A Product Market Fit Show | Startup Podcast for Founders
00:11:17 · Pablo Srugo, Shmulik Fishman
[00:11:17] S1: Are you charging Avis at this point, or what are you trying to get out of this first pilot? [00:11:21] S0: Yeah, there's billing or invoicing set up at that time, so that was not in place. This is a company – it's still the culture of the company today, but we're really building backwards because we're not industry veterans. I don't come from an income verification background. I have no experience in credit bureaus. I've never worked in any of these businesses, and so we've been so client led from the very beginning. We've just been getting feedback from each one of our clients. The first client that paid us was a business called Level Goals, which I – is still around. They're a small dollar lender, and Level Goals only was interested in doing lending to gig workers. We had already built the first integrations into Uber and Lyft for the Avis use case, but they just wanted to use those same data points but to issue a small dollar loan for $100. I believe they gave us $1 per verification. That was the first paying client, and that is a much bigger business just for the industry, for income verification. That's also become a major business for us as well. [00:12:35] S1: How much time went through from the first pilot with Avis until you signed that first paying customer? [00:12:40] S0: It took a while. The first year of Argyle, I believe, until the late part of 2019 – Argyle was found in 2018. Until the late part of 2019, we had no paying customers. We were just building, and I think we got our first couple of customers in late '19, including Level Goals.
00:12:40 · Shmulik Fishman, Pablo Srugo
[00:12:40] S0: It took a while. The first year of Argyle, I believe, until the late part of 2019 – Argyle was found in 2018. Until the late part of 2019, we had no paying customers. We were just building, and I think we got our first couple of customers in late '19, including Level Goals. [00:12:58] S1: What was hard about it? What are you building? From the outside looking in, you're like, okay, we've got this resume builder thing. It's got the main KPIs on it. I feel like somebody should pay for that.
00:12:58 · Pablo Srugo, Shmulik Fishman
[00:12:58] S1: What was hard about it? What are you building? From the outside looking in, you're like, okay, we've got this resume builder thing. It's got the main KPIs on it. I feel like somebody should pay for that. [00:13:05] S0: What is novel about Argyle, what we had to build literally from scratch is connectivity into payroll processors. Some history or some stats to make it interesting, the IRS says that there's about 6,000 registered payroll service providers, which is a really wonky term, but basically, it just means there's 6,000 entities that are allowed to submit payroll on another company's behalf. That's a super fragmented market, and in order to provide income verification, Argyle needs to connect into every one of those platforms. For any business of value to use a verification service, you need to have most of that connectivity built out. The initial wedge and, again, totally on accident just because there was this huge correlation between people that worked for Avis and gig workers, we had this natural wedge at the beginning where we were building out connectivity just into gig platforms so Lyft and Uber and Instacart and DoorDash, Lugg. At the time, Postmates was still around, and that is tough to do if you're doing it all by yourself in the very beginning. You're having to build microservices into third-party applications that you are not under the control of, and so you need to build into their spec.
00:14:33 · Pablo Srugo, Shmulik Fishman
[00:14:33] S1: How did you think about when to charge? I'm on this question because I think the lean startup methodology would tell you, as soon as you deliver a unit of value, try and figure out what it's worth, especially, as a first-time founder, what you're going in with or what you're hammered with, if you go to an incubator accelerator, all those sort of things. You didn't do that, and you didn't do that with the experience you had in your previous company. It was deliberate. Curious if you could just dive a bit more into how you thought about it back then and where you were trying to get to in that first year before you started charging. [00:15:05] S0: I'm a big advocate for having a pilot customer that you're giving the product away for or the feature away for free for a period of time. They're helping to design the first version of it with you, and we still do this today when we release new product. The first couple of people that use the new feature, we give it to them for free for a little bit because we want them to see value in it and for it to be part of – when it becomes part of a general release, then you should start to charge for it. I think it's important to get people using your product and to be able to get feedback on your product, even if it's negative, and then put the pricing part on it. That's been a big part of my strategy, and I think it works at different levels of scale as well.
00:15:05 · Shmulik Fishman, Pablo Srugo
[00:15:05] S0: I'm a big advocate for having a pilot customer that you're giving the product away for or the feature away for free for a period of time. They're helping to design the first version of it with you, and we still do this today when we release new product. The first couple of people that use the new feature, we give it to them for free for a little bit because we want them to see value in it and for it to be part of – when it becomes part of a general release, then you should start to charge for it. I think it's important to get people using your product and to be able to get feedback on your product, even if it's negative, and then put the pricing part on it. That's been a big part of my strategy, and I think it works at different levels of scale as well. [00:15:52] S1: It's funny; you talk about design partners. I was actually just watching the review of the AI Pin and also the Rabbit one or whatever, and that was the comment. These products are over MVPing, like the humane pin. It's just coming out half-baked. It's worse with the hardware, obviously. You come out half-baked with all these promises of features in the future. You've just taken the MVP thing too far, right? I think what you're saying is, especially I think with enterprise, the design partner helps you get to figure out what the real MVP is that you can really charge for. [00:16:26] S0: Yes.
00:16:26 · Shmulik Fishman, Pablo Srugo
[00:16:26] S0: Yes. [00:16:27] S1: The risk is you're building something that actually is never a top priority, never going to become a top priority. You're getting stringed along. What signals are you looking out for to make sure that there's something real here? When you decide to charge for it, the money will be there. [00:16:42] S0: To use your AI Pin as an analogy, I like building tools that are just a better mousetrap instead of something that is actually new, and I think the AI Pin is like that's a net new product, which makes it even harder to extract value out of or to price around. Everything that Argyle builds, even in the early days, is just a better, faster, cheaper, more automated version of something that a business is already paying for. Every lender already pays for income verification services and to know income by year. We're just giving you the next generation of it. I think it's easier to price around that because there's already a reference of value with a client, but they're already spending on it. They're not allocating new spend, or having to get a new PO, or get a new approval. I like to say that every business spends on income verification services already. You can go to your P&L. We do this as part of our sales cycle. How much money do you spend on this today? We're going to cut your cost. That’s a very different way of communicating with a client, with a prospect than trying to convince them that they need a widget that they're not buying today. We just have a better widget.