Column · what not to build

How to decide what not to build

2026-10-19 · InfoGap

The three filters

  1. Does it serve the one sentence the product promises?
  2. Has a specific person reported the failure it addresses?
  3. Can it be removed later without breaking the promise?

Why the third matters

Because reversible decisions should be made quickly and irreversible ones slowly. Anything that can be removed without breaking the promise is reversible and does not deserve deliberation.

The thing that is hardest to refuse

The request from a paying customer. It is real demand and it is also a single data point. Build it if it serves the sentence; decline if it does not, and say why.

The cost of saying yes

Each accepted request becomes part of the product's surface area, and surface area is what makes the next change expensive.

See the data this is built on

Open the snapshot

map · column · data · pricing