Good market research questions each feed one business decision: is there demand, how big is the market, who buys, what do they use today, what will they pay, and which version wins. This page lists 24 questions for market research under those six decisions, with the answer options and the mistake each one usually ships with.
By the Wavefield Research team · Published Nov 11, 2026
Most lists of market research questions are sorted by topic: demographics, brand, product. That order makes it easy to write a long survey and hard to say what any answer is for. The U.S. Small Business Administration frames market research around six things a business needs to know: demand, market size, economic indicators, location, market saturation and pricing. The bank below follows that frame, folds the two profile items into one group about buyers, and adds a sixth group for testing a concept.
The test for every question is the same. Before it goes in, write down the decision it informs and which answer would change that decision. A question that survives the test earns its place. One that fails it is a question someone was curious about, and curiosity is how a ten-question survey becomes forty. The method side of the work, from choosing a sample to reading the tables, is in our guide to market research surveys.
Each group answers one decision. Take the groups your decision needs and leave the rest: a pricing study has no use for the concept group, and a demand check can stop after the first two. Replace the bracketed words with your own category, product and price.
Decision: whether to build or sell this at all.
“In the past three months, how often have you [had the problem]? (Never / Once or twice / About monthly / About weekly / Daily)”
Problem frequency is the base of every demand estimate. Name a fixed recent window, because 'how often do you usually' invites a flattering average.
“The last time it happened, what did you do about it?”
An open-end you code into three piles: did nothing, used a workaround, paid for something. People who did nothing are telling you the problem is tolerable.
“How much of a problem is [issue] for you? (Not a problem / A minor annoyance / A real problem / One of my biggest problems)”
Severity belongs next to frequency. A daily minor annoyance and a yearly emergency are two different businesses.
“In the past year, have you spent money or more than an hour trying to solve it? (Yes, money / Yes, time only / No)”
Past spending is the strongest demand evidence a survey can collect, and it takes one question. Opinions about the problem cost the respondent nothing; this answer did.
Decision: whether the market is large enough to be worth entering.
“Which of these have you done in the past 12 months? (Select all that apply: a list of six to eight activities, one of them yours)”
The incidence question. The share who pick your activity, multiplied by the population the sample represents, is the market in people. Hide the target among other activities so nobody can guess the answer you want.
“How many times did you buy [category] in the past 12 months? (0 / 1 / 2 to 5 / 6 to 11 / 12 or more)”
Purchase frequency turns people into volume. Offer ranges, because nobody remembers an exact count and a blank box collects guesses.
“About how much did you spend the last time you bought [category]? (Under $25 / $25 to $49 / $50 to $99 / $100 to $249 / $250 or more)”
Spend per purchase. Incidence times frequency times spend is the market in dollars, built from three answers people can actually recall.
“Who makes the final decision to buy [category] in your household or company? (I do / I share the decision / Someone else does)”
Separates the people who use the thing from the people who pay for it. In business markets, skipping this question means interviewing the wrong job title for the whole study.
Decision: which segment to target first, and where to reach it.
“Which of these best describes your household income before taxes last year? (Ranges that match the census brackets, plus Prefer not to say)”
Matching the census brackets lets you size a segment against published population counts. Ask it near the end, and always offer a way to decline.
“What are the first three characters of your ZIP or postal code?”
Location at the precision you need and no finer. A full address in a market survey costs you completes and tells you nothing the first three characters do not.
“How many people work at your company? (Just me / 2 to 9 / 10 to 49 / 50 to 249 / 250 or more)”
The business-market equivalent of household income. Company size predicts budget, buying process and how long a sale takes.
“Where do you look first when you need [category]? (A search engine / A specific store or site / Social media / Someone I know / Somewhere else)”
The channel question. It decides where the marketing budget goes, so randomize the option order to keep the first choice from winning by position.
Decision: who you are really competing with, and how hard they are to displace.
“Which of these have you heard of? (Select all that apply: a randomized brand list that includes yours, plus None of these)”
Aided awareness. Some researchers add one made-up brand to the list; the share who claim to know it shows how much of everyone else's awareness is politeness.
“Which one do you use most often today? (The same list, plus Something I put together myself and I do not use anything)”
Share of current use. The two extra options matter most: a spreadsheet and doing nothing are competitors, and they are usually the biggest ones.
“How satisfied are you with what you use today? (Very dissatisfied / Dissatisfied / Neither / Satisfied / Very satisfied)”
Satisfied users of a rival are expensive to win. Your first customers are in the bottom two boxes, so this question is really a segment finder.
“What would have to be true for you to switch from what you use today?”
An open-end about switching triggers. Three respondents naming the same trigger outrank any feature ranking you could ask for.
Decision: the launch price, or whether the price you need is one the market accepts.
“What do you pay now for [current alternative]? (Nothing / Under $10 a month / $10 to $24 / $25 to $49 / $50 to $99 / $100 or more)”
The anchor. What people already pay for the alternative bounds every other price answer, and it is a fact they can report without imagining anything.
“At what price would [product] start to feel expensive, but still worth considering?”
One of the four Van Westendorp price questions. Asked as a set, the four answers give an acceptable price range without ever showing the respondent a price.
“At what price would [product] be too expensive to consider at all?”
The upper bound from the same set. On its own it shows where demand falls off a cliff, which is the number a pricing meeting most needs.
“How likely would you be to buy [product] at [price]? (Definitely would not / Probably would not / Might or might not / Probably would / Definitely would)”
Purchase intent at a named price. Show each respondent one price and compare the groups; asking one person about three prices teaches them to bargain.
Decision: which idea, feature set or message goes forward.
“After reading this description, how interested are you in [concept]? (Not at all / Slightly / Moderately / Very / Extremely)”
The core concept score. Show each respondent a single concept and compare across groups, so nobody is rating the second idea against their memory of the first.
“What is the one thing you like most about it?”
Tells you which sentence of the description is doing the work. It is often not the feature the team is proudest of.
“What, if anything, is unclear or hard to believe?”
The believability check. A concept can score well on interest while half the sample doubts the main claim, and that doubt is cheap to find here and expensive to find after launch.
“How different is this from what is already available? (Not at all different / Slightly / Somewhat / Very / Completely different)”
Uniqueness. High interest with low uniqueness describes a product people want and can already buy somewhere else.
Three of these groups have a page of their own: the full four-question price set is on Van Westendorp pricing, the awareness questions are on brand awareness surveys, and the profile questions are on demographic survey questions.
The questions above only work on the right people, and potential customers are harder to find than existing ones. Market research questions for customers you already have can go to your own list. Questions for people who have never bought from you need a screener first: one or two questions that confirm the respondent has the problem or buys the category, asked before anything else and worded so the qualifying answer is not obvious. How to write those is covered in survey screening questions.
A potential customer has no experience of your product, so every question about it is a guess. Questions about the last time the problem came up are testimony. Fill the survey with testimony and keep the guesses to the last three questions.
Write the concept description the way a neutral stranger would: what it is, what it does, what it costs. Sales copy raises the interest score and tells you how persuasive your copywriter is.
People on your mailing list already like you. Their answers describe your fans. For a market estimate, the sample has to include people who have never heard of you, in roughly the proportions they exist.
Small businesses asking these questions for the first time will find the cheaper routes to a sample in market research for small business.
The best market research questions still produce answers of unequal value. A report that prints a purchase-intent percentage next to a past-purchase percentage, in the same font, is treating a guess and a fact as the same kind of number. Sort the answers before anyone reads them.
| Kind of answer | Trust | How to read it |
|---|---|---|
| Past behavior (bought, spent, switched, did nothing) | Highest | Report as shares with a margin of error. This is the evidence to build the market size on. |
| Awareness and current use | Fairly high | Brand lists inflate recognition a little. Compare brands against each other before trusting any single level. |
| Attitudes (severity, satisfaction, interest) | Good for comparisons | Compare segments or concepts measured the same way. The absolute level means little without a benchmark. |
| Purchase intent | Lowest | Stated intent runs ahead of real purchases, because saying yes in a survey is free. Use it to rank concepts or prices, never as a sales forecast. |
| Open-ended price guesses | Low | People asked to name a price name a low one. Ask the Van Westendorp set, or show one price per respondent. |
The second limit is arithmetic. Market research is read by segment, and segments are small. With 50 respondents in each of two segments, a gap between them has to be about 20 percentage points wide before it can be told apart from sampling noise at the usual 95 percent confidence. With 400 in each, about 7 points is enough. So decide which segments you need to compare before you field, and size the study for the smallest one with the sample size calculator. The tables that show those gaps, and the significance tests on them, are explained in cross-tabulation.
Ten questions from the bank, ordered the way they should be asked. The order matters as much as the wording. Problem questions come before the concept so the description cannot shape what people recall, price comes after the concept so there is something to price, and the profile question comes last.
Two wording rules do most of the work in a questionnaire like this one. Pew Research Center advises asking one question at a time, because a question that bundles two ideas is hard to answer and hard to interpret, and randomizing the order of listed options, which does not remove order effects but spreads them evenly. The first rule protects the price and concept questions. The second protects every brand list and channel list in the bank. For the open-ended questions, the craft is in open-ended survey questions.
Good market research questions each feed one named decision and ask about something the respondent can actually report. The strongest ones are about past behavior: how often the problem came up, what the person did about it, what they use today and what they pay for it. Good questions for market research also come with a fixed time window, labeled answer options and a way to say none of these.
Start by confirming they are potential customers: a screening question about what they have done or bought in the past year. Then ask how often the problem comes up, what they did the last time, what they use today, how satisfied they are with it, and what they pay. Show your idea only after those answers are in, and finish with interest, believability and purchase intent at one price.
Ten to fifteen for a study with one decision behind it, which takes about five minutes to answer. Each added decision adds a group of about four questions. Long surveys lose people part way through, and the ones who finish are the most patient respondents, which is a bias of its own. Two short surveys usually serve better than one long one.
For a new product the best market research questions are the demand group and the concept group: how often the problem happens, whether the person has already spent money or time on it, how interested they are after reading a plain description, what they find hard to believe, and how likely they would be to buy at a named price. Treat the intent answer as a ranking tool for comparing versions.
Survey questions for market research are mostly closed: fixed answer options asked the same way of a few hundred people, so the answers can be counted and compared between segments. Interview questions are open and asked of a handful of people to learn the vocabulary and the reasons. Interviews come first, because they supply the answer options the survey needs.
Yes. The ten-question starter on this page is a complete sample questionnaire for market research: one screener, three questions about the problem, two about the current alternative, three about your concept and one profile question. Replace the bracketed words with your category and price, keep the order, and you have a questionnaire that can be fielded as written.
Related: market research methods · product survey questions · concept testing software · survey question examples
Tell the agent the decision and the market. It drafts the screener, the questions, the answer options and the skip logic for you to edit. Building and previewing is free; fielding starts at $99 per study.
Build a market research survey free