By Vellora Labs · Updated October 8, 2026
A concept test helps you see what people think your idea does, where it fits their work, and what they still need to know. Use it when a proposition or workflow is uncertain and building the full product would be premature.
Write a decision, not a validation goal.
“Validate our AI assistant” leaves too much room for interpreting any compliment as success. A more useful decision is: “Should the first screen lead with reviewing a summary or assigning follow-up actions?” Record the alternatives, the audience, and the observation that would change your mind.
Worked example — fictional: a team is considering a meeting follow-up product. It wants to learn whether product managers understand that they must review suggested actions before sharing them. There are no participant results behind this example.
Prepare one stimulus and its evidence boundary.
Use a short description, storyboard, or product video. Label prototypes and simulated features. Keep a copy of the exact version shown. List what the product demonstrably does separately from planned capabilities. Avoid a narration that answers the very question you want to test.
A product video brief can keep the audience, one message, approved assets, and claims together. For a first round, choose one concept per person. If you show multiple concepts, vary their order and record it; the first can change how people understand the next.
Recruit around the task.
Choose people who have recently done the work the idea addresses. For the fictional assistant, job title alone would not establish experience with meeting follow-up. Ask about recent actions, responsibility, and the existing workaround. Use the screener builder to draft questions and review exceptions manually.
Include relevant differences in experience and tools. State who is absent from your sample. Set your session target and budget as planning assumptions, not as a guarantee that every important issue will appear.
A 20-minute session outline
- 0–4 minutes: context. Ask for the last relevant task, what triggered it, and what happened. Confirm consent and whether recording is permitted.
- 4–6 minutes: exposure. Show the stimulus without a sales introduction. Record whether the participant replays, pauses, or asks for help.
- 6–11 minutes: comprehension. Ask them to explain the concept to a colleague. Probe what they think happens automatically and what they must do themselves.
- 11–16 minutes: fit. Return to the real task they described. Where could the idea fit? What would prevent it? What would they keep using?
- 16–20 minutes: unresolved questions. Ask what evidence they would need before trying it. Close without promising features or access you cannot provide.
Record the observation before the interpretation.
| Capture | Example field | Why it matters |
|---|---|---|
| Stimulus | Version and scene timestamp | Connects the issue to what was shown |
| Evidence | Participant’s exact words or observed action | Keeps your interpretation separate |
| Classification | Misunderstanding, missing proof, or workflow mismatch | Suggests different changes |
| Alternative explanation | Unfamiliar term, task mismatch, moderator hint | Prevents a premature conclusion |
Make a bounded decision.
If an important capability is misunderstood, revise the explanation and test it again. If people understand the concept but cannot connect it to a recent need, investigate fit before rewriting the pitch. If they understand and identify a relevant task, invite a realistic next-step test with clear consent. Interest alone does not establish adoption, payment, or product-market fit.
Choose a decision rule before sessions, based on the risk of being wrong. For example, any observed misunderstanding about automatic sharing might justify revising that scene. This is a risk-based editorial choice, not a statistical threshold or a result from this example.
Continue the workflow.
Test a video explanation when the stimulus is a video, or build a plan before recruiting.
Sources and scope
Maze’s concept-testing survey guide provides background on clear test goals, neutral framing, and exposure methods. The session outline and example here are Vellora’s suggested workflow, not a tested client outcome. Source checked October 8, 2026.