100 Survey Questions Examples, Organised by the Goal They Serve

Sorted by type, a list of survey questions tells you the grammar: this one is open, that one is a rating scale. Sorted by goal, it tells you what to ask on a Tuesday afternoon when you need an answer. The second sort is the useful one, because a question is only good if it earns its place against a decision you are trying to make.

So the 100 survey questions examples below are grouped by the job each one does: measuring satisfaction, getting feedback on a product you already sell, testing something before you launch it, finding a price, understanding a market, working out why someone left. Pick the goal, take the questions, change the wording to fit your product. Before the list, the short version of what makes a question good and the five ways a question quietly breaks. After it, how to write your own and how to read a few hundred answers without drowning.

This is the broad guide covering every question type. If the questions you need are specifically the ones people answer in their own words, the open-ended questions list goes deeper on that one format, and if you are still deciding whether a survey is the right instrument at all, start with the market research methods guide.

What makes a survey question good

One rule does most of the work. A good question is written to a decision. Before you write anything, finish this sentence: “If I knew the answer to this, I would do X.” If you cannot finish it, the question is decoration, and decoration is what makes surveys long and answers thin.

The second rule is one idea per question. The moment a question asks about two things, the answer becomes impossible to read.

The third rule is neutrality, and it matters more than most people believe. The wording of a question changes the answer you get back. Pew Research Center ran an experiment where half the respondents were asked whether there were plenty of “jobs” available in their community and the other half were asked about plenty of “good jobs”. Sixty per cent said jobs were available, against 48 per cent for good jobs. One adjective moved the result 12 points. In another Pew study, 51 per cent favoured letting doctors “give terminally ill patients the means to end their lives”, but only 44 per cent favoured letting doctors “assist terminally ill patients in committing suicide”. Same policy, different words, seven-point gap. If a single word can do that to a national poll, it can do it to your checkout survey.

Five ways a survey question quietly breaks

Before the examples, the failures to design out. These show up in more than half the surveys I see.

Leading questions point at the answer. “How much did you enjoy our excellent service?” assumes you enjoyed it and that it was excellent. You will get a warm number and learn nothing. Ask “How would you describe your experience with our service?” instead.

Double-barrelled questions ask about two things and accept one answer. “How satisfied are you with the speed and accuracy of our support?” cannot be answered by someone who found it fast but wrong. Split it into two.

Jargon quietly narrows your sample to people who speak your internal language. “How would you rate the UX?” means little to a normal buyer. “How easy was it to use?” means the same thing to everyone.

Too many questions poison the answers you already have. SurveyMonkey’s analysis of its own data found respondents spend about 75 seconds on the first question and only around 19 seconds by questions 26 to 30, and that abandonment rises for surveys taking more than seven to eight minutes, with completion rates dropping by anywhere from 5 to 20 per cent. Every question you add makes the earlier ones worse.

That decline is worth sitting with, because it is a data-quality problem dressed up as a courtesy one. If attention falls by three quarters between the first question and the twenty-sixth, then the order you put questions in decides how good your answers are, not just how polite the survey feels. The practical rule that follows is one most guides never state: put the question you most need a real answer to first, while people are still willing to think, and let the logical flow come second. Demographics, satisfaction ratings and anything you are only asking because it would be nice to know all belong at the bottom, where a rushed answer costs you least.

The wrong type wastes the answer. If you want to count and compare, use a closed question. If you want reasons and language you did not anticipate, use an open-ended one, but use it sparingly: on Pew’s panel, open-ended questions go unanswered around 18 per cent of the time against 1 to 2 per cent for closed questions.

The 100 survey questions, by goal

Take what fits, swap in your product name, and never use all of them at once. Ten questions well aimed beat forty scattered.

Customer satisfaction and experience

1. Overall, how satisfied are you with [product]? (Very satisfied to very dissatisfied)
2. How likely are you to recommend us to a friend or colleague? (0 to 10)
3. How easy was it to get what you needed today? (Very easy to very difficult)
4. How well did [product] meet your expectations? (Far better to far worse than expected)
5. What is the one thing we could do to make your experience better?
6. How would you rate the value for money of [product]?
7. Compared with your last visit, has your opinion of us improved, stayed the same, or got worse?
8. How responsive were we when you had a question or a problem?
9. What nearly stopped you coming back?
10. How would you feel if you could no longer use [product]? (Very disappointed, somewhat disappointed, not disappointed)

That last one is the product-market-fit question popularised by Sean Ellis: the share who would be “very disappointed” is a sharper read on whether you have something people need than any satisfaction score.

Question 2 deserves a caveat the industry rarely gives it. The 0 to 10 recommendation question is the most widely used metric in this list and also the most argued over: its claim to predict growth better than other measures has been challenged repeatedly in the academic literature since it was published. It is a perfectly good tracking number, in that a fall in your score means something changed. Treat it as a thermometer rather than a forecast, and do not restructure a business around a single digit.

Feedback on a product you already sell

11. Which feature do you use most often?
12. Which feature do you never use?
13. What is missing that would make this a must-have for you?
14. What were you using before [product], and why did you switch?
15. How easy is [product] to use? (Very easy to very hard)
16. What almost made you choose a competitor instead?
17. Describe a time [product] saved you money or effort.
18. What is the most confusing part of using [product]?
19. If you could change one thing about [product], what would it be and why?
20. How would you describe [product] to someone who has never heard of it?

Testing a new product or concept before launch

21. In your own words, what does this product do?
22. Who do you think this product is for?
23. How different is this from what you use now? (Very different to not at all different)
24. How likely would you be to buy this? (Definitely would to definitely would not)
25. What would you expect to pay for this?
26. What questions would you want answered before buying?
27. What, if anything, would stop you buying this?
28. What does this remind you of?
29. Looking at these two versions, which do you prefer, and why?
30. What would make this an easy yes for you?

A word of warning on this group. Questions 24 and 30 ask people to predict their own behaviour, and stated intent is a soft signal: people say yes to a survey far more often than they reach for a card. If the decision is expensive, treat these answers as a starting point and confirm the demand for real before you commit to the launch. More on that below.

Pricing and willingness to pay

31. At what price would this be so expensive you would not consider it?
32. At what price would this be so cheap you would question the quality?
33. At what price would this start to feel expensive, but you would still consider it?
34. At what price is this a genuine bargain?
35. Would you buy this at [price]? (Yes or no)
36. And would you buy it at [higher price]?
37. How does this price compare with what you expected to pay?
38. What would justify a higher price for you?
39. Which of these three bundles is the best value to you?
40. If the price went up by 10 per cent, what would you do?

Questions 31 to 34 are the four questions in the Van Westendorp Price Sensitivity Meter, a 1976 method for finding an acceptable price range from what people say. The four questions are useless on their own, so here is how to read them. Plot each answer as a cumulative curve across your price range: too cheap, cheap, expensive and too expensive. Where the “too cheap” and “too expensive” curves cross, you have the point at which the fewest people object on either side, which is usually taken as the optimal price. Where “cheap” crosses “too expensive”, and where “expensive” crosses “too cheap”, you have the two ends of the range most buyers will tolerate. Anything outside those bounds is either leaving money on the table or losing you the sale.

Questions 35 and 36 are the start of a Gabor-Granger sequence, which walks a single buyer up or down the price ladder to find the point where they drop out. Both methods give you a read on price from stated answers, which is useful for framing and dangerous if you treat it as a forecast.

Understanding a market

41. What is the biggest challenge you face with [problem area]?
42. How do you solve that problem today?
43. How much time or money does that problem cost you in a typical month?
44. Where do you look first when you want to learn about [category]?
45. Whose opinion do you trust when choosing a [product type]?
46. What would have to be true for you to switch to a new [product]?
47. How often do you buy in this category?
48. What triggered your most recent purchase in this category?
49. Which brands did you consider last time, and why those?
50. What do you wish existed in this category that does not?

Brand and positioning

51. What three words would you use to describe [brand]?
52. In your view, how is [brand] different from the alternatives?
53. What do you think [brand] stands for?
54. When and where did you first hear about us?
55. What made you remember us afterwards?
56. What would you tell a friend who asked whether they should buy from us?
57. What almost put you off us at first?
58. Which of these statements best describes why you buy from us?
59. What is one thing you wish more people knew about us?
60. If you stopped buying from us, what would you miss least?

Post-purchase, why they bought

61. How did you first hear about us?
62. What were you doing or feeling when you decided to buy?
63. What almost stopped you completing your order?
64. Was there anything you expected to find but could not?
65. How easy was the checkout? (Very easy to very hard)
66. What other products did you look at before choosing this one?
67. What finally convinced you to choose us?
68. Who is this purchase for?
69. How likely are you to buy from us again?
70. What would make you buy sooner next time?

This block belongs at the moment of purchase rather than in an email a week later, which is the whole design principle behind a post-purchase survey. Question 61 is doing more work than it looks: asked properly, “how did you hear about us” is the one attribution read no analytics dashboard can model, because it comes from the customer rather than a pixel.

Churn and cancellation

71. What is the main reason you are leaving?
72. What could we have done to keep you?
73. Did the product stop meeting your needs, or did your needs change?
74. Where are you going instead, if anywhere?
75. How long had you been thinking about leaving before you did?
76. Was price the deciding factor, or something else?
77. What one change would make you reconsider?
78. How likely are you to come back in future?
79. Would you still recommend us to someone whose needs are different from yours?
80. Is there anything else you want us to know?

Website and user experience

81. What did you come to the site to do today?
82. Were you able to do it? (Yes, partly, or no)
83. What, if anything, got in your way?
84. How easy was it to find what you were looking for?
85. Was there anything you expected to see that was missing?
86. What nearly made you leave without buying?
87. How clear was the pricing?
88. Did anything make you hesitate before buying?
89. Overall, how easy was the site to use? (0 to 10)
90. What one change would make this site better for you?

Demographics and screening

91. Which of these best describes your role?
92. Which age range do you fall into?
93. Where are you based?
94. How often do you buy in this category? (Screening)
95. Which of these have you bought in the last 12 months? (Screening)
96. Are you the main decision-maker for [category] in your household?
97. Which of these best describes the size of your business?
98. What is your approximate household income? (Ranges, with a “prefer not to say” option)
99. How did you come across this survey?
100. Is there anything about you we should know to make sense of your answers?

Put sensitive questions like income last, always offer “prefer not to say”, and only ask a demographic question if you will actually cut the results by it. Asking someone’s age and then never using it is just tax on their patience.

How to write your own

The examples are a shortcut, not a substitute. When you write your own, three habits keep them honest.

Start open-ended questions with what, why or how. Those words ask for a reason or a story. “Why did you choose us over the option you were also looking at?” earns you a sentence. “Did you compare us to others?” earns you a yes.

Match the type to the goal. If you need a number you can chart over time, use a rating scale. The familiar 1 to 5 agree-to-disagree scale comes from Rensis Likert’s 1932 monograph, “A Technique for the Measurement of Attitudes”, in Archives of Psychology No. 140, and it endures because it turns an attitude into something you can average and track. The 0 to 10 “would you recommend” question comes from Fred Reichheld’s 2003 Harvard Business Review article, and it endures for the same reason. Reach for a rating when you want to benchmark, and an open-ended question when you want to understand.

Design the scale, not just the question. The “good jobs” experiment shows that a word in the question stem moves the result, and the same is true of the answers you offer. Labelling every point on a scale gets you different numbers from labelling only the two ends. Offering a midpoint gives people somewhere to hide, and taking it away forces a lean they may not feel. Including “don’t know” reduces the number of people who guess, which lowers your response count and raises your data quality at the same time. None of these has one right answer, but each is a decision, and if you change it between waves you have broken your own trend line.

Keep it short, and mean it. Given how fast attention drops after the first handful of questions, treat every question as something you have to earn the right to ask. A tight survey of eight questions that people finish beats a thorough one of thirty that they abandon halfway.

How to read the answers without drowning

Closed questions read themselves: you count them. Open-ended answers are where people give up, because a few hundred free-text replies feel like a wall.

The method that works is coding. Read a sample first to learn the language people actually use, group the answers into a handful of recurring themes, then tag every response against those themes. Now you can say “a third of leavers mentioned delivery speed” instead of “people seemed unhappy”, and you keep the sharpest verbatim quotes for the wording. AI tools now do a decent first pass of that tagging on hundreds of answers in minutes, which removes the main reason people avoid open-ended questions in the first place. If you are building a habit of listening to customers at scale, it is worth comparing the voice-of-customer tools that specialise in exactly this.

When a survey is the wrong tool

Here is the limit worth naming, because it decides whether the answers are safe to act on. A survey measures what people say. For a satisfaction score or a churn reason, what they say is close enough to the truth to use. For a buying decision that has not happened yet, it often is not. The gap between “yes, I would buy this” on a form and an actual purchase is wide and well documented, which is why so many launches that surveyed brilliantly still landed flat.

So if you are about to spend real money, on a new product, a pack, an ad, a claim, or a price, a survey is a good way to sharpen your questions and a poor way to bet the launch on the answers. That is the gap TestFeed is built for: you put the concept, the price, the ad or the pack in front of the shoppers you are actually targeting and get back a purchase-intent read, a clear verdict, and their reasons in their own words, in days rather than weeks. It is directional pre-launch signal, not a guaranteed sales number, and it will not tell you how something tastes or feels. But it moves the expensive decisions off stated intent and onto something closer to real behaviour, which is exactly where a survey runs out of road. If you want to see how a purpose-built test audience differs from a survey panel, that comparison is worth reading next.

Frequently asked questions

What makes a good survey question?

A good survey question is written to a specific decision, asks about one thing only, and uses plain, neutral wording. It avoids leading the respondent, avoids jargon, and matches the question type to what you need: a rating scale when you want to count and compare, an open-ended question when you want reasons in the customer’s own words. If you cannot say which decision a question informs, cut it.

What are the main types of survey questions?

The common types are multiple choice, rating scales such as the 1 to 5 Likert scale, Net Promoter style 0 to 10 questions, yes or no questions, ranking questions, and open-ended free-text questions. Closed types are fast to answer and easy to count. Open-ended types surface reasons and wording you did not anticipate but take more effort to answer and analyse, so use them sparingly.

How many questions should a survey have?

Fewer than you think. SurveyMonkey’s data shows respondents spend around 75 seconds on the first question but only about 19 seconds by questions 26 to 30, and abandonment climbs sharply past roughly seven to eight minutes. Keep most surveys to five to ten focused questions, and if you genuinely need more data, run two short surveys rather than one long one.

What are examples of good survey questions?

Specific, single-idea questions such as “What almost stopped you buying today?”, “How likely are you to recommend us to a friend, from 0 to 10?”, “What were you using before us, and why did you switch?”, and “At what price would this feel too expensive to consider?”. Each is aimed at one decision and asks about one thing.

What is a leading question and why should you avoid it?

A leading question quietly points the respondent towards an answer, for example “How much did you enjoy our excellent service?”, which assumes both enjoyment and excellence. It inflates positive responses and gives you data that flatters rather than informs. Rewrite it neutrally: “How would you describe your experience with our service?”. The same applies to double-barrelled questions, which ask about two things at once and cannot be answered cleanly.

The next step is not to write more questions. It is to open a blank survey, write the one decision you need to make at the top, and delete every question that does not serve it. Whatever survives is your survey.