Product-market fit is more transient than ever

Jason Lemkin (SaaStr) and Rory O'Driscoll (Scale) went on 20VC recently and shared a number of interesting observations about the startup landscape today.

SaaS investing, as we've known it, is basically done. This is the key portion of the transcript:

Rory: There was a twenty year period where it was pretty obvious what to build, and you built it. And because the broad direction was obvious, all that was left to analyze was the math. It was pretty clear the direction of travel was take X, move it to the cloud, compound for a long period of time, get a great outcome. Pretty straightforward. So the only thing you had to do was – it was actually fairly simplistic – evaluate the relative growth rates of different things and pick the thing that's growing the most, at the most efficient level. There was a ten year period where these companies didn't change. I invested in Box in 2010, and it didn't change. They're still around – in the eSignature market, it's DocuSign. The thing we invested in in 2010 is the exact same thing in 2024 – it's stunning. So there was no conceptual thinking about, "What should we build next?" It was just about – "Build this thing, and sell as much of it as you can." And that's done now... [because] two things happened at the same time. The existing markets saturated, so all the growth rates flattened out. Anyone who needed a Zoom account or a DocuSign account has a DocuSign account, because they all got one during Covid. We're done. At the same time, these new AI startups took off, where... unlike the SaaS thing where stuff was the same for 20 years, this shit changes every six months. I've had companies acquire and lose product-market fit two or three times in a two-year period. It's terrifying! So it's way harder now. When it works, it's way better. But, oh my god.

Jason: It used to take you five years to fall out of product-market fit. Now it can be five weeks.

Rory: Absolutely.

Jason: When we started, if you hit product-market fit, and you had a decent team, you had five years to run. You had to reinvent yourself around year four or five. But you can't count on any of that today.

Rory goes on to explain the two underlying causes of this as (1) the rapid pace of model progress and (2) the fact that we're undergoing a paradigm shift.

Two examples of startups that come to mind are Jasper and Tome. Both had huge momentum and hype in 2022-2023, but later underwent major pivots due to changes in the landscape.

It's an open question what "good" AI revenue looks like. We're seeing startups like Cursor set new records for revenue growth, but revenue retention remains to be seen.

Rory and Jason's observations underscore the importance of moving quickly, something that small teams leveraging agents do best.