Prototype testing turns a hunch into something people can try
A prototype test puts an early version of a product in front of the people it is for. You give them a task or a focused question, then look at how they respond before the expensive parts become expensive.
It works with rough screens, clickable designs, and working web prototypes. The fidelity can change. The job stays the same: learn what the next version needs.
The point is not to prove the prototype is good. It is to find out what the next version needs.
Start with the decision, not the questionnaire
A useful prototype test is narrow enough to change something. Name the decision first, then choose the lightest method that can answer it.
| Decision you need to make | Method to use | What it gives you |
|---|---|---|
| Can people find where to begin? | First-click test | The first place people choose on a screen |
| Does the screen communicate what you meant? | Five-second test | What people notice and remember |
| Which direction makes more sense? | Preference test | A choice, followed by the reason behind it |
| Can people complete the important path? | Live prototype task | Captured task results and follow-up answers |
| Why did people hesitate or stop? | Open text and rating questions | Their explanation in their own words, alongside the score pattern |
How to test a prototype
1. Name the next decision
Write the decision in plain language. Are you choosing between two directions, fixing a confusing path, or checking whether the main task makes sense?
2. Use the lightest prototype that can answer it
A still screen can answer a first-click question. A clickable flow is useful when sequence matters. A working prototype earns its keep when people need to complete a real task.
3. Build one focused study
Set the context, give one clear task at a time, and ask only the follow-up questions that help explain the behaviour. More questions do not rescue a vague decision.
4. Preview the whole route
Take the study as a respondent would. Check the prototype link, task wording, follow-up questions, and every ending before anyone else sees it.
5. Launch and share
Launching freezes the version respondents will see and gives you a link to share. Respondents do not need a Fiuto account. Fiuto does not recruit them, so send the link to your customers, waitlist, community, panel, or other people you can reach.
6. Read the evidence by method
Fiuto organises results around each question and task. Look for the places where behaviour and explanation agree, the places where they pull apart, and the uncertainty that needs another test.
Use the prototype you already have
No ceremonial rebuild required. Choose the setup that fits what exists today.
A screen or screenshot
Use a first-click test when the first move matters. Use a five-second test when the message or hierarchy matters. Use an image-click test when you want people to react directly to a static design.
A clickable design or public prototype
Use a live prototype task to open a public browser URL inside the study. Give respondents a clear task, then add the follow-up questions that explain what happened. If recording is used, Fiuto asks for consent before capture starts.
An agent-built web prototype
If a coding agent built the prototype and can edit its code, it can install Fiuto tracking and attach the prototype to a Fiuto launch. Tracker events are reviewed separately from individual respondents’ study answers.
Prototype testing is one part of MVP validation
Prototype testing can show how people understand and use an early experience, and it can capture what they say about it. MVP validation is wider. Demand, pricing, acquisition, and repeat use may need evidence of their own.
Do not make one prototype session carry every job. Use the test to answer the product question in front of you, then pair it with customer and market evidence when the business question is bigger.
One study, whichever way you build
In the web app
Build the study from question and task blocks, preview it, launch it, share the respondent link, and read the results by method. Use this route when you want to work directly in Fiuto.
From an MCP-ready builder
If you are already working in Claude Code, Codex, Cursor, Lovable, v0, or Replit, ask the builder to create and launch the Fiuto study from where the prototype is being built. The prototype stays in front of you and the study follows the same create, preview, launch, share, and review loop.
Frequently asked questions
01What is prototype testing?
Prototype testing puts an early version of a product in front of real people so you can see how they understand it, use it, and explain their choices before you build the next version.
02How do you test a prototype?
Name the decision, choose the lightest prototype that can answer it, build a focused study, preview the full route, launch it, share the respondent link, and review the evidence by question and task.
03What should I test first?
Test the part that is both important and uncertain. That might be the first step in a flow, the meaning of a screen, the path through a key task, or the choice between two directions.
04Can I test a Figma prototype or a live website?
Yes. Use image and click methods for static screens, or a live prototype task for a public browser URL. Choose the method around the question, not around how polished the prototype looks.
05Does Fiuto recruit testers?
No. You bring your own respondents, such as customers, a waitlist, a community, or a panel. Fiuto gives you the study and the respondent link.
06Do respondents need a Fiuto account?
No. Anyone with the respondent link can take the launched study without a Fiuto account.
07Can my coding agent create the prototype test?
Yes. An MCP-ready builder can author and launch a Fiuto study from where the prototype is being built. If it can edit a live web prototype, it can also install Fiuto tracking.
08Is prototype testing the same as MVP validation?
No. Prototype testing is one input to MVP validation. It can show how people understand and use an early experience, but wider questions about demand, pricing, acquisition, and repeat use need additional evidence.
09What results does Fiuto show?
Fiuto shows a result card for each question or task in the launched study. The shape follows the method, such as counts, distributions, written answers, tested-asset results, or captured live-task results.
