A Product Market Fit Show | Startup Podcast for Founders
00:28:12 · Pablo Srugo, Colin Zima
[00:28:12] S2: And then by the time you come back out and you get those six out of that podcast, that's when? [00:28:17] S0: Yeah, that was June. They signed in June, demoed in May and I would say before that, in probably September of '22. We had the moment where internally we were using the product and we're like, wow, this thing feels good. It was just like we didn't have dashboards at the time. Visualization was sort of young. Then we were like, OK, we've got some differentiation. Now we have to do all of the debt work so that we can actually exist as a BI tool in the existing market. [00:28:44] S2: And maybe just to bring it home in terms of the value, right? You mentioned the analogy, but take me to an example of a real data problem, right? At least my idea for this, is somebody somewhere in the business is like, hey, can you figure out whether ACV affects churn? Or can you figure out whether people in this market are more likely to buy this thing if this condition is true? And then that goes to the data team. Is that the right construct? [00:29:08] S0: It is, it's everything from we're walking into a board deck and we need to get financials to how are we doing against different competitors and what words. When we mention them, do we perform better with. [00:29:20] S2: And then flowing through that to this is how Looker handles it or existing handles it. This is what you were able to do.
00:29:20 · Pablo Srugo, Colin Zima
[00:29:20] S2: And then flowing through that to this is how Looker handles it or existing handles it. This is what you were able to do. [00:29:25] S0: Yep, so AI has transformed a ton of that. Now it's just you put it into a chat box and it just goes, and figures out everything for you. Before, it would be like, I need to make a dashboard that has a SQL tile over here and a governed metric over here. I know that sounds like the simplest thing ever, shockingly hard to do in every single BI tool that existed and we could do a demo where we could show you a SQL tile next to a governed financial metric that the finance team approved. And from there, we were sort of off to the races. [00:29:55] S2: You brought up AI, so I think it's good to talk about it. ChatGPT comes out near the end of '22, how is that lived for you guys? How much does that affect the roadmap or not?
00:29:55 · Pablo Srugo, Colin Zima
[00:29:55] S2: You brought up AI, so I think it's good to talk about it. ChatGPT comes out near the end of '22, how is that lived for you guys? How much does that affect the roadmap or not? [00:30:05] S0: I mean, I think that early, just talking personally. I used early ChatGPT for data and I was sort of like, this is not that great. It's not very interesting. I'm not going to use this and so I'd say for six months, I was just completely unconvinced. Kind of in the haiku stage for Claude. When Sonnet came out in October, our engineering team started adopting Claude code at that point and it was so obvious that the world had sort of changed in an instant. In terms of the way that people were writing code, that it became obvious that it was going to start happening for data too and even though we didn't believe in it. We started putting these AI features in the product, and our customers started using it. [00:30:42] S2: How early did you put these features out?
00:30:42 · Pablo Srugo, Colin Zima
[00:30:42] S2: How early did you put these features out? [00:30:44] S0: I'm forgetting the years now. So I don't know whether it was just this past year or not, but it was October. Whenever Sonnet came out, we had started landing some of that stuff. So I think it was October of '25. But we had a handful of customers that had started using the AI and they were like, no, it's actually really good. You should be trying this internally. We did an implementation internally where we're like, OK, let's set things up so it actually works well and literally that same week, we stopped using the UI. And it was one of those moments where we had to test ourselves. And it was so good that it was just really obvious the world was changing. And then, you just sort of playing catch up and trying to see what the models can do, and how you can incorporate them. [00:31:21] S2: It's noteworthy in the sense that in '23, and especially the second half of '23 and all of '24. Almost every new company was just jumping on AI, jumping on AI, just building with it. So, you're building a new company. By that point, like you said, by mid-23, you're just starting getting new customers. So you're really at the stage of these companies and you kind of, I don't know, I'm going to say ignore it but you kind of just kept going. You were basically on a path. You didn't really go off the path or change a roadmap dramatically until considerably later on. [00:31:50] S0: Yep, that's right. It was about a year later.
00:31:50 · Colin Zima, Pablo Srugo
[00:31:50] S0: Yep, that's right. It was about a year later. [00:31:53] S2: Maybe tell me a little bit about the go to market, since. You start off with more this ad hoc, go through and just find people that use Looker that you could pitch this to, et cetera. That gets you to that product market fit moment. Now, once you've got something that people use and onboard and seem to like, you need to scale that, scale in a small way, but get the $1 million, $2 million ARR. What do you use to do that? How do you do it?
00:32:15 · Colin Zima, Pablo Srugo
[00:32:15] S0: It's really funny, I've talked to our head of growth about this and our head of marketing. And everyone hopes that there's just some magical channel that they haven't thought of, that when they turn it on will inflect the business. What we found, that is the least satisfying but sort of most obvious conclusion that you could possibly imagine, is you just need to do more of literally everything. So early we had some organic, we had a little bit of outbound that was mostly the founders through LinkedIn, and we had a little bit of partnership. Which is our product needs to get implemented. There's a big data network of practitioners that help people with their data. Some of them would sort of recommend Omni and bring us in. And as we grew we've really just sort of leaned into each completely independently. So the outbound, we've scaled a big BDR org and they're on the phones sending email, scaling up founder LinkedIn. And so it's that but more. On inbound it's just mechanizing the machine but it's also we do a lot of outbound programmatic email and spinning up sort of enrichment techniques and things like that. Just identifying people to go talk to and then on partnership, we spent a lot of time with the hyperscalers. So Amazon, but also Snowflake and Databricks. And it was really just building relationships more broadly across the ecosystem. And then events is probably the last one that we started turning on. [00:33:30] S2: Did you find partnerships helped early on? I often find they help later. [00:33:34] S0: Enormously. [00:33:35] S2: They did, interesting.