The useful question
People like a statistic about how many products fail. This article does not use one. The useful question is narrower: what must be true in 90 days for the next three months of building to be worth paying for?
The address of this page keeps a phrase readers search for. The claim itself, that “most MVPs fail,” is not a finding we are willing to print without a source we would stand behind.
One user, one job, one number
Name the user tightly enough that you could list ten of them. Name the job as a sentence with a start and a finish. Pick one number: activation, a paid invoice, or weekly use. Do not pick all three and then celebrate whichever moves.
A marketing site can ship in the same quarter. It should describe the job the application actually does. A site that promises a platform, sitting on an application that completes one task, trains the wrong buyer.
Patterns that stall a build
These are patterns from delivery work, not a census.
- No specific buyer, so every feature request sounds equal.
- The workflow was guessed in a workshop and never watched.
- Distribution was left until the software felt finished.
- A platform was designed before a single path worked end to end.
What the 90 days should produce
A path a real user completed, notes on where they stalled, and a decision. Continue, narrow, or stop. Stopping is a successful use of the 90 days when the number did not move and you can say why.
Scope that does not serve that decision is cost, even if it looks like progress in a demo.
Related service: Website development