A Product Market Fit Show | Startup Podcast for Founders
00:04:11 · Alex Halliday, Pablo Srugo
[00:04:11] S0: I knew Sam, yeah. So we met when I first moved to San Francisco, walking down the street. Actually, during San Francisco Pride. We were just wandering down the street with ten thousand other people and he was walking next to me. We got chatting. We've been friends ever since. [00:04:25] S1: No way! That's just crazy coincidence. [00:04:28] S0: Yeah, it's crazy. Well, that time in SF as well, you just would walk into a coffee shop and Jack Dorsey would be there or Kiefer Boyce or Peter Thiel. I look back now and I realize how insane it was. But yeah, that was the origin story there. So he got me onto LLMs, and I just totally fell down the rabbit hole, and the business started moving in that direction. And that was kind of one of the big inflections in the business that changed its direction. [00:04:53] S1: Yeah, tell me more about that time because I can imagine as you're playing around with these tools is right before ChatGPT. On the one hand, I'm like, yeah, I could see how you play around. You're like, wow, this is so impressive. It's not that easy to connect that to a business opportunity, like you figure out, OK, here's how I'm going to use this. What were some of your initial guesses?
00:05:12 · Alex Halliday
[00:05:12] S0: Yeah, it's such a good point. I remember I was on a plane flying to Atlanta and I was like, OK, I'm going to go do this. I'm going to try this OpenAI thing and sign up for an account. So I remember my first few prompts I put in were actually around SQL generation. It was like, can it create credible SQL for actually working with Shopify stores randomly? That was my initial throw at it and when it started coming back with kind of joins and, let's say, plausible SQL in those days. I just felt like the laws of physics that govern software were changing. It was just one of those moments where you're like, wow, there's this general purpose intelligence, and I didn't really have. Honestly, a fantastic business case for it back then. We started applying it initially to generating SQL and explaining data tables for customers. We released a little Chrome extension on Product Hunt. We started, you know, just kind of creating little experiences for existing customers. But there was this pulling of a thread that took us about six months that led us to our eventual customer, which were marketing teams. But what I realized was that there was going to be this explosion of use cases and businesses that would use this technology. And for me as a founder, the idea of being truly horizontal and trying to do everything for everyone just kind of lowkey terrified me. I remember feeling very strongly that we needed to pick a customer that we wanted to build for a decade and we had people using our tool in all these different ways. And I just looked at this map, and I felt if we try and serve all these people, no one is going to be happy. So we ended up picking marketers. It was a great choice. They found us, and then we doubled down on them.
00:06:51 · Pablo Srugo, Alex Halliday
[00:06:51] S1: Just to be clear, what were marketers doing with the product initially? [00:06:55] S0: Yeah, it's a great point. So we had a four-step workflow template for writing an article like a blog post template. It was very throwaway. We put it on the platform and I remember thinking, no one's going to need something that's four whole steps. That seems like a lot and then we would log into the platform on a Sunday evening. And we would just see all this activity, and we would click into it. And it was marketers building like fifty to eighty step workflows, to try and do really precise content work with LLMs, breaking down the problem into little steps. [00:07:23] S1: And this was, by the way, timeline? Was this before or after ChatGPT? What are we talking about? [00:07:27] S0: But at this point, this was very early in 2023. So this was just after ChatGPT. [00:07:32] S1: Just after. OK.
00:07:32 · Pablo Srugo, Alex Halliday
[00:07:32] S1: Just after. OK. [00:07:33] S0: Yeah, yeah, so it was people who played with ChatGPT, and they wanted to do more with it. So they were like, OK, like this, I want to orchestrate it, I want to do that extra mile and so they were coming to our platform. And I remember just thinking, is this the customer that we want to build the business against? And what I loved about marketers is they are so creative with the technology, and they have really high taste. One piece of advice is build against the high-taste user, and they will keep rejecting you till you pass their very, very established exit criteria for a problem. And it's the best way in the early days to get really good at something, in my view. It's very hard for a team in the abstract, unless they have recently practiced that role, to know what very, very good looks like. So that was kind of how we found our true customer that we ended up building a whole business against. [00:08:20] S1: And what exactly were they doing eighty steps for? What was the end product of that?
00:08:20 · Pablo Srugo, Alex Halliday
[00:08:20] S1: And what exactly were they doing eighty steps for? What was the end product of that? [00:08:25] S0: Yeah, so in those days, it was a lot of content creation for pretty basic stuff, right? But it would be like creating local pages for a national franchise or things that we typically call like programmatic SEO. So that was our first real cluster of interest. So I remember the CMO over at Deepgram created a workflow where he would look at Product Hunt every day for new products. He would then research those companies and create pages for each startup that launched on Product Hunt that met his criteria. And because those companies had nothing written about them, he would rank number one for all of them. And because Deepgram had domain authority. And so he'd figured out this sort of early use case for LLMs to create interesting, novel content in a semi-automated way. And I think it's 35x Deepgram's traffic to their website. So that was an early opportunity. Now, the space has moved on a lot from that, but it was incredible to see both the energy they were putting in and the wins that they were getting.
00:09:24 · Pablo Srugo, Alex Halliday
[00:09:24] S1: So this is interesting because it's this mix of you've got a product, you've got some data, you've got people using your product in a certain way, and you're actually able to leverage that to double down. Let's go deeper on that because I think this is not an easy thing to do and a very important one. How do you take that next step? Do you go to everybody that's using it on the marketing side and have fifty conversations? Do you start maybe more from like first principles, you know, strategy? What is kind of that next move to say, OK, if we're going to go all in on marketing. What does that really look like? [00:09:55] S0: Yeah, I mean, I think it's a good question. The first piece of it for us is, can we solve the problem for one customer really, really well? So it's that true solution acceptance by a high-taste user. Where they go, this is it, this is great. I will be a referenceable customer. I really, really like this, and then you have to figure out whether or not the pain that this problem that you're solving is widespread enough as you need it to be, and also acute enough. You don't want to be people's number three or four problem. You want to hopefully attack their number one problem. So whilst you're figuring out if you can actually solve the problem, what the solution looks like, starting to learn about the shape of the sales motion. How you would end up selling this to more people. How broad-based this problem is in the market is incredibly important as well.