and he produced 0 dollars in direct revenue for the business and had no prospect of really ever doing so. Over time, we, we eventually hired, two more software engineers as we continued to scale.

But it was like incredibly clear that that was not going to work. Cause what we were doing was we were taking all the money from the consulting business and like shoveling it over here to pay software engineers that weren't making us any money. And that meant that we were probably underpaying our consultants and it was just like not sustainable.

but no, the product roadmap was things that we needed to deliver consulting engagements. And things that people in our community raised their hand and said like, Hey, this is a big problem for me. I would love to see DBT fix it. And sometimes because it was open source, they would fix it themselves.

Not all the time, but sometimes. 

Brett: For those that haven't thought that much about starting as a consultancy, oftentimes, because investors say the number one risk is you're just going to be a consultancy. But what's some of the goodness that other people might not know about in just starting as a consultancy versus trying to sort of start a more traditional venture backed software company?

Tristan: VC is wonderful if you are really going to build a company that can potentially IPO someday. and maybe it does, or maybe it doesn't, but like if it has the potential, then, then VC can be really wonderful. As an entrepreneur, there are a ton of really, really good outcomes that can happen well below an exit event.

It is wonderful to just be a capitalist where you build a thing that people want, they pay you money for it, maybe you hire some people. you have a P and L every month, you watch the cash balance go up and down and hopefully it goes up and you can maybe, you know, you can build a consulting business and sell it for two to three times revenues.

Or you can just like continue to operate it because you love what you do. But fundamentally, like you get all of these choices that are literally not on the table, if you're a venture backed company, There's not a quote unquote venture scale exit potential in a consulting business, but there is a really satisfying founder scale, potential in, in those businesses.

and they, they can be like really just fun to, operate. Even when in 2019, we had a thousand companies using DBT and continued, you know, 3x year over year growth rate. I still didn't want to take venture, but it was so clear we were doing the product and the community a disservice by under investing in it.

We had three engineers supporting a thousand companies and DBT was generating at the time, literally hundreds of millions of dollars of Underlying compute spend from the warehouses that it plugged into. It was too important and it was too widely used and it was too under resourced.

And we were just like, Okay. It's, it's just like not that responsible of us to continue operating in the way that we've, we've been operating, but it was, we had to really convince ourselves to, to do something other than run the business the way that we were Cause it was, it was fun.

Brett: The fact that you were a consultancy enabled you to get so close to so many customers and see the problems in such a deep way. How important was that to sort of the early days and eventually sort of becoming a more traditional software company? Kind of getting to be that close with that many customers.

Tristan: There's two superpowers you get if you come via this route and there's plenty of costs. I'm not suggesting that this is like the only right answer that it's for everybody. . one of them is that you as founder, CEO live and breathe the technology in a way that you fundamentally can't, or at least can't for very long as a venture backed software company. Like, I spent all of my days focused on, like I did client work, up to the day that we raised venture money. and so I was like deeply, deeply in the weeds of our product and the problems that it was solving.

I was like writing code as a user would every day. The other one it gives you time. the thing that you need in order to unlock a venture scale outcome is you need some type of exponential growth curve that you're sitting on top of. And sometimes you can't just like magically will those into existence.

Sometimes you have to discover them and the process of discovery, I understand that there's a lean startup approach and there's Y Combinator and all this. I actually am not a big supporter of these approaches, because I think fundamentally what they're trying to say is you can take the process of innovation and you can put a methodology around it.

And like, if you only approach the process of innovation with this methodology, then you are maximizing your chance of success. And I think that's fundamentally like not, correct. I think what you need to do, is you need to give yourself time and earning revenue gives you time and you need to make sure that you stay close to customers and having a consulting business keeps you close to customers.

Brett: So what was going on between the hundred customers using DBT and the thousand, which seemed like you changed your mind. Were you just sort of making the product a little bit better, taking on consulting customers. what was going on between that period of time?

Tristan: Yeah, we continued to, make the product better. We, had to rewrite it a couple of times for various different reasons. the kind of thing that would be very challenging if you were, you know, you had some aggressive product roadmap that you had to hit. We supported new adapters, we continued product innovation, but mean, that was like three people doing that stuff. I was spending my time operating a consulting business and, you know, doing, doing marketing stuff. So I worried about things like, you know, this client is pissed at us, or, you know, this person that we hired maybe isn't growing into the role in the way that we had hoped, or, you know, all of this kind of boring stuff, but the point, if you were sitting on top of and exponential, then you don't need to be heroic in the process. Like what you need to do is continue doing the things that support that exponential curve. Do the next, build the next feature, make the next customer happy. Your mindset changes pre and post recognition that like you're sitting on top of an exponential curve.

Brett: Talk more about sort of that observation around a thousand where you made the call that we're going to change the direction and become a more traditional software company. What did that, what were like the conversations? What was that like?

Tristan: It was super stressful. Connor and Drew and I fundamentally made that, made that decision. And, the funny thing is kind of naively thought that everybody, the business was going to be really excited about that. It was not true, we churned about a third of our team over the next year.

Brett: How many people were there at that point before you made the change?

Tristan: There are like, I don't know, 18 or 20 of us, something like that. And I remember we actually closed the fundraising and shared the news immediately before an offsite, or I guess at an offsite. And, when I shared that news, I thought that everyone was going to be really excited. And I was like, oh, that's just some weird energy in the room.

I underestimated, and this is probably true of any like meaningful change in strategy for a business, I underestimated the amount of work that would be needed to explain the strategy, explain the reasoning behind it, explain how different people fit into the new world, uh, explain why it made sense for them to be a participant in that new world, but like had a productive, transparent conversation about that, as opposed to just kind of dropping a bomb and saying like, Hey, isn't that exciting?

It was not only the scale of the thing that, that drove us in this direction. It was also we had enterprises show up in our community for the first time in 2019 and we just like had no way to help them. Like they couldn't use open source software by itself without a support contract.

And just didn't have the scale to support, a huge company like USAA or Coca-Cola or Chesapeake energy, like some of our early customers. So it was, it was like, Hey, everybody, I understand that, this is the business that we have today, but if we want to live into the potential of this thing that we've created, here's the road that we've got to go down.

Brett: What was it about that thousand customer mark that pushed you to go do that versus 900 customers, 600 customers, 200, like what got you to flip?

Tristan: things were breaking. If we felt like we could have just continued to do this in the way that we were doing it, I don't think we would have changed, and now AWS has, has kind of changed to become more open source friendly in the five years since then. But in the 2017 through 19 period, it was like a massive concern for every open source founder that AWS was just going to take your thing.

and put you outta business. And, now they put you outta business. Maybe it was like a little bit of an exaggeration, but that was the fear. And because all of our innovation was in the open source, uh, it was, very plausible that somebody with dramatically more reach and resources could come and just kind of take the thing over.

Brett: What were the negatives of starting the company in this consulting fashion?

Tristan: I'm now, an angel investor and I've invested in, three or four dozen companies and every once in a while I invest in a company that has such a clear idea from the outset of what they're trying to do. And oftentimes second time founders. Or at least like folks who have, really solid careers ahead of that.

they set out, they raise money, they build a team, they whiteboard exactly the thing that they're trying to build. And then they build exactly that. And then it works. And that's like the wonderful thing about being a second time founder is that you like, you can plan all this stuff out really nicely.

If you have the market insight, right. 

Brett: That's a big if.

Tristan: It is, but you know, I've. I have a lot of respect for, Eric at Modal and Nick at, at Dagster. And there's like people who like really do this, even, you know, we were talking about Bob Moore before with, with cross beam. I think he's, he's gotten a lot right, it's cause he's a second timer. And, when you grow in this like much more organic way, well, it does allow you to discover things that are really special. it means that. You live with all of your legacy organizational debt and tech debt, for a long time. so I think that aspects of our product had a hard time scaling to become a commercial product, even though we had an unbelievable amount of market pull, as we initially commercialized, it was not super easy for us to transform this open source command line tool into like a, a product that, you know, Enterprises would pay a million dollars plus for.

Brett: You know, a few minutes ago you chose around a thousand customers of the open source project. you chose to go all in in building a software company, basically, as opposed to a consultancy, you raise money, you talk to the team, some people left

and then what happened?

Tristan: We have had many of the bumps and bruises that you would expect along the way. the initial, leadership team that I had grown the business with scaled for a long way, but over the past 18 months, I've, replaced every single person that, reports to me with like the folks that can scale to the, next level.

we've, tried to figure out how to marry open source culture with commercial culture. yeah, we, so we've, faced plenty of challenges along the way, but. Honestly, the, how we got to that point in the first place was very unusual. And the stories since then are, ones that you could hear from other founders.

Brett: But when you raise the capital and then at some point I assume needed to build a closed source or cloud offering to monetize it, what was that process like? Or aside from not consulting any more, what are the first handful of important things that you did?

Tristan: Got the modestly sized software engineering team in a room and, built a browser based integrated development environment, built a, orchestrator, like we knew what we needed to build because they were again, like the things that we needed ourselves. It was a very different motion to build proprietary software versus open source software.

And so we had to kind of. Learn the basics of that. You know, you had to build enterprise features like role based access and control and SSO integrations and all this stuff. but it wasn't rocket science. It was like, okay, now it's time to build the cloud version.

Brett: You were continuing to improve the open source version as well. 

Tristan: Yes.

Brett: And so you split the team basically.

Tristan: yeah, we, for a long time had, three streams of innovation. three parts of our engineering team. We had the people focused on the open source, we had people focused on our orchestration product and people focused on our, code authoring ID product.

Brett: When you started to build the cloud offering, did you build it with a very specific subset of customers that looked like a certain size and certain use case or were all the things very obvious? So you didn't, focus on, I mean, you had so many types of companies doing so many types of things at so many different types of scales at that point.

Tristan: We didn't have that luxury this time around, or, or maybe I like, I just wasn't clever enough to navigate it. Now we had lots and lots of inbound demand.

We were not culturally that good at saying no, and that's probably my fault. but so we built very horizontally, you know, the same product was used by, I don't know, think of your direct to consumer e commerce company series B in 2020 and like the largest, soft drink maker in the world. very, a very wide gap to span.

And, and I think that honestly, don't know in the grand scheme of things, if that was a good decision or not. But I do think it was hard for us to try to do that. We had a team of, I don't know, in 2020, we'd probably hired up to 15 software engineers. I think that parts of our community got irritated at us because they could see that a lot of our innovation was going to service this new group of users that we'd never prioritized before.

Brett: How long did it take you to build the first version of the cloud offering?

Tristan: we had parts of it in the works in 2019. So we got it out the door in 2020, shortly after fundraising. Like we, we knew what we were going to fundraise on, we were going to build a commercial business on top of. It was, I mean, I think our seed investors, which that seed investment closed in November of 2019, must've been very happy because within two months of the round closing, we had landed two fortune 500 companies as, as customers.

And so, that helped them out.

Brett: did the cloud offering to your point you're making here just immediately click? Like it was such linear execution that you knew what you had to build it. You shipped an early version. You had your first five, 10, 15 paying customers. It was relatively smooth.

Tristan: think one of the things about product market fit is that people don't realize that you have to keep solving for it.

Constantly. you, you have to solve for it at zero and at one and at 10 and at 50 and at a hundred and, different parts of that journey for different companies feel hard and different parts feel easy. so we had solved a lot of the product market fit questions that we needed to solve to go from zero to whatever, 40 or so, during that early phase.

But then we had all of this next set of questions, like how does the product expand from there? How do we convince more and more people to, pay for a commercial version if there is a, an open source version that they can use for free? And there's, how do we explain the value of the commercial version?

Like all, all of this stuff that. Honestly, when you, when you have enough product market fit, sometimes it allows you to get away with not being super tight on different things, like, product marketing or sales motion or this kind of stuff. And so all of that kind of hit right around 22, which is the same time that software purchasing was under the most pressure.

and so it was just like from this gigantic acceleration curve, it was like overnight, okay. You got to like sit at the adult's table and you got to figure things out real fast.

Brett: I mean, we could probably spend hours talking about just those few months in 22, but what were the most important things that you figured out?

Tristan: We had to have a very clear, explainable answer to the question, why should I use the commercial product and not the open source product? And it had to be digestible by someone with a C in their title. and for a long time, we actually didn't know if we were going to be a PLG business or a, sales led business.

our first, let's call it 10 million in ARR was like very PLG oriented. and so it felt like maybe that would just continue to be true. in data, that's like almost never true. when you sell to software engineering, sometimes you can go PLG for a lot deeper in the journey and, you know, Stripe and, Twilio and others have, have done that very successfully.

but data for a variety of reasons that are probably too boring to get ends Needing to be sales led and it, and most of the dollars come from the enterprise. and so, we couldn't just build a broad set of tools that developer, that, practitioners really loved. We had to really focus our efforts and, build cohesive stories that we could explain to, to senior data leaders.

the big unlock for us was As people use DBT for longer, the complexity of their code, went up and up and up and it became really a problem for the most sophisticated users and oftentimes they were at the, largest companies. And, so there was a real opportunity for us to step in and solve that.

And so we, for the first time started talking about complexity as the biggest differentiator or like solving complexities, the biggest differentiator for, DBT cloud. and that, that's a storyline that like, you know, it was clear to show it in the data and it was a storyline that like enterprises could understand and pay money for.

Brett: How did you get to that unlock? Was it just spending time with enterprises?

Tristan: It was literally like this one chart. And we, we showed it on stage in our, I think it was in our, I can't remember 22 or 23 user conference, but it was, a chart of the number of DBT projects that had over a certain number of models in them. You can see over a hundred, over 500, over a thousand, et cetera.

we watched that number climb and we knew as users, like, Oh my God, trying to work in a DBT project with 5, 000 models in it is like kind of challenging. And so then, we went out to, you know, we started with that quantitative data point and then we went out to, folks in our community and talk to them about their experiences with these like very large, complex DBT projects and validated that like, yeah, this is a pain in the ass.

And if we didn't find answers to some of these questions that long term somebody was going to disrupt us here, because, this is not a great experience.

Brett: Where do you think you were the biggest beneficiary of good luck in this eight year journey thus far?

Tristan: I certainly think there's luck all over the place, but, but I think that I didn't understand how good our timing was going to be, in, at the beginning. I think that

there are a lot of humans who have had ideas that were very similar to the idea for DBT. And it just so happens that the community aspect of DBT Is highly defensible. Like people want to use programming languages that other people use. And so once there is growth curve behind one that does a certain job, then the incentives are generally to use that one because then you can hire people.

to do it, you can get resources on it, et cetera. And so I think that the opportunity window would have closed if we were, I don't know, a year later, two years later. I don't know that it's possible to know that kind of thing at the outset. And I think that this is where the, like being in the arena quote really matters.

You just, sometimes you have to just step in the arena and see what happens. cause all the experience and, all the, you know, business school courses in the world are not going to like answer some of these questions for you. 

Brett: How would you break down those underlying tailwinds? You kind of talked about them throughout the conversation.

Tristan: It's the move to the cloud. Enterprises really started moving to cloud for their data in 2020. There was like Hadoop stuff going on before that, but that like never really achieved any level of maturity or real penetration. and yes, Redshift and Snowflake like did exist prior to, and BigQuery did exist prior to 2020, but, none of the platforms were like really ready to scale.

And it really started to be 2020 when large digital transformation efforts and, and, you know, got pulled forward by COVID. but large digital transformation efforts kicked off that said like, okay, we're going to take all our on prem data processing and we're going to move all of it to the cloud, which is then like got people to say like, okay, cool.

I've been using this legacy vendor for 20 years. am I going to use in my new cloud environment? And that's what unlocks the opportunity for a company like us.

Brett: I think it's going to be maybe a tricky for sort of you to answer this given sort of how you articulated your values. You're in some ways, this. begrudging CEO, right? And that you didn't sit down and say, I want to go build a company and eventually be a publicly traded CEO.

That's not my, I don't, it wasn't even a part of the early incarnation of the company, but like the body of work, if you assume that it's not all blind luck, indicates that you're a very special person that was really well fit to this opportunity. So like, if you had to pick it apart, what's the why behind you personally?

Like, why, have you been successful?

Tristan: yeah, I, accept that. and for what it's worth, I'm actually now having a lot of fun in my job. The first couple of years of, the venture funded journey for us were like extremely stressful for me and like a big transition, but, I'm having a lot of fun now. The rise of the consumer internet and mobile and the era when like the social network movie was made was that it kind of convinced people that founders were supposed to be college dropouts and they were supposed to pound on the table and say, we're going to change the world and all this stuff.

And, I really deeply, I'm curious about how the world works. I have an MBA, I have five years of consulting experience at Deloitte. I worked at three different venture backed startups, for seven years in executive roles. we talked about stepping into the arena.

Well, I never imagined that I was going to be a founder. I was a student of the game for a long time and I spent most of my time reading. blog posts about, the startup ecosystem. I have like this deeply curated art, RSS feed all the humans that I deeply respect and read every single thing that they ever write.

I think that experience is just like extremely undervalued in startup culture. And I think that's a huge mistake. And one of pieces of advice that I most frequently find myself giving to founders are people that are thinking about starting companies is like, go work at a company before starting your own, because like, you don't need to actually experience every single failure yourself.

Like you, you can actually accelerate that process through other people's learning,

Brett: What are the other things that, you tend to tell folks who are thinking about starting companies about their journey or the early work trying to get into product market fit.

Tristan: That it's not a playbook, you know, for a long time, people would get in touch with me saying like, Oh, DBT created this great community. How do I create a great community? I'm like, well, tell me your context right now. And inevitably it would be like, well, you're, you're actually not situated create some giant community.

so you have to like fit those puzzle pieces together. I think in, general, like get a coach really soon, even when like that coach feels super expensive. I, at one point in time, I paid 2 percent of Of our monthly revenue to a coach. cause like, that's how important I consider that.

I think that egos get in the way tremendously, or like maybe not just egos, but like our own psychological weirdness, which everybody has, gets in the way of success I think there shouldn't be any CEO out there that doesn't have a, who doesn't have a coach. 

Brett: Is there anything else kind of, when you think about the topic of understanding that your, key point is there's no paint by numbers to get to product market fit, but are there certain things that someone can do that will increase the chance? 

Tristan: I think it can be very challenging when every single customer call can be. Oh, everything is terrible or, oh my God, maybe we have a billion dollar company here. this emotional rollercoaster in the early days can be so significant

if what you're trying to do is create a billion dollar company. And sometimes it's actually this nice psychological hack to lower your standards. And instead of being so attached to the outcome is like, I'm building a company. The outcome is like, I'm solving a problem in the world. I don't know what size company that's going to be, but if I'm attached to the problem that I'm solving and I'm super curious about the problem and so obsessed with it that I couldn't imagine not spending all my time on this, then I think you increase your chances of actually being financially successful.

The success of the company has to be downstream of all of that stuff, but you, you can't constantly be focused on that. Or I think it's self defeating.

Brett: You mentioned this passing comment that I was curious about what, what happened in the couple years after you raised capital that was so personally challenging for you?

Tristan: Well, we raised capital right before, COVID started and so. Um, we shut down our office, went distributed. Everybody else was figuring that out too, but we were, we had a 15 person team growing to a 50 person team that year.

And we were trying to do that and all the hiring and all the onboarding in an environment that was like brand new for us. I was trying to figure out how to lead a larger group of people. Like I had never seen myself as a person who was like great at standing in front of large groups of people.

Inspiring and motivating them. So I had to figure out my own journey there. we had an acquisition offer for more money than I would have ever been able to use in like three generations of my family. like things came at us too quickly for me to process them effectively. But it just felt like a constant stream of emotional overwhelm.

Brett: Maybe we could just wrap up where, where we often like to, which is when, when you think about a bunch of the topics that we talked about today. Is there someone in and around the company or or in your life that's had a disproportionate impact on how you figured this stuff out or has kind of shared things with you that are a big part of the way that you approach this work?

Tristan: , The person who like probably top of that list for me is a guy named Jerry Colonna, who, was a VC became a coach, and got me into the idea of CEO coaching

and I, I ended up working with his firm for a little while. And essentially via blog posts and podcasts, taught me the version of CEO that. I wanted to become. More empathetic, more introspective, more written, less verbal. authentic, how do I be my authentic self in front of everybody as opposed to constantly be playing CEO on stage? so I think it's great if you live in a hub of startup activity, great, like develop your own mentors, but if you live in a place like Philly that, you know, has some entrepreneurship and not a ton. you can get a lot out of relationships at a distance too.

Brett: Good place to end. Thank you so much for spending the time with us. I really, really enjoyed the conversation.

Tristan: Thank you. This was fun.