Research before you name, name before you brand

The common failure is starting with the name. It is the most fun part, so it is the part people do first, and then the research arrives later and contradicts it. By then the name feels like a decision already made.

Research first

Define the market narrowly — a topic, a target market, a niche, and the phrases a real buyer would type. Then search the live web for evidence that the pain exists and that someone pays to solve it. What comes back is not an idea yet, it is evidence.

Only then generate concepts, grounded in that evidence, and score them on revenue potential, starting cost and difficulty. Five of them, kept, so you can see what you rejected.

Then the landscape

Before naming anything, map who is already there. This changes names: you will not pick something adjacent to an incumbent once you know the incumbent exists. It also changes the concept more often than people expect.

Then the name

Now naming is constrained by something real — the audience, the job, the competitors to avoid sounding like. Every candidate is domain-verified, so the list you choose from is a list you can actually act on.

Then the brand, then the build

The identity follows the name, the repository follows the identity. By the time code exists, the README can state the opportunity, the estimates and their provenance, because all of that was established before anyone opened an editor.

Each step is cheap to redo and expensive to skip. Reversing the order does not save time, it just moves the cost later.