if you look at the last decade of second-time founder companies, you'll see a striking pattern. people who built successful consumer companies the first time around, second time: infrastructure. people who built successful b2b saas the first time around, second time: developer tools. people who built successful fintech the first time around, second time: some kind of financial primitive. it's so consistent that it's almost funny. the question is why.
the obvious answer is that a second-time founder has become more sophisticated about technical problems and wants to work on harder things. this is partly true but doesn't fully explain it. the deeper reason, i think, is that every founder discovers during their first company a list of problems they'd have paid real money to have solved, which nobody was solving, and which they themselves had to build internal tools for. after they sell or step back from that first company, the problems they remember most vividly are the boring infrastructural ones they had to cobble together ad hoc — the monitoring, the deployment, the billing, the authentication, the data pipelines. when they decide what to build next, the memory of those pains is a much stronger signal than whatever consumer insight they had the first time.
the first-time founder has a different calculus. they want to build something new and exciting that resonates with users. they are picking from the universe of ideas, which is mostly consumer and application software — the stuff a smart person outside the industry can observe and form opinions about. the second-time founder picks from a much narrower universe — the pile of infrastructure pain they personally remember, which almost nobody else can see clearly because it's only visible from the inside.
this produces an interesting asymmetry in the startup market. application companies get started by people with bright ideas but no specific insight into what's broken in the infrastructure layer. infrastructure companies get started by people who are painfully aware of what's broken because they lived it, but who wouldn't have been able to articulate it before they lived it. this is why infrastructure companies tend to be built by operators rather than observers, and why they tend to appear in successive waves as new application patterns create new categories of infrastructure pain.
the other factor is that dev tools and infrastructure businesses have a distribution pattern that suits second-time founders well. they sell to engineers, usually bottom-up, starting with individual contributors who bring the tool into the company. this is a channel a second-time founder can seed from their existing network — they know a lot of engineers at good companies, those engineers will try anything interesting, word spreads. a consumer business requires paid acquisition or a growth-hacking channel that probably doesn't exist yet. infrastructure spreads through trust and social proof. second-time founders have lots of both.
the narrower, less charitable version of the pattern is that second-time founders are often less willing to take consumer risk. consumer is a graveyard. most consumer products don't find a market. the ones that do require enormous taste and timing and luck. infrastructure is more legible. if your tool solves a real engineering problem, a finite number of teams will buy it for a finite number of reasons, and you can forecast revenue within a factor of two. for a founder who has already done one impossible thing, a bet on a consumer product is uncomfortably correlated with ruin. infrastructure is a survivable bet.
there's a meta-observation in here about how founders should think about what to build. the first company is often a bet on a thesis about the world. "people will want this." the second company is more often a bet on a deficiency you personally experienced. "there was no good way to do x, and everyone who tries to do x knows it." the second kind of bet is empirically more robust. it's also more boring, less likely to be a generational consumer success, and more likely to become a solid long-running infrastructure business. this is the right trade-off for a lot of second-time founders, but not for everyone.
my informal advice to people starting a company now. if you are a first-time founder and can only see application-layer opportunities, that's fine — pursue one, and accept that you're betting on a thesis. if you are a second-time founder and you remember five unsolved problems from your last run, you probably should pick the most painful one and start. the market rewards founders who have the patience to solve problems nobody else could see because they didn't have the scars. scars are a moat, actually.
the pattern — consumer-then-infra, saas-then-devtools — is consistent enough that it's become a running joke among investors. it stops being funny when you realize it's an efficient market sorting founders into the problems they're uniquely qualified to solve. the second company is usually the one they should have built first, except they didn't know enough yet.