Concept testing questions fall into five groups: comprehension (does the idea make sense?), appeal (do people want it?), relevance (is it for them?), differentiation (is it better than what exists?), and purchase intent (would they actually buy it?). Ask at least one from each group and you can tell not just whether a concept works, but why it does or doesn’t.
Most concept testing guides stop at the question list, because that’s the easy part. The decisions that actually determine whether your results are usable come before and after: how you show concepts to respondents, and how you analyze what comes back. This guide covers all three.
The 25 concept testing questions
Use these as written or adapt the wording. The grouping matters more than the exact phrasing — a concept test that only measures appeal tells you which idea won, but not what to fix.
Comprehension: does the concept land?
- In your own words, what is this product?
- Who do you think this product is for?
- How easy or difficult was this concept to understand?
- Was there anything confusing or unclear about it?
- What questions would you want answered before deciding whether to buy it?
Comprehension questions come first for a reason. If a third of respondents can’t describe the concept back to you, every score that follows is measuring their confusion rather than the idea.
Appeal: do people want it?
- Overall, how appealing is this concept to you?
- What do you like most about it?
- What do you like least about it?
- How interested would you be in learning more?
- How does this concept make you feel?
Relevance: is it for them?
- How well does this product fit your needs?
- How often would you use a product like this?
- In what situations would you use it?
- Does this solve a problem you actually have?
- How does this compare to how you currently handle that problem?
Differentiation: is it better than what exists?
- How different is this from products already available?
- What makes it different, in your view?
- Which existing product would you give up to use this one?
- How believable are the claims made about this product?
- What would make you choose this over the alternative you use now?
Purchase intent and pricing
- How likely are you to buy this product if it were available?
- What would you expect to pay for it?
- At what price would this be too expensive to consider?
- At what price would it be so cheap you’d question the quality?
- If this were available tomorrow, what would stop you from buying it?
Questions 23 and 24 are the anchors of a Van Westendorp price sensitivity meter — including them turns a concept test into a light pricing read at no extra fieldwork cost.
Choose your concept testing design first
The questions matter less than how you show the concepts. There are three standard designs, and picking the wrong one is the most common way a concept testing survey produces unusable data.
Monadic testing shows each respondent one concept only. Because nobody sees a comparison, their reaction is uncontaminated by contrast effects — the cleanest read on how a concept performs in the real world, where consumers don’t see your alternatives side by side. The cost is sample: with monadic testing you need a full sample cell for every concept, so testing six concepts means six times the fieldwork.
Sequential monadic shows each respondent several concepts, one at a time, in randomized order. It’s the pragmatic middle ground — far cheaper than pure monadic, and randomization spreads order bias across the sample rather than letting it favor whichever concept came first. The trade-off is that later ratings are influenced by earlier ones, so absolute scores drift.
Comparative testing shows concepts together and asks respondents to choose or rank. It produces the sharpest discrimination between concepts and the least realistic reflection of how people actually encounter products.
The rule of thumb: use monadic when the absolute score matters (will this concept clear our launch threshold?), sequential monadic when you’re screening several ideas on a budget, and comparative when you only need to know which of a shortlist wins.
How to analyze concept testing results
This is where concept testing research either earns its budget or becomes a deck nobody acts on.
Start with significance, not averages. Concept A scoring 4.1 and Concept B scoring 3.9 is not a result — it’s a rounding difference until statistical testing says otherwise. Every comparison in a concept test needs significance testing applied, including comparisons between subgroups.
Cut by the audience that matters. A concept that underperforms overall but wins decisively among your target segment is a different decision from one that fails everywhere. Concept tests are usually fielded to a broad sample precisely so you can cut them, so cut them.
Read the open-ends properly. Questions 4, 7, 8, and 17 generate the verbatims that explain the scores, and they are the most-skipped part of concept testing analysis because coding them by hand is slow. AI text analysis codes them into themes automatically, which turns “Concept C scored badly” into “Concept C scored badly because people didn’t believe the durability claim.”
Prioritize features, don’t just rank concepts. When your test includes a list of features, claims, or benefits, MaxDiff produces a far more discriminating ranking than rating scales, which tend to compress everything into “somewhat important.” And when you’re choosing a portfolio rather than a single winner, TURF analysis finds the combination that reaches the most people with the least overlap — often a different answer from simply picking the top three scorers.
Set the reporting up once. Concept testing rarely happens once. If your analysis rebuilds itself when the next wave of concepts is fielded, round two costs hours rather than days.
Claims testing and message testing use the same method
Concept validation testing, claims testing, and message testing are the same research design with a different stimulus. Instead of a product concept, respondents see a claim (“clinically proven to last 24 hours”) or a marketing message, and the questions shift toward believability and persuasion:
- How believable is this claim?
- How relevant is it to you?
- Does it make you more or less likely to consider the brand?
- What does this claim tell you about the product?
Everything else — monadic versus sequential monadic, significance testing, open-end coding — carries across unchanged. If you already run concept tests, claims testing costs you a questionnaire, not a new methodology.
Frequently asked questions
What are concept testing questions?
Concept testing questions are survey questions used to evaluate a new product, service, or idea before launch. They measure five things: whether people understand the concept, whether it appeals to them, whether it’s relevant to their needs, how different it is from existing options, and whether they’d buy it. A complete concept testing survey covers all five rather than only measuring appeal.
How many concepts can you test in one survey?
With sequential monadic testing, most researchers show three to five concepts per respondent before fatigue degrades the data. Beyond that, either split the concepts across sample cells or use a monadic design with a larger total sample. Testing ten concepts in one sitting produces data that mostly measures how tired respondents were by concept eight.
What is monadic testing?
Monadic testing shows each respondent a single concept, so their evaluation isn’t influenced by seeing alternatives. It gives the most realistic read on absolute performance – the way a consumer would actually encounter the product – but requires a separate sample cell for each concept, making it the most expensive design. Sequential monadic, which shows several concepts one at a time in random order, is the common compromise.
What sample size do you need for concept testing?
A common working minimum is 100 to 150 respondents per concept cell, which is enough to detect meaningful differences on rating scales. If you plan to analyze subgroups – and you usually should – size the cells so each subgroup you care about still has around 100 respondents, not the total.
Can you provide an example of concept testing?
A drinks brand testing three new flavors would field a sequential monadic survey: each respondent sees all three in random order, rates each on appeal, relevance, uniqueness, and purchase intent, and answers open-ends on what they liked and disliked. Analysis then tests the differences for significance, cuts results by category buyers versus non-buyers, and codes the verbatims to explain why the losing flavor lost.
What’s the difference between concept testing and usability testing?
Concept testing evaluates whether an idea is worth building, using survey research with a representative sample. Usability testing evaluates whether a built product is easy to use, typically with a small number of participants observed completing tasks. Concept testing answers “should we make this?”; usability testing answers “can people use what we made?”
The bottom line
Good concept testing questions are necessary but not sufficient. The design decision — monadic, sequential monadic, or comparative — determines whether your numbers mean anything, and the analysis determines whether anyone acts on them. Get all three right and a concept test stops being a scorecard and becomes a decision.
