What questions should you ask in a concept test?
TL;DR
- Screen first, then show the concept, then ask reaction, comparison, intent, substitution and one open question.
- Answer scales change the answers. In one classic experiment, the share reporting heavy TV viewing more than doubled when only the scale range changed.
- Cut every question that doesn't change a decision. Long tests produce tired, flat answers.
- Neon Blue runs the same questionnaire on a simulated audience and returns scores and reasons by segment, labeled as simulated. Run your first study free
Neon Blue builds simulations of human behavior. A good concept test asks about ten questions in a fixed order: who the person is, how the idea lands, how it compares with what they use, whether they would buy it, and what they would change. Each question should map to a decision your team will make. Below is a questionnaire you can copy, with the reason each item is there and the wording mistakes that quietly bend the answers. Neon Blue fields it on a simulated audience of your buyers.
What is a good concept test questionnaire?
- In the past 3 months, which of these have you bought? [Category list plus "none of these"]
- How often do you buy [category]? [Never, less than monthly, monthly, weekly, more than weekly]
- How much do you like this idea overall? [Dislike a lot, dislike, neutral, like, like a lot]
- How believable is what it says it does? [Not at all to completely, 5 points]
- Compared with what you use now, is it better, about the same, or worse?
- What, if anything, is new to you here? [Open]
- At $[price], how likely are you to buy it? [Definitely not to definitely, 5 points]
- If you bought it, what would it replace? [Products used now, plus "it would be an extra"]
- Who in your household would use it? [Self, partner, kids, other]
- What is the one thing that would stop you from buying it? [Open]
| Question | Why it is there | Decision it feeds |
|---|---|---|
| 1 and 2 | Defines the segment you read results by | Which audience to target |
| 3 and 4 | Separates "nice idea" from "I believe it" | Whether the claim needs proof |
| 5 and 6 | Tests whether it is meaningfully different | Positioning |
| 7 | Intent at a stated price | Go or no-go, price band |
| 8 | Shows where volume comes from | Cannibalization risk |
| 9 | Finds the real user, who may not be the buyer | Pack size and message |
| 10 | Names the barrier in the buyer's words | What to fix before launch |
Questions 1 and 2 should match the segments you plan to target, and question 7 feeds straight into a launch go or no-go call.
Why does the wording of answer scales matter so much?
In 1985, Norbert Schwarz and colleagues reported an experiment in Public Opinion Quarterly. They asked German adults how much TV they watched a day, using one of two answer scales. When the scale ran from "up to half an hour" to "more than two and a half hours", 16.2% picked more than two and a half hours. When the scale started at "up to two and a half hours", 37.5% reported watching that much. Same people, same behavior, different answer box.
People treat the scale as a hint about what is normal. In a concept test, that shows up in three places:
- Frequency questions. Set ranges from real category data, not from a guess about your best customer.
- Price. Showing the price before the description makes it the frame for everything after. Show it once, at the intent question.
- Lopsided scales. "Good, very good, excellent" has no room for a negative. Use balanced scales.
Which concept test questions should you cut?
- Double questions. "How appealing and useful is this?" mixes two answers in one score.
- Leading questions. "How much would this new, healthier option improve your routine?" assumes the answer.
- Feature wish lists. "What features would you like?" produces long lists nobody will pay for. If you need feature priorities, use a ranked task.
- Willingness to pay as an open number. It reads high. Use a price ladder, explained in measuring willingness to pay.
What does a concept test read look like on a simulated audience?
Hypothetical: a personal care brand has a refillable deodorant concept and wants to know if the refill story is believable and what it replaces.
- Audience: US adults who bought deodorant in the past 3 months, split by those who buy natural or clean-label products and those who don't, and by age.
- Design: the ten questions above, one concept per respondent, with the brand's current best-seller as a control.
- Readout: liking, believability, intent at the stated price, what it replaces, and the main barrier, by segment.
Example output layout (illustrative, not a real result)
| Segment | Like a lot or like | Believable (top 2) | Would replace current brand | Main barrier (coded) |
|---|---|---|---|---|
| Clean-label buyers | __% | __% | __% | Coded theme |
| Mainstream buyers | __% | __% | __% | Coded theme |
| Age 18 to 34 | __% | __% | __% | Coded theme |
If believability is low but liking is high, the concept needs proof, not a new idea. That is the moment to bring in a few live interviews. To read the intent numbers properly, see how to measure purchase intent.
Questions about concept test questionnaires
How many questions should a concept test have?
About ten for one concept. Past 15, answers flatten as people tire, and open-ended answers get shorter.
Should the price be shown in a concept test?
Show it if the decision involves price. Put it at the intent question, not in the headline, so it does not color every other answer.
What is the most important concept test question?
The open barrier question. Scores tell you whether there is a problem. The barrier, in the buyer's own words, tells you what to fix.
Can Neon Blue use our existing concept test questionnaire?
Yes. Neon Blue fields your questionnaire on a simulated audience of your buyers so the read lines up with your past studies. Start a free first study.