A product function is not a queue of tickets. One person can hold it end to end, if that person has done all of it before.
What it covers
- Development direction, including the decisions nobody enjoys making: what is rebuilt, what is left alone, which debt is paid and which is tolerated knowingly.
- Deployment and operations: the servers, the backups, the pipelines, the failures at the wrong hour.
- Testing and quality, in proportion to what the product actually risks.
- The market side: positioning, pricing, launch, and the feedback loop from users.
- The business side: suppliers, budgets, and a roadmap that survives contact with reality.
The proof: a product rebuilt in-house
An investor-backed property business has had its whole product function run this way since 2024: v2.0 shipped, then v2.5 and v2.6, with v3.0 now being built on a rebuilt backend. Development was brought in-house after years of external contractors, and the product started moving again. The project write-ups are on this site, with the versions and the decisions in them.
The same person also runs two businesses of his own, so the operating advice is not academic.
Capacity, stated plainly
At most three engagements run at a time, and two are taken. So the honest answer to most enquiries is no. That is a signal about the work, not a sales tactic.
The engagements that exist get the attention they were promised. The ones declined are declined before anyone wastes a quarter discovering it.
The next step
A conversation about what the product needs and what it would take. It can end in an engagement, in a narrower version of one, or in no. Start at how we work.