6 Pivots to $1.4b: PostHog Origin Story (and How They'd Do It In 2026)

Zero to one felt brutal; pulling that first dollar out of anyone was harder than going from one million to tens of millions. I’d gladly rerun the scale-up path compared to reliving that first stretch.

Welcome to Before It Clicked, I’m Sunny Recky. Today’s guest is James Hawkins, co‑founder of PostHog, which recently raised 75 million at a 1.4 billion valuation; they help product engineers understand behavior, ship with feature flags, run experiments, and act on the same customer data, and this episode gets into the messy pivots before it clicked.

PostHog is a developer platform with a broad suite of tools built on unified customer data. We’re product‑led, adding thousands of companies weekly, remote by design, small autonomous teams, and on track to hit a nine‑figure run rate with about 180 people across dozens of countries.

Roll us back to the start: why leave your jobs, and what did day one actually look like?

I’d wanted to start something for years, learned hard lessons at a venture‑backed company, and treated doing my own thing as a high‑upside bet with limited downside. I saved a year of runway, teamed up with Tim, a very fast builder, we were far from Silicon Valley, and we just filled out YC to see what would happen.

So, that first day?

Pretty scrappy—coffee shops, no office, I consulted a couple days a week to fund basics, tracked personal burn, sold my car, and cut costs to extend time. Tim built nonstop and I chased conversations and early users.

Where did the early ideas come from, and why build instead of only interviewing?

I kept a long list of pains from past roles, but our early validation was poor because we lacked product taste and talked to people who told us what we wanted to hear. We switched to developers for blunt signal, put product in hands fast, and learned by usage; in today’s AI world, building something real quickly is even more important.

You’ve said you chased the first paying customer; would you change that focus now?

Back then we were desperate for any revenue, which was a mistake; I’d now start with an ambitious, meaningful vision and work backward on funding and go‑to‑market. The bar is higher, people expect magic, and we leaned into that by going multi‑product earlier than conventional wisdom.

Could the earlier you have raised for a big swing?

Maybe, but networks like YC help, and the real unlock was confidence—stop people‑pleasing investors and speak with a strong point of view about risks, path, and why it will work.

How did you decide when to persist versus pivot?

We treated ideas like trying different sports and looked for natural ease—quick usage, buyers who paid, momentum, and that we enjoyed building it. Our technical debt tool had usage but not willingness to pay, selling felt like pushing a boulder, so we moved on, guided by intuition plus simple gates: can we get users, will they use it, will they pay?

Did anything change in your process as you iterated?

Switching to developers was huge, and we learned the website is the sales team—so we invested more in clarity, docs, and quality. Persistence mattered most; we just kept taking swings.

Give an example of pain leading to the next idea.

We tried territory management and a CRM angle—neither landed. Repeated frustration setting up analytics and dealing with unfriendly vendors pushed us to a developer‑first, open‑source approach; YC’s advice that janky internal systems hint at demand also rang true.

Some advise staying in a single market so insights compound; you hopped—mistake or feature?

We hopped out of urgency and personality, and a hot market pulled us along; hard to recommend universally, but the wide search worked for us.

With AI, how should founders think about MVPs now?

Your first version needs to impress, and it’s faster than ever to reach that bar; the bigger hurdle is believing you can build something remarkable and then doing it.

What about competition and defensibility?

Competitors validate demand, and a sharp twist on a known category is easier than inventing from scratch, but you must stand out with product and storytelling; platform risk worries me more than rivals. We also used radical openness—public strategy and open source—to pull competitors into our game while we stayed a step ahead; in AI, real defensibility leans on distinctive UX and data, not thin wrappers.

How did you design the successful launch?

We did a Hacker News pre‑mortem, pre‑answered common critiques, and over‑invested in a substantial site, docs, and use cases; we were transparent about who we were and how we’d make money, and treated the site like part of the product.

Walk us through the origin of PostHog and the moment it felt real.

We were pitching engineer retention, sensed it was mushy, and flipped meetings to ask about in‑house analytics; teams wanted control and flexibility, so we shipped open‑source self‑host first to earn traction. Self‑host created huge support drag, so we built cloud, it quickly became the vast majority of revenue with far fewer tickets, we sunset paid self‑host licenses, most customers moved, and growth accelerated as we layered more products and leaned into AI‑native workflows.

Quick plug—what are you hiring for and building now?

We’re working on a desktop app that ties into your product data and lets you choose the model, then drafts pull requests and suggests merges based on behavior, issues, and feedback; if you want early access, email me at james at posthog dot com.

Thank you for the candor; the big takeaways are build more, notice your own pains as you go, and keep going.

Pivoting feels existential, but product‑market fit is reachable; extend your runway, keep taking shots, and commit fully together rather than half doing two things.