Portfolio-level product strategy for a one-person studio
How to run five products without a hundred-person org: treat them as positions, not pets.
PromptXchange Labs has shipped five products across proptech, fintech and consumer AI. The question I get most is some version of “how?” — and the honest answer is that the studio runs like a small fund, not like a company.
The loop
Every product goes through the same lifecycle: validate → prototype → ship → measure → scale. Nothing is novel about those words. What makes it work is that measure is a gate with published criteria, and scale has an alternative branch called deprecate.
- Validate means a demand signal before a repo exists. A concierge version. A landing page. A paid pilot. Perch was a hand-curated weekly email before it was an agent.
- Prototype means AI-assisted builds that get to a working MVP in days, not quarters.
- Ship means real users in production early, because the only feedback that counts is behaviour.
- Measure means activation and retention are defined before launch, along with the thresholds that would cause the product to lose its engineering time.
- Scale or deprecate is then a mechanical decision.
Why stop criteria are the whole thing
Most product organisations have no shortage of ideas. What they lack is a way to kill the ones that aren’t working before they quietly eat the roadmap. A one-person studio can’t afford a single zombie product, so the kill decision has to be made in advance, when nobody is attached yet.
Publishing the stop criteria at launch does two things. It makes the decision emotionally survivable later. And it forces the activation and retention definitions to be real, because you have to be willing to be judged by them.
Capital allocation, not prioritisation
In a single-product company, the roadmap question is “what next?” In a portfolio, it’s “where does the next unit of engineering time earn the most?” That reframing changes behaviour. A feature request on a product below its retention threshold is not a backlog item; it’s a question about whether the product should exist.
It also creates optionality. Perch’s screening agent earned its keep as a standalone app, then turned out to be worth more as a buyer-side module inside Studio 3Sixty5. A portfolio lets you make that move. A single product with a roadmap would have had no place to put it.
The constraint is the point
I’d rather run five positions with clear exit rules than one product with a hundred-item backlog. The constraint of one person’s time is what makes the discipline necessary — and the discipline is what makes five products possible.