The Product Learning Cycle

A sequence of questions

Building a new product is a sequence of questions, each answered with evidence before the next one is worth asking:

1. Is the problem real and important? If not, no product will succeed.
2. What do we believe about the solution? Written as hypotheses that could be wrong.
3. What is the smallest thing that tests the riskiest belief? This is the minimum viable product (MVP).
4. What did we learn, and what do we change? This is iteration.
5. Do customers stay, use it and tell others? This is product-market fit.

This course follows that sequence. Teams that skip a step usually pay for it later. A team that builds before checking the problem may produce an excellent product that nobody needs.

Which risk is largest?

Every product carries all four risks, but usually one is much larger than the others. The first work should go to the largest.

Example 1: a spare-parts ordering app for small garages. The technology is ordinary: a catalogue, a basket and a delivery booking. Feasibility risk is low. The big question is desirability: will busy mechanics stop phoning their usual supplier and use an app instead? The first test should focus on behaviour change, not on building a perfect catalogue.

Example 2: a system that predicts pump failures from vibration data. Maintenance managers clearly want fewer breakdowns, so desirability risk is lower. The big question is feasibility: can the model predict failures accurately enough, early enough, from the data sites actually have? The first test should be a technical proof using real historical data, with no interface at all.

The same team, using the same method, would build two completely different first products.