The Exchanges

Q Um, and those are your own models, uh, or like, how do you go about building those?

A No, it turns out that the hard problem here is actually, I mean, there's enough SQL written in the world. There's enough. Even dbt code written in the world where just your standard foundation model will, will do this stuff pretty well, which will be interesting for us to figure out, like, can we over time scale this stuff with lower cost, whatever, like the right now, you know, you plug in GPT, 3.5 and it's not as fast as you want and it costs more than you want. Um, but like something is as stupidly simple as like, Uh, insert formula and you describe the English text to what you want the formula to do and, uh, having it write the regex without you having to go to stack overflow and say like, remind me how to do email splitting regex. Uh, it's just such a big performance accelerator. Um, and so I think that, I think analytics engineering is going to Change meaningfully over the next two years. And beyond that, I have no idea what anything is going to look like. So I don't try to predict beyond that.

AI assessment note: “No, it turns out that... just your standard foundation model will, will do this stuff”

Q And obviously a lot of the excitement around dbt from, you know, market and investment respective has been that dbt labs, the company and dbt core, I guess the project has, has, has owned this transformation layer. Do you want to explain what dbt actually is and what it does?

A dbt is the T in ELT. I was just talking about how the, this re-architecture. So dbt does not ingest data into your warehouse. It transforms it once it's in your warehouse. The funny thing about that is that if the data is already in the warehouse, then the only thing that you need to do to, uh, transform that data is write SQL, and you can do that in a couple different ways. You can, like, Create a view that abstracts some business logic, or you can create a table that stores the results of a query, or you can incrementally update the data in a certain table, but what dbt does is it allows data analyst, analytics engineers, data engineers to write these small bits of logic, modular business concepts, and slowly build up a Directed acyclic graph, a DAG of these concepts, and you go from left to right, and you start at the source data, and you slowly build up all of these concepts where you, and you eventually get to a place where you're dealing with business concepts that can be productively analyzed, and dbt is the framework that allows you to both, like, express all of that in code, but then also to run it against your database and, like, materialize all that stuff.

AI assessment note: “dbt is the T in ELT. I was just talking about how the, this re-architecture.”

Q And in layman's term, what is it actually a transformation? What are some examples of operations that one does?

A Yeah. Um, the, the one that I, uh, always like to talk about is, uh, we have, Uh, a self-service SAS product. Uh, we use Stripe for our payment processing. Stripe gives us invoices. The invoices have start date and end dates. Um, but if you just add up all of the invoice totals for a given month, you don't actually get monthly recurring revenue. In order to get monthly recurring revenue, you have to take your invoices and amortize them over the period that they are for. So if you have An annual subscriber. You need to recognize that revenue over the course of 12 months. And so a data transformation is, and it's actually like reasonably a complicated one is to take your invoices and amortize them into monthly revenues such that you can just have this table that every month you just add up the numbers and that, that gives you your monthly recurring revenue.

AI assessment note: “take your invoices and amortize them into monthly revenues”

Q So that's one part of the discussion of the intersection of the modern data stack and, and, and AI, which is, uh, okay, can AI, uh, you know, help you build superpowers to do data engineering better? Another part of the conversation is, uh, is the modern data stack. Part of the emerging stack around generative AI, are the things overlapping? Are they complementary? How do you think about it?

A Yeah, ok, so. I don't think we have good words for this yet, but We're seeing it happen internally and anecdotally I'm hearing it happening at other companies too. Um, one of the first places that we have invested from an AI perspective is making our support team more efficient. Not, we don't have our customers interact directly with AI agents, but we have AI agents that are there to accelerate response times on the part of our, our support reps. And that has made a massive improvement in support. Uh, We primarily from an efficiency standpoint, it's not that previously we were giving terrible responses. Now we give good responses. It's that we're now able to take a brand new support rep and ramp them up in half the time. Um, we are also able to, uh, get response times down and all of that relies on data that lives in our data, like our data warehouse. Um, dbt has kind of no awareness today that that is happening. And so in some ways you could say, well, the quote unquote modern data stack has nothing to do with this, but in fact, Infrastructure is exactly as valuable as the business value that is created on top of that infrastructure. So to a certain extent, the fact that there are people developing AI use cases all over the place that dbt generated data is powering is, is very, very good for us.

AI assessment note: “all of that relies on data that lives in our data, like our data warehouse”

Q And then what happened in 20, 20 and 20, 21, which, um, you know, I think in your piece, you talk about the memification of, uh, the modern data stack. It sort of feels like those were the years when it happened. So what happened then? VCs.

A Yeah, VCs. It's your, it's your fault. Um, well, it, it just, I think it became clear to enough people, both investors and practitioners and, and even enterprise buyers that there was something going on here. And I think that, um, the Snowflake IPO was a big part of this. I mean, the biggest enterprise software IPO of all time. Um, and All of a sudden that unlocks. I mean, it creates a huge marketing event for this technology ecosystem broadly. Um, but, but it means that all of a sudden enterprise buyers take this wave seriously in a way that they hadn't before. And as a result of that. There was increased demand from buyers and there was increased interest from investors. And all of a sudden now, I mean, I don't, I don't know your portfolio, uh, particularly intimately, but, but my guess is that you probably felt some need also to get exposure to the, the overall like data stack trend. Um, and, and when you have that belief from the VC side, then it creates Uh, it was very common for many early stage VCs to actually go out into big tech companies and find people who like worked on X or Y or Z modern data stack type problems. Yes. And then they would just like pull them out and say like, we'll give you five million bucks. Go, go make this. And There was just, there were too many companies created too quickly. Um, and, and I think that created some, um, I don't know, like unplea…

AI assessment note: “There was just, there were too many companies created too quickly.”

Q Still, to make this interesting for everyone, do you want to explain what data transformation actually is? Like a couple of examples, like two minutes.

A So, My favorite answer, my favorite example of what data transformation is, is that we worked for a grocery delivery company And one of the most challenging problems in the, that this company, ah, experienced was that they needed to calculate cost of goods sold for their orders. And cost of goods sold, they, every order was different, so the cost of goods sold needed to be able to go down to the individual product skew level. So like, you needed to say, what's the cogs for a, one of those little bunches of green onions? And, and like, it turns out that Calculating the cost of goods sold for, like, a bunch of green onions was tremendously complicated. It, like, you relied on all this, like, inputting cost data, and how, how big were the bunches, and all this stuff, and so this group of, you know, three or four of us would have these long conversations about what does it mean, like, what does that even mean, cost of goods sold for green onions, and you, you then eventually get to a place where you've kind of sorted that out. You've defined what that means, And you save all that knowledge into one table or a small sense, small number of tables. And then literally nobody else in the business has to ever think about that again. Like that's this really tremendously annoying problem that thankfully a small group of people can solve. And then if you've documented it well and you've lik…

AI assessment note: “data transformation is really this process of, like, taking this raw data, and applying business context”

Q much a, uh, a promoter and central Piece of it. So, um, I love the idea that you would ask the question in the first place. So maybe let's start with the concept of the modern data stack itself. Uh, is that, is that a functional thing, uh, for products working together? Is that a marketing concept? Is that a, an industry alliance? Uh, how do you think about it?

A I won't say that I coined the term modern data stack. I will say that I think I was a part of the community that started using that term, and I did some of the writing and thinking that led to that its popularity. And, um, it, it became a thing that people frequently said in the, like, 2016 through 20 19 period. The, the way that I experienced it was that, um, I had a decade plus data career prior to 20 12, and that data career felt very different to my data career post 20 12. And the, the big dividing line there for me was Redshift and not specifically Redshift because there have been now been other databases that have come out that are, that are like Redshift, but it's the, it is the year that the cloud first came for analytics. And the cloud really changes a lot about how you build an analytics product. It changes a lot about how you as a practitioner want to go about the work that you do. There's just a lot of assumptions that when you kind of start off, uh, you know, baking the cloud in, you have to kind of reevaluate them. And so in, in 2016. I, I was a consultant. I helped people implement quote unquote, the modern data stack. And at that point in time, it really meant something like, um, I, I had a strong preference to use tools like mode or looker over a tool like tableau because tableau was pre cloud and it assumed that you were going to be able to Now it did have som…

AI assessment note: “at that point in time, it really meant something like”

Q So where, where, where do your ambition start and stop in terms of, like, roadmap for the next couple of years?

A The thing that is, is neat about the position that we're in right now is that we get to, um, we get to ask the question, like, what, how should all this stuff work? Not like, what is the one piece that we can build, but like, oh gosh, we, we actually have a lot of people using this thing, and that gives us an opportunity to say, like, let's build something that maybe no one's actually Been able to build before. Um, one of the nice things about dbt is that it allows you to create this map that spans the entire graph of computation inside of an organization from, like, the data landing in the warehouse all the way through to, like, people using the data on the other side. But dbt actually understands, like, hey, this is a data source. This data is coming from Fivetran. And then it knows this is a data transformation. It's executing on Snowflake, or this is a Python-based data transformation. It's executing on Databricks. Um, and then here is a Looker dashboard that's querying this table, you know, et cetera. So anybody in the data ecosystem that's building a product or, or like in-house tooling can query this API and say like, hey, tell me the state of my data. You can ask questions like, Is this data source outdated, or does this data source, or does this transformation power a downstream dashboard? So the, one of the things that most of the practitioner space in, in the dbt com…

AI assessment note: “We're looking to be the infrastructure that powers this ecosystem”

Q one question from Jill. We didn't ask that question. Jill. Um, can you speak to the balance between the value of data democratization And the importance of data reliability and governance, particularly for large enterprises. Philosophically, what parts of the data stack do you think should be more democratized versus more controlled? What features and processes can you build into tools that democratize data to maintain data reliability and governance?

A Yeah, I love this question because the The assumption that we always run into is that democratization and data quality or data reliability are going to be intentional with one another, and I think that that is a natural assumption to make, um, but if you look at the population of software engineers that Exist today and push code to production applications. That has, I don't know what, oh, more than 10 X in the past 10 years. Um, and, and the way that that has happened is not that you have, like, a very Extremely tight controls on like who can push to production. What you do is you have mature CICD processes. You have like DevOps workflows that, so you like build these guardrails that create high quality processes around code releases to production where you have your cake and eat it too. You have democratization. You also have governance.

AI assessment note: “build these guardrails that create high quality processes... You have democratization. You also have governance.”

Q I read was, uh, that, uh, you've, um, made some, some really key hires recently, a new CTO, a new president and COO, a new vice president of product. Um, how did you think of navigating this transition, um, of, you know, one team to the other, one senior layer to the other, and what have you learned over the years about, um, hiring at different stages of the business?

A It's something I've been thinking about a lot. I, I think that one of the Everything about being a founder is, is about context. It's about, um, you know, there's some times that you want to be patient and then there's some times that you want to move very fast. Um, and I think to the extent possible, you almost always want to be patient. And I think that was very, very successful for us for a very long time. We waited, waited a long time to raise any venture funding at all. We had a thousand companies using the product before we raised a single cent. Um, and, and that gives you a ton of power and control. Uh, but we, with all of the market changes over the past Let's say, 18 months. And with the, like, very fast changes in scale of the business, It became very clear, uh, kind of very quickly that we needed people around the table that had a different set of experiences. And so if you look today, um, the only people who are consistent presence on our executive team from a year ago are me and our CFO and everybody else around the table is, is new. Um, and it's something I'm thinking a lot about right now. How do I, you know, there's a lot of strengths that come with that. These folks know way more than I do about operating in that scale business. I mean, every one of them has operated at larger scale than we're at now. Um, and so then where are their weak spots and how do I make…

AI assessment note: “we needed people around the table that had a different set of experiences.”

Q and all of this, which, you know, look, you could argue whether that's part of the, you know, very, uh, precise definition of the modern data stack or not, but this is clear path to expansion from a functional standpoint. So are we going toward a world where people will get most of their tooling from a small number of vendors or what would you think is going to happen?

A So, yeah, this is the question. This is like, Like my job and George's job at Fivetran and Ali's job and Frank's. Oh, this is like the thing that people are trying to, to figure out and kind of elbow each other around right now. Um, I mean, really, you know, kind of AWS and, and, uh, and Azure and Google kind of sit at the top of this food chain and then people like me and George just get to figure out what everybody else above us is doing. Um, but, uh, here's, here's some things that I, I would guess. Um, my guess is that if you sell compute and storage. Most of your money is going to come from selling compute and storage because it's just, it's a great business model. Um, it's, it's all five of those companies. It's, it's their primary business model. And it's been funny to navigate the, the partnership there because they have actually my read, it's not like I, I know the minds of the folks at Snowflake or Databricks so, so clearly, but like my read is that people like us, people like Fivetran, while we are critical to their success as platforms, because we drive a ton of consumption. They don't actually really care about the dollars that we make as businesses. Cause they're so tiny relative to the like actual business of, of compute and storage. Um, the way that you, like when you talk to the folks at the hyperscalers, they will say, we are, we're, we're like making sure tha…

AI assessment note: “This is like the thing that people are trying to, to figure out”

Q I think the fish channel analytics analytics, which was the, which morphed into dbt labs was a consulting company. So it was a popular open source project. You're, you're now a super well-funded startup. Uh, and there is now a product called dbt cloud, which is the, I guess, commercialization effort around dbt. What, what does that do, and how do you think about it versus the open source project?

A The original thing that dbt core did was it provided a language to express data transformations, and it provided a, a command line interface to actually execute them. Um, You know, we're, we were out in the world actually doing consulting projects, so I was, you know, the back story with me and venture funding was that I had worked prior to starting Fishtown Analytics, I had worked for seven years in three different VC-backed companies. Um, I don't know if any of you work at VC-backed companies, but it can be a reasonably high burnout environment. Um, so I was a little bit burned out, and I was like, no external capital, no external expectations. I'm going to, um, I'm going to fund this on revenue. And so we did that for three and a half years. We were, like, paid the bills via consulting. We, we, like, at, at the time, the only thing that existed was dbt core, and we clearly needed a way to operationalize this. As we're, like, working with clients, we, like, we've got all these great jobs described, but you need to actually, like, update data on whether it's 4:04 hours, or once an hour. It's not, it's not twice every second. Um, And, and so that was, that was like, we originally called it Sinter. We didn't even anticipate that it was going to be like a, an associated commercial product. Um, but it got more and more users over time, and what we've realized is that dbt core pres…

AI assessment note: “dbt cloud is, like, how you actually make that stuff true in reality.”

Q a, a big part of your overall positioning as a company. And in fact, your, your, your blog and podcast is called the analytics engineering podcast. So, uh, sort of like spearheading and creating that, that profession is, is like a thing that you guys have done wonderfully well. But like, so, uh, just to double click on the point, you think, um, analytics engineers become prompt engineers over time?

A Um, well, the thing that we are, one of the core beliefs about the, that we have about the, the profession is that data and software are not that different. And when you're building a production data system, you're building a production software system, which means that, uh, we should be taking lessons as a profession from things that software engineers often figured out decades ago. Now, in the case of AI, we're not so far behind because a lot of these capabilities have only existed for the past, whatever, a year or something. Um, but I don't think that Anyone believes that software engineers are going to go away. I think that they are going to get dramatically more efficient. So, um, if, if you, there, there are so many things in when, when you try to express an idea as code that you don't have loaded up. Into your prefrontal cortex, right? Like, and, and that's for a long time. We've used get Google and stack overflow and constant recompilation until you get a success message, uh, as, as kind of the workflow there. And, uh, that I think is. Like really what the future looks more like, like you express the thing that you're trying to accomplish. You go in there and you Fiddle around with some of the bits that require proprietary knowledge that only lives in your brain, and then you ask the thing to, uh, repeatedly recompile itself and fix errors until it gets to a compilable …

AI assessment note: “I don't think that Anyone believes that software engineers are going to go away.”

Q to be sure we, we touch upon is, um, building an open source company. Uh, so both of you are building your companies, uh, with a very strong open source components and a commercial product on top. How do you, I guess, how does that work? And how do you think about, uh, what needs to be in the open source and what needs to be in the commercial product?

A Can I answer that first? Because my answer is going to be a non-answer. Um, uh, I, uh, we are still early and I don't know. I, it's a very, very hard question. And I think that, uh, it is like something that the, you know, the entire B to B software industry has like embraced open sources like this wonderful formula, but, but the last generation of open source companies used a particular model. And in a lot of ways, I think the industry has said like, no, we're looking for, for new models. Um, so we have certainly have things that we are doing and that we're, uh, testing out, but we are still very early as a company. The thing, Jeremiah, that I'm interested in hearing from you, um, you, I heard you on a podcast, uh, uh, invest like the best. You talked about this insurance concept and I love it. It's like been very instructive in my own thinking. I'm curious. How uncertain do you know when you presented that it felt like you were just like, I know all the answers. Like we've, we've, we've answered this monetization question. Um, my guess is that that's probably not as, as true as that. Is that.

AI assessment note: “my answer is going to be a non-answer. Um, uh, I, uh, we are still early”