I like starting from a small impulse. It might be a page, an automation script, a knowledge structure, or just a recurring inconvenience that keeps asking to be fixed.
But an impulse is not enough. It quickly turns into a list of desired features, and then into a project that feels impossible to finish. When that happens, I try to stop and write three sentences: who this is for, what specific problem it solves, and what the first version must prove.
Once that is written, the project becomes smaller, but also more real. It becomes easier to see which features are decoration, which assumptions need to be tested first, and which parts can remain manual instead of becoming a system too early.
Independent products are not mainly limited by resources. They are limited by attention. Writing the problem clearly is a way to protect that attention.
Comments