ChatGPT Prompts for Market Research: 12 Templates (2026)

We need to decide [the decision, e.g. whether to launch a cheaper tier for small agencies] by [date]. Turn this into a research plan: the 3 to 5 questions whose answers would change the decision, for each the method that answers it fastest (desk research, customer interviews, a survey, review mining, our own usage data), the sample or sources needed, and what answer would push us each way. Put the questions in the order that lets us stop early if the first answer is decisive.

Build a bottom-up estimate of the market for [product] among [customer type] in [geography]. Start from a countable unit (number of businesses, practitioners, or households), then apply each filter step by step: the share that has the problem, the share that would pay to solve it, and a realistic annual price. Show each step as a line with the number, the source or assumption behind it, and a low and high value. Mark which assumption moves the total most. Do not use top-down industry report totals as the starting point.

Here is the text of [competitor]'s homepage, pricing page, and two feature pages: [paste]. Extract their target customer, the main problem they claim to solve, their pricing structure and the value metric they charge on, the three claims they lean on hardest, and anything they conspicuously avoid mentioning. Then compare with our positioning: [paste ours]. Only use what is in the text I pasted, and mark anything you are inferring.

Here are [N] customer reviews of [product or category] from [source]: [paste]. Group them into the jobs customers hired the product for, the outcomes they praise, and the complaints, with a count for each group. Quote the exact phrases customers use for each, since I will reuse their words in copy. List complaints that appear across more than one competitor, because those are gaps in the category rather than in one product.

Write a 30-minute interview guide to learn how [customer type] currently handles [problem]. Use questions about specific past events (tell me about the last time you...) rather than opinions or hypotheticals. Include a warm-up, 8 to 10 core questions with follow-up probes, and a closing question. Flag and rewrite any question that is leading, that asks them to predict their own future behaviour, or that mentions our product before they have described the problem.

Here are notes or transcripts from [N] customer interviews: [paste, separated by interviewee]. For each interviewee, summarise their situation, the trigger that made them look for a solution, what they tried, and what they paid or would pay. Then across all interviews, list the patterns that appear in at least three, the notable disagreements, and the direct quotes that best represent each pattern. State clearly how many of the [N] interviews support each finding.

Draft a survey of no more than 12 questions to measure [what we want to learn] among [audience]. Use a screening question first, put behavioural questions before attitude questions, and avoid double-barrelled, leading, and loaded wording. For each question give the answer format and why it is there. Then review your own draft and list every question where the order, the wording, or the answer options could push respondents toward a particular answer.

Here are [N] responses to the open question '[question]': [paste or attach]. Read a first batch of 50 and propose a codebook of 6 to 10 themes with a one-line definition and an example for each. Apply the codebook to every response, allowing up to two codes each. Report counts and percentages per theme, list the responses that fit no theme, and quote the three most representative answers for the largest two themes.

Write a price sensitivity section for a survey of [audience] about [product], using the four Van Westendorp questions (too expensive, expensive but worth considering, a bargain, too cheap to trust) adapted to our product description: [paste]. Then explain how to read the results, what sample size is enough for a rough range, and the two main ways this method misleads for a product people have never used.

Here is an export of our customers with [columns, e.g. industry, company size, plan, signup source, months active, revenue]: [paste or attach]. Propose 3 to 5 segment definitions that differ meaningfully in retention or revenue, with the numbers for each. For each segment say what we would need to believe about their needs for the difference to make sense, and what interview or survey question would test that belief.

Using web search, find the most recent public figures on [topic, e.g. number of independent accounting firms in the US and their average headcount]. For each figure give the number, the source name, the publication date, and a link. Prefer government statistics, regulators, and company filings over blogs and press releases. Mark any figure older than two years, and say where sources disagree rather than picking one.

Turn these research findings into a one-page readout for [audience] deciding [decision]: [paste findings]. Open with the recommendation and the one finding that most supports it. Then give the three strongest pieces of evidence with their sample sizes, the most important finding that cuts against the recommendation, and what we still do not know. End with the decision we need from them and by when.