Market Research Survey Template

Use this market research survey template when a decision — positioning, pricing bands, feature priority, market entry — is waiting on evidence rather than opinion. It is written for product marketers, insights teams, and founders doing their own research, and every section below is a sample you can copy and change the category terms in.

The methodology overview, including when to choose a tracker over a one-off study, lives in the guide on market research survey types and methods. This page is the instrument.

Before you write a single question

Write these four lines down first. If you cannot fill them in, the survey is not ready.

1. The decision. “We will lead with X instead of Y in Q1 messaging if…” A survey without a decision attached produces a deck nobody opens.
2. The population. Who counts as a qualified respondent, stated tightly enough that a screener can test it. The population of interest covers how to define it without accidentally defining your existing customers.
3. The incidence estimate. What share of your source will qualify. This determines fieldwork cost more than anything else.
4. The cuts you must read. Which subgroups the decision depends on. Size for those — the sample size calculator gives you the completes per cell, and quota sampling covers how to hit them.

The sample questionnaire

Section A — Screener

Short, factual, and never trainable. Respondents learn quickly how to qualify, so avoid stems that announce the right answer.

• In the last 12 months, which of the following have you purchased for your household or business? (multi-select list of 6–8 categories, yours among them, “none of these” required)
• How often do you [use the category]? (daily / weekly / monthly / less often / never)
• Which of the following best describes your role in choosing [category] for your organisation? (sole decision-maker / part of a group decision / I recommend but do not decide / no involvement)
• In which country do you currently live? (single select)

Terminate on the screening criteria silently. Set quota caps with survey quota limits so a cell closes on its own rather than by someone watching a dashboard, and route disqualified respondents straight out with skip logic.

The screener mistake that costs you the study: listing only your own category. If the qualifying question is “do you use project management software,” everybody says yes to be helpful. Embed the category in a list where “none of these” is a normal answer.

Section B — The brand funnel, in this order

Order is not a style choice here. Every question that names a brand contaminates the ones after it, so the unaided items must come first and there is no going back.

1. When you think about [category], which brands come to mind? — open text, no list, no examples in the stem.
2. Which of these brands have you heard of? — the rotated aided list.
3. Which would you consider buying next time?
4. Which have you purchased or used in the last 12 months?
5. Which do you use most often?

Rotate the brand list with random order of questions and answers; an unrotated list reliably inflates the brands at the top. Keep the wording identical wave to wave or the trend line is decorative.

Section C — Usage and behaviour

• How long have you been using [category]? (banded)
• Roughly how much does your organisation spend on [category] per year? (bands, with “prefer not to say”)
• Which of these do you use [category] for? (multi-select, plus “other — please specify”)
• When did you last change provider, and what triggered it? (single choice for timing, free text for the trigger)

Ask behaviour only where you cannot observe it. If your product analytics or CRM already records usage, do not spend a respondent’s attention re-asking it — people mis-remember their own behaviour, and your logs do not.

Section D — Drivers and barriers

• How important is each of the following when choosing a provider? (matrix, 1–5, 8–12 attributes, rotated)
• Which single one of these matters most? (forced single choice from the same list)
• Which of these would stop you from choosing a provider? (multi-select barriers list, “none of these” required)
• What would a provider have to do to win your business? (free text)

The forced single choice earns its place. Importance matrices produce nine attributes rated 4.2 and no ranking; one forced choice separates them. When you need a full priority order, use a ranking question rather than adding more scales.

Section E — Concept or message test (optional)

• Here is a description of a product. How appealing is it to you? (1–5) — one concept per respondent, cells balanced.
• How different is this from what is available today? (1–5)
• How believable is this claim? (1–5)
• What, if anything, is unclear or not credible about it? (free text)

Show one concept per respondent unless you have deliberately designed a sequential test. Showing three in a row measures order effects as much as appeal.

Section F — Price (optional)

Price questions fail when they ask for a single number in a vacuum. Use a band structure, or the four-question Van Westendorp price sensitivity framing:

• At what price would this be so expensive you would not consider it?
• At what price would you consider it expensive but still worth considering?
• At what price would you consider it a bargain?
• At what price would it be so cheap you would question the quality?

Stated price intent is not purchase behaviour in either direction. Treat the output as a range to test, not a price to set.

Section G — Demographics and firmographics, last

Age band, region, company size band, industry, role — only the fields feeding a planned cut, and only fields you are lawfully allowed to collect in each market. Asking these first costs completes; asking them at all without an analysis plan costs goodwill.

Leading questions, before and after

This is where most market research surveys quietly break. Each pair below is the same intent, written twice.

The last row is the one people skip. If your survey introduction says “a short survey from [Brand]”, every unaided awareness answer afterwards is worthless. Keep the intro generic and identify the sponsor on the end page. For the broader failure modes, survey response bias covers acquiescence, social desirability, and order effects.

How to read the results

Base sizes in every cell, every time. A percentage on n=14 is a number, not a finding. Thirty completes is the usual convention before reading a cell at all; a hundred before comparing two. Suppress everything below your stated floor rather than footnoting it — footnotes get cropped when the chart goes into a board deck.

Compare against the decision, not against everything. Run the crosstabs your hypothesis needs. Cross-tabulation will happily generate hundreds of cells, and at conventional significance levels some of them will look meaningful by chance alone.

Name the sample honestly. An online panel is an online panel; a customer list is a customer list. Neither is “the market”. State source, device mix, field window, and incidence next to the headline, and weight only if you can document the assumptions.

Read verbatims systematically. One or two open-ended questions in a quant survey are enough, but they need real coding. Athena clusters them into themes and scores sentiment, which is the difference between quoting three comments you happened to like and reporting what the open text actually said.

Soft launch first. Field 10% of the sample, check the drop-off curve, the straight-liners, and the median completion time, then fix and launch the rest. Testing the survey covers the mechanics; the discipline is worth more than the mechanics.

When a market research survey is the wrong tool

• When you need to know why. Surveys measure how many. If you cannot yet articulate the options, run interviews or a focus group first and survey to size what you learn. Qualitative versus quantitative research covers the sequencing.
• When the decision is already made. Research commissioned to ratify a decision produces a defensible deck and no information.
• When your only sample is your own customers. That is a customer study. It tells you nothing about the people who chose someone else, which is usually the question.
• When incidence is very low. Below a few percent, panel costs rise faster than the value of the answer. Consider a targeted list, a partner’s audience, or expert interviews.
• When behavioural data already answers it. Do not ask people how often they log in.
• When the result would be a regulated claim. In healthcare and finance, a survey finding is not clinical or financial evidence and must not be presented as one.

Where teams take it next

• Use the competitor research survey template when the question is specifically how buyers compare you with named alternatives.
• Use the brand perception survey template when you need attribute association and positioning gaps rather than category sizing.
• Use the consumer preference survey template for assortment, format, and feature trade-offs.
• Use the product-market fit survey template once you have customers and want to know how many would be disappointed without you.

Build the questionnaire in the survey maker or draft a first pass with the AI survey generator and edit it down, field it across email, link, QR, or website intercepts through survey distribution, and see how research teams run the whole cycle on the research teams page. For sizing the segments the study produces, market segmentation is the practical follow-on.