Prototyping

We build prototypes to settle arguments. Something clickable, in front of real people, early enough that changing course costs a week rather than a quarter.

The point is never the artefact. It is the question: will people understand this, will they trust it, will they finish it, is it worth building at all.

One question per prototype

A prototype built to test everything tests nothing. We decide the question first, then build only as much fidelity as answering it requires, which is often far less than expected.

Sometimes that is a clickable flow, sometimes a single screen, sometimes a fake front end over a manual process, which is the fastest way to learn whether a service is wanted before any of it is automated.

Tested with people who are not you

Five or six of the right participants surface most of what is wrong. We run the sessions, watch where people hesitate rather than ask them what they think, and report what changed our mind.

The output is a decision with evidence behind it: build this, cut that, or do not build it at all. That last answer is the most valuable one a prototype can produce, and the hardest to reach any other way.

Have a question worth answering before you build?

Let's talk.