Usability testing with prototypes: methods and tools
Prototype usability testing means asking real users to complete tasks with an early version of a product, from paper sketches to clickable mock-ups, before it is built. Teams choose the fidelity that suits their question, prepare realistic flows and content, test with moderated or unmoderated methods and measure task success, errors and confidence to decide what to change.
Why test prototypes
A prototype is a simplified version of a product or feature that lets people experience how it might work before it is fully built. Testing prototypes with real users is one of the most cost-effective things a product team can do. Changing a prototype takes minutes or hours; changing a launched product can take weeks and involve engineering, quality assurance and release cycles. Finding problems early saves time, money and frustration.
Prototype testing also helps teams settle debates with evidence. Instead of arguing about which design is better, the team can test both and see which one users understand and complete more easily. Our guide to running a usability test covers the basic session process; this post focuses on what is specific to testing prototypes.
Choosing the right fidelity
| Fidelity | What it looks like | Best for testing |
|---|---|---|
| Paper or sketch | Hand-drawn screens | Early concepts and overall flow |
| Low-fidelity wireframe | Grey boxes and placeholder text | Structure, navigation and content order |
| High-fidelity mock-up | Near-final visuals, clickable | Detailed interactions, labels and visual clarity |
| Coded prototype | Working code with real or sample data | Complex interactions, performance and data-heavy flows |
Lower fidelity prototypes are faster to make and encourage honest feedback about the overall idea, because participants do not focus on visual details. Higher fidelity prototypes are better for testing specific interactions, wording and visual hierarchy. Many teams start low and increase fidelity as their questions become more detailed.
Preparing a prototype for testing
A prototype only needs to support the tasks you plan to test. Build the main path for each task, plus the most likely alternative paths, rather than every possible screen. Use realistic content, names, prices and data wherever possible, since placeholder text can confuse participants and hide problems. Make sure links and interactions behave reliably, and decide in advance how you will handle clicks on areas that are not built yet.
- Build the main path for each task
- Add likely alternative routes and back navigation
- Use realistic text, prices and data
- Check every link and interaction works
- Prepare a polite message for unbuilt areas
- Test the prototype yourself on the target device
Moderated and unmoderated testing
- Moderated in person: A facilitator observes and can ask follow-up questions; best for complex flows and early concepts.
- Moderated remote: The same, via video call and screen sharing; easier to recruit across locations.
- Unmoderated remote: Participants complete tasks alone while software records them; fast and scalable for clear tasks.
- Guerrilla testing: Quick, informal tests with people nearby; useful for very early checks.
Moderated testing allows facilitators to explore why people behave as they do, while unmoderated testing returns results quickly from more participants. Many teams use moderated sessions for early and complex work, and unmoderated tests to validate specific changes before release.
Useful tools
Most design tools can create clickable prototypes. Figma is the most widely used, with links, transitions and interactive components that make realistic flows possible; our Figma basics guide explains how. Specialist prototyping tools handle advanced interactions, sensors and animations. Remote testing platforms let teams recruit participants, run unmoderated tasks, record sessions and collect metrics. Video call tools with screen sharing work well for moderated remote sessions.
| Tool type | Used for |
|---|---|
| Design and prototyping tools | Building clickable flows |
| Advanced prototyping tools | Complex interactions and device features |
| Remote testing platforms | Recruiting, recording and measuring unmoderated tests |
| Video call software | Running moderated remote sessions |
| Research repositories | Storing clips, notes and findings |
What to measure
Prototype tests usually combine observation with simple measures. Task success shows whether people could complete each task. Errors and wrong paths show where the design misled them. Time on task can reveal friction, though prototypes may be slower than real products. Confidence or ease ratings after each task capture how participants felt. Combining these with notes and quotes gives a clear picture of what to fix.
- Task success: Did participants complete the task?
- Errors: Where did they go wrong or hesitate?
- Time: How long did each task take?
- Ease rating: How easy did it feel, on a simple scale?
- Comments: What did they say while thinking aloud?
Limitations to keep in mind
Prototypes are not the real product. They may not show real data, error states, performance issues or how people behave over weeks of use. Participants may also be more forgiving of a prototype than of a finished product. For these reasons, prototype testing should be followed by testing in later stages and by measuring the live product. Our guide to measuring user experience covers post-launch measurement.
Frequently asked questions
What is a prototype in UX?
A simplified, often clickable version of a product or feature used to test ideas with users before building it.
Should we test low or high fidelity prototypes?
Low fidelity is best for early concepts and flow; high fidelity is best for detailed interactions and wording.
How many participants do prototype tests need?
Five to eight per user group usually reveals the main problems in moderated tests; unmoderated tests often use more.
Which tool is best for prototyping?
Figma is the most widely used. The best tool is one your team knows and that supports the interactions you need to test.
Can we test prototypes on mobile devices?
Yes. Most prototyping tools let participants open prototypes on their own phones, which gives more realistic results for mobile products.
What if participants click on parts that are not built?
Note it as a finding, explain politely that this part is not ready and ask what they expected to happen.