The product outgrew its founder

The demo was good. Worth saying upfront, because this is going to sound like a rescue story, and it is not.

A founder walked us through a subscription product she had built entirely alone. Next.js on the front, Supabase underneath, Stripe wired for the monthly plan, the whole thing deployed and taking money. Every line of it hers.

If the company had an org chart, it would be one box. Head of engineering, head of payments, head of data, head of customer support on Tuesdays. One rectangle, four job titles, one name.

For getting to market, that org chart is a cheat code. No standups, no handoffs, no waiting. She shipped in months what funded teams ship in years, and paying customers proved it.

Then the roadmap stopped being polite.

The next stretch of work, written down, looked like this. Stripe needed one-off purchases stacked on top of subscriptions, which is where billing stops being a checkout button and becomes proration, refunds, and dunning emails. PostHog was installed and had not recorded a usable event in months. Analytics that track nothing are very cheap to run. The email flows existed as a list titled “email flows.”

And the heart of the product, an async engine that turns a paid webhook into a generated report, a PDF, and a notification, had been scheduled for next month three months in a row.

Each module on that list is a solved problem for somebody. None of them was solved for her.

She was brilliant at the skills that get a product to market: deciding alone, gluing services together, shipping before the doubt catches up. The list wanted depth in four directions at once, and she was one person on one ladder.

The reflex move is hiring. At a company this size the math rarely closes. Months of interviews, a salary leaning on the runway, onboarding she has no hours for, all to cover a gap that might be six weeks wide. A job posting is a prayer that the problem will still matter by the time someone arrives to solve it.

The other move is gutting it out alone, which is quieter and costs more. The engine slides another month. The analytics stay dark. A roadmap like this does not fail loudly. It just ages.

Honesty about the first call: her budget and our estimate were not on speaking terms. She had a number built from optimism. We had one built from having shipped async pipelines and remembering what the second week always finds. The useful part of that hour was not the quote. It was running every line on the list through two questions. Is this missing skill, or missing hours? And is it a phase, or a permanent part of the product?

Missing hours on a permanent thing is what hiring is for, eventually. Missing skill on a phase is not a hire. It is six weeks of somebody else’s depth, borrowed, and then gone. The billing work, the events, the emails all turned out to be her: hours she did not have yet, skills she did. The engine was the only line in the other bucket, the one with a specialist-shaped hole in it.

You can run those two questions on your own roadmap tonight, no vendor required. Founders who built the whole thing alone almost never need four engineers. They usually need one, briefly, in the right corner.

Most lists shrink the moment you sort them by what kind of missing they are.

So one engineer embedded for six weeks, in the corner where the product was stuck. In week three she hit a wall of her own, a failure mode in the queue design she had not met before. At midnight she had people behind her who had met it. By morning the wall was a footnote.

The code lives in the founder’s repo. The decisions stayed her. The specialist left on schedule.

The org chart is back to one box, and we would love to claim the engine was the hard part. It was not. The hard part was a founder admitting the box had limits, and she did that before anyone got on a call. Everything after was just engineering.

The skills that get a product to market are not the skills the roadmap asks for next. Rent the gap. Keep the product.