The easiest part of a new idea to build is not always the most important part to test. You can spend a week making a beautiful interface while avoiding the uncomfortable question of whether anyone understands what the product is for.
A prototype becomes useful when it encounters the uncertainty that matters. Its job is to produce evidence that could alter a decision. It is allowed to be incomplete because completeness is not the question yet.
GOV.UK's prototyping guidance describes approaches with different levels of fidelity, including prototypes that support realistic interactions. The useful choice depends on what you need to learn, not how impressive the artifact will look in a progress update.
Name the uncertainty before choosing the format
Write the assumption that would make the project difficult to justify if it proved wrong. Perhaps intended users cannot explain the offer. Perhaps the workflow needs information they do not have. Perhaps delivery requires more specialist effort than the price can support.
Use a sentence with a condition: “This idea depends on a first-time user being able to complete the task without a facilitator.” Now the prototype has a job. It must expose the parts of the task where a person needs help.
Avoid trying to answer every question with one test. Comprehension, usability, willingness to pay and long-term value are related but different uncertainties.
Match the artifact to the question
Consider these possible pairings:
| Uncertainty | A prototype that could help |
|---|---|
| Do people understand the offer? | A short description and a sample outcome. |
| Can they complete the task? | A clickable or working version of the relevant steps. |
| Does the result support a real decision? | A manually produced example used in a realistic scenario. |
| Can we deliver it consistently? | A limited internal delivery rehearsal with recorded effort. |
A static picture cannot establish that a complex interaction works. A working interaction cannot establish that someone will return next month. Keep the conclusion within the evidence the artifact can provide.
Define what you will observe
For a fictional assessment product, a useful question might be whether intended buyers can interpret a priority recommendation. Show a sample, explain that it uses fictional answers, and ask what action they would take from it.
Observe where they look, what they understand and what they cannot resolve. “They liked the colors” may be valuable design feedback, but it does not answer the interpretation question.
Decide what would lead you to revise before conducting the session. For example, if participants cannot connect a recommendation to their answers, the report may need clearer evidence. Avoid moving the standard afterward so every result supports proceeding.
Keep the test honest about its limits
Tell participants what is real, what is simulated and whether their input will be saved. Do not collect payment for a promised finished product when you are only testing an undisclosed mockup.
If you provide a manual service behind a simple interface, explain the relevant delivery arrangement. Protect private information and obtain permission before recording. A fast learning exercise still involves real people whose time and expectations matter.
Small qualitative sessions can expose problems worth investigating. They do not automatically establish market size or a reliable conversion rate. Record the participant context so you know whose experience you observed.
End with a changed decision or a sharper question
After testing, write three short statements: what we observed, what remains unknown and what we will do next. Sometimes the next step is another prototype. Sometimes it is a narrower offer. Sometimes the useful result is deciding not to build.
Keep the artifact and the notes together. A beautiful prototype with no record of what it taught can become a false sign of progress. A rough prototype with a clear finding can save the team from committing to the wrong work.
For your current idea, identify the question you are most tempted to postpone. Build the smallest honest artifact that can bring that question into contact with reality. That is a more demanding creative act than polishing the parts you already know how to make.
Begin with a clear creative brief if the underlying task is still unresolved.
Questions worth asking
Does a prototype need to look finished?
Only when appearance or realistic interaction is part of the question. A sketch or manual demonstration may be enough for an earlier uncertainty.
What should I do after testing a prototype?
Record what you observed, what it does and does not establish, and whether to revise, continue or stop. Do not treat a positive reaction as proof of demand.
