The prompt
Score 10 of 10, Excellent
Use it when You need short, unbiased survey questions for customer feedback, staff input or market research.
- {{survey_goal}}
- {{decisions}}
- {{respondents}}
- {{channel}}
- {{time_limit}}
- {{draft_questions}}
Full prompt and why it works
Prompt
What the survey is for: {{survey_goal}} Decisions the results will inform: {{decisions}} Who will answer it: {{respondents}} How it will be sent: {{channel}} Time people will give it: {{time_limit}} Draft questions I already have, if any: """ {{draft_questions}} """ Task: Write a survey for the people above that gets honest, usable answers, so that I can make the decisions listed above. Every question should earn its place by feeding one of those decisions. Format: 1. Introduction text, under 70 words: why we are asking, how long it takes, whether answers are anonymous, and how the results will be used. 2. The questions, numbered, each with its type (single choice, multiple choice, rating scale or open text) and its exact answer options. 3. After each question, one line in brackets naming the decision it informs. I will delete these lines before sending. 4. Cut list: questions from my draft that you removed or rewrote, and why. 5. Before you send: a short checklist for testing the survey with three to five people from the group. Rules that make answers honest: - One idea per question. Split anything that asks about two things at once, such as "How satisfied are you with our prices and service?" - Neutral wording, with no leading or loaded questions such as "How much did you enjoy our new menu?" - Ask about past behavior ("In the last month, how often did you...") rather than predictions ("Would you use..."), because people are poor at predicting what they will do. - Use balanced rating scales with every point labeled, the same direction throughout, and a "Not applicable" or "Don't know" option where some people cannot answer. - Answer options must not overlap and must cover every case. - No more than two open text questions, placed near the end. - Ask only for personal details we will actually use, put them at the end, and make them optional. Size: only as many questions as fit in {{time_limit}}, at about three closed questions per minute and a minute or more per open question, and never more than 20. If this is a staff survey, add a note to agree with HR how anonymity will be protected, for example by not reporting results for any group smaller than five people, because staff will not answer honestly if they think they can be identified. If the goal is too broad for one short survey, say so and suggest which decisions to cover now and which to leave for later. If a decision cannot be informed by a survey at all, for example because it depends on what people actually do rather than what they say, say so too.
Variables
- {{survey_goal}}
- What you want to learn, in a sentence or two.
- {{decisions}}
- Each decision the results will feed, one per line, for example "whether to keep weekend opening hours".
- {{respondents}}
- For example "customers who bought in the last six months" or "all staff at two sites".
- {{channel}}
- For example an email link, a QR code in store, or a pop-up in the app.
- {{time_limit}}
- For example "5 minutes".
- {{draft_questions}}
- Questions you already have. Write "none" to start from scratch.
Why it works
- Who it's for. Every question has to name the decision it informs, which is the quickest way to cut the questions that make surveys long and answers thin.
- What to avoid. It rules out double questions, leading wording and predictions about the future, the three habits that most often produce answers you cannot use.
- If unsure. If the goal is too broad for one short survey, or a decision needs observed behavior instead of opinions, it says so rather than writing questions that cannot help.