100+ Survey Question Examples for Every Goal (and How to Write Your Own)

This page is a bank of 113 survey question examples you can copy right now. The right questions to ask in a survey depend on what you are trying to decide, so the bank is organized by goal, not by question type: pick your goal below, jump to that section, and adapt the wording to your context. After the bank, you’ll find a short guide to question types, seven rules for writing your own questions (each with a weak version and a rewrite), and the survey-level mistakes that quietly ruin data.

• Customer satisfaction survey questions
• NPS follow-up questions
• Product feedback questions
• Employee engagement and eNPS questions
• Event feedback questions
• Market research questions
• Demographic questions
• Education and course evaluation questions
• Lead qualification questions

Good survey questions share one property: each one measures a single thing, in language your respondent would use. When you adapt an example, keep that property. It’s the difference between data you can act on and answers you have to guess at.

Customer satisfaction survey questions

Use these after a purchase, a support interaction, or at a regular interval with active customers. Mix one scale question with one open-ended question per topic; the scale gives you the trend, the open answer tells you why.

1. How satisfied are you with your most recent experience with us? (1 to 5, very dissatisfied to very satisfied)
2. Have you contacted our support in the last 90 days? (yes / no; on no, skip questions 3 to 5, which cover your most recent case)
3. How long did you wait before getting a first response from a person, not an automated acknowledgement? (multiple choice: under 1 hour / 1 to under 8 hours / 8 to under 24 hours / 24 hours or more / no response yet; on no response yet, skip questions 4 and 5)
4. How well did the support team understand your problem? (1 to 5)
5. Was your issue resolved on the first contact? (yes / no / partially)
6. How easy was it to find what you were looking for? (1 to 5, very difficult to very easy)
7. How would you rate the value you get for what you pay? (1 to 5)
8. What almost stopped you from buying from us?
9. Which part of the experience most needs improvement? (multiple choice with your own touchpoints as options, plus “no improvement needed”)
10. What did we do well that we should keep doing?
11. If you could change one thing about our service, what would it be?
12. How does your experience with us compare to what you expected? (worse / as expected / better / had no expectations)
13. Is there anything else you’d like to tell us?

A note on question 13: generic catch-all questions produce few answers, but the ones they produce are often the most useful in the whole survey, because respondents use them to report the thing you didn’t think to ask about. Keep it optional and last.

NPS follow-up questions

The Net Promoter Score question itself is fixed: “How likely are you to recommend us to a friend or colleague?” on a 0 to 10 scale. The score alone doesn’t tell you what to fix, so the value of an NPS survey is concentrated in the follow-up. Show these based on the score the person just gave.

14. What is the main reason for your score? (everyone)
15. What would we need to change for you to score us higher? (detractors, 0 to 6)
16. What was missing from your experience? (detractors)
17. Which problem should we solve first? (detractors, multiple choice with known issues as options, plus “other” with a write-in)
18. What kept your score from being a 9 or 10? (passives, 7 to 8)
19. What do you value most about us? (promoters, 9 to 10)
20. What almost made you choose a different provider? (promoters)
21. In your own words, how would you describe us to a colleague? (promoters)
22. May we quote your answer publicly? (promoters, yes / no)
23. How likely are you to renew or keep using us over the next 12 months? (0 to 10, everyone)
24. May we contact you about your answers? (everyone; yes / no, with an email field shown only on yes)

Branching by score is what separates an NPS survey from an NPS form letter. Asking a detractor “what do you love about us” reads as tone-deaf, and asking a promoter to list problems wastes your best goodwill. Conditional logic handles the routing automatically.

Product feedback questions

These work in-product, after onboarding, or as a recurring pulse to active users. They diagnose the product itself rather than a single interaction.

25. How easy is the product to use day to day? (1 to 5)
26. How well does the product cover what you need it for? (1 to 5)
27. Which feature do you use most often? (multiple choice)
28. Which feature did you expect to be better? (multiple choice, plus “all met expectations” and “had no prior expectations”; on either, skip question 29)
29. What did you expect that feature to do better? (open, shown only when question 28 names a feature)
30. What do you use alongside our product to get the full job done?
31. What were you using before us? (if nothing, answer “nothing” and skip question 32)
32. What made you switch?
33. How disappointed would you be if you could no longer use the product? (very / somewhat / not disappointed)
34. What is the main benefit you get from the product, in your own words?
35. How would you rate the product’s speed? (1 to 5)
36. In the last 30 days, how many times did a failure (an error, an outage, anything that blocked you) stop you from using the product? Count each incident once. (not once / 1 or 2 times / 3 to 5 times / more than 5 times)
37. How clear is our pricing for what you get? (1 to 5)
38. What was the hardest part of getting started with the product?
39. If you could wave a wand and change anything about the product, what would it be?
40. What kind of company do you think would get the most out of the product?

Question 33 is the Sean Ellis product-market-fit question. Track the “very disappointed” percentage over time and by segment rather than judging a single round; the trend carries more signal than the absolute number. Questions 30 to 32 map your real competitive landscape, which is often not the one in your pitch deck.

Employee engagement and eNPS questions

Run these quarterly or twice a year, anonymously. The eNPS variant of the recommendation question opens the set, and agreement statements (a Likert scale) carry the diagnosis.

41. How likely are you to recommend this company as a place to work? (0 to 10, this is the eNPS question)
42. What is the main reason for your score?
43. I feel my work is recognized. (1 to 5, strongly disagree to strongly agree)
44. I have the tools I need to do my job well. (1 to 5)
45. I get the information I need to do my job well. (1 to 5)
46. I receive feedback that helps me improve. (1 to 5)
47. I see a path to grow at this company. (1 to 5)
48. My manager genuinely cares about my well-being. (1 to 5, plus “not applicable” for respondents without a manager)
49. I can voice a dissenting opinion without fear. (1 to 5)
50. My workload is sustainable. (1 to 5)
51. I understand how my work contributes to the company’s goals. (1 to 5)
52. The company’s leadership communicates openly about where we’re going. (1 to 5)
53. What is one thing we could change to make your work better?
54. What should we be careful not to change?

Anonymity has to come from the survey’s design. In a 12-person company, asking for team and tenure together identifies people. Collect only the segments you’ll actually analyze, and state at the start of the form what is and isn’t recorded.

Event feedback questions

Send within 48 hours of the event, while the memory is fresh. Two of these belong before the event, in the registration form, where they shape the program instead of grading it.

55. What do you most want to get out of this event? (registration form, multiple choice plus “other”)
56. Which topics would you like to see covered? (registration form, open or multiple choice)
57. How would you rate the event overall? (1 to 5)
58. How well did the content match what the event promised? (1 to 5, not at all to completely)
59. Which session was most useful to you? (multiple choice, one per session; for multi-track events, show only the sessions the respondent attended)
60. Which session could be cut next time? (multiple choice, with a “none” option; for multi-track events, show only the sessions the respondent attended)
61. Was the pace of the sessions right for you? (too slow / about right / too fast)
62. Was the overall length of the event right? (too short / about right / too long)
63. How would you rate the venue or, for online events, the platform? (1 to 5)
64. How clear was the communication before the event? (1 to 5)
65. How worthwhile was the event for the time you invested? (1 to 5, not at all worthwhile to completely worthwhile)
66. How likely are you to attend the next edition? (0 to 10)
67. What is the one thing we should do differently next time?

Watch the gap between question 57 and question 65. Attendees often rate an event highly and still say it wasn’t worth their time. Treat that gap as a signal to investigate with a follow-up question, not as a diagnosis by itself.

Market research questions

For validating a product idea, sizing demand, or understanding how a market talks about a problem before you build for it.

68. When was the last time you dealt with [the problem]? (under 7 days ago / 7 to under 30 days ago / 30 days to under a year ago / a year ago or more / never / can’t remember; on never, skip questions 69 to 74; on can’t remember, skip question 74)
69. How do you handle [the problem] today?
70. What have you already tried for this?
71. If you stopped using any of those, why?
72. About how many hours did dealing with [the problem] take the last time it happened, start to finish?
73. Setting time aside, about how much money did that last occurrence cost you?
74. Walk me through the last time it happened: what triggered it, and what did you do?
75. If a tool solved [the problem], what would it need to do for you to adopt it?
76. What would make you distrust a new solution in this space?
77. At what price would [the product] be so expensive you wouldn’t consider it? (Van Westendorp pricing question)
78. At what price would it be expensive, but still worth considering? (Van Westendorp)
79. At what price would it be a bargain? (Van Westendorp)
80. At what price would it be so cheap you’d doubt the quality? (Van Westendorp)
81. Who decides on purchases like this? (multiple choice: just me / me together with others / someone else)
82. Where have you discovered new tools in this space in the last 6 months? (select all that apply: search / recommendations / social / review sites / other; plus “none in this period” as an exclusive choice)

Questions 68 to 74 deliberately ask about past behavior instead of future intent. “Would you use this?” invites politeness; “what did you do last time?” produces facts. The four Van Westendorp questions (77 to 80) together map an acceptable price range, and they only work as a set.

Demographic questions

Demographic questions segment your results; they rarely stand on their own. Put them at the end of the survey, make them optional, and only ask for segments you’ll actually use in the analysis. Every question you drop here lifts completion.

83. What is your age group? (under 18 / 18 to 24 / 25 to 34 / 35 to 44 / 45 to 54 / 55 to 64 / 65 or older / prefer not to say)
84. What is the highest level of education you’ve completed? (options adapted to your country’s system, plus “prefer not to say”)
85. What is your main employment status, the one that takes most of your week? (employed full-time / employed part-time / self-employed / student / retired / not currently working / prefer not to say; on student, retired, or not currently working, skip questions 86 to 88)
86. What is your role or job title?
87. What industry do you work in? (multiple choice with an “other” option)
88. How many people work at your company? (just me / 2 to 10 / 11 to 50 / 51 to 200 / 201 to 1,000 / more than 1,000)
89. Where are you located? (country or region list)
90. What is your household income range? (ranges adapted to your market, always with “prefer not to say”)
91. How long have you been a customer? (less than 3 months / 3 months to under 1 year / 1 year to under 3 years / 3 years or more)

Always offer “prefer not to say” on personal items, and make brackets mutually exclusive: ranges like “25 to 34” and “34 to 44” overlap at 34 and force some respondents to pick arbitrarily. We’re preparing a dedicated guide to demographic questions with complete answer-option sets for each item.

Education and course evaluation questions

For course, class, or training evaluations, whether you’re an instructor, a course creator, or an L&D team.

92. How would you rate the course overall? (1 to 5)
93. The course delivered what its description promised. (1 to 5, strongly disagree to strongly agree)
94. The instructor explained concepts clearly. (1 to 5)
95. The pace of the course was right for me. (too slow / about right / too fast)
96. The course materials supported my learning. (1 to 5)
97. The practical exercises reflected situations I actually face. (1 to 5, plus “not applicable” when the course had no exercises)
98. Which module was most valuable to you? (multiple choice)
99. Which module needs the most work? (multiple choice, with a “none” option)
100. What can you do now that you couldn’t do before the course?
101. How likely are you to recommend this course to a colleague? (0 to 10)
102. What almost made you not enroll? (voluntary enrollment only; skip for mandatory training)

Question 100 is the one instructors skip and shouldn’t: it captures the outcome as students perceive it, not just the experience. Strong answers become testimonials (ask permission separately), and vague answers on an otherwise well-rated course are a prompt to check retention with a direct assessment, not proof that students retained little.

Lead qualification questions

A qualification form is a survey about fit that respondents complete one at a time. These questions decide who gets a sales call, who goes into an email sequence, and who receives a polite no. Ask the fewest you can act on: each extra required field costs conversions.

103. What are you trying to solve right now? (multiple choice with your main use cases, plus “other”)
104. What is your main area of responsibility? (multiple choice: running the whole company / marketing / sales / operations / other)
105. How many people work at your company? (just me / 2 to 10 / 11 to 50 / 51 to 200 / 201 to 1,000 / more than 1,000)
106. What tool or process do you use for this today?
107. What’s prompting the change now?
108. When do you want a solution in place? (this month / this quarter but not this month / this year but not this quarter / next year or later / no date yet)
109. What budget range have you set aside for this? (ranges, plus “not defined yet”)
110. Who will be involved in this decision? (select all that apply: me / my team / leadership / procurement / other)
111. What would make this project a success six months after it’s in place?
112. How did you hear about us? (multiple choice: organic search / a paid ad / a recommendation from someone I know / our social media content / other)
113. Anything we should know before we talk?

Not every lead should see every question. Conditional logic can send high-fit answers straight to a calendar page and route “no date yet” leads to a nurture sequence instead of a sales call, so the form qualifies without interrogating.

Survey question types 101

Every question in the bank above is built from a small set of types. Knowing which type produces which kind of data is what lets you write your own instead of copying.

Survey goal	Question types that carry the weight	The data you get
Customer satisfaction	Rating scales plus one open “why”	Scores you can trend, quotes that explain them
NPS and loyalty	The 0 to 10 NPS question, branched open follow-ups	A benchmarkable score and sorted reasons
Product feedback	Multiple choice, PMF-style categories, open text	Feature priorities and verbatim needs
Employee engagement	Likert agreement statements, eNPS	Dimension scores comparable across rounds
Event feedback	Ratings per session, one open improvement question	A ranked list of what to keep and cut
Market research	Behavioral open questions, Van Westendorp pricing	Facts about behavior, an acceptable price range
Demographics	Closed brackets with “prefer not to say”	Clean segments for cross-analysis
Course evaluation	Likert statements, outcome open question	Teaching diagnosis plus self-reported learning outcomes
Lead qualification	Closed fit questions, conditional routing	A score you can route on

The types themselves, briefly:

• Open-ended questions collect free text. Maximum signal per answer, highest effort per answer. Use them where you need the “why”, never as the default.
• Closed questions (multiple choice, yes/no, dropdowns) collect structured data you can count, filter, and chart. Use them wherever you already know the plausible answers.
• Likert items measure agreement with a statement on a symmetric scale, usually 5 points. Our Likert scale guide covers writing and analyzing them, including the mean score calculation.
• Rating scales (1 to 5 stars, 0 to 10) measure intensity of a single perception. The NPS question is a rating scale with a fixed benchmarkable format.
• Ranking questions force respondents to order options by preference. Costly to answer; use only when trade-offs are the thing you’re measuring.

How to write your own: 7 rules with rewrites

The bank covers the common goals. When you need a question that isn’t there, these rules produce one that measures what you think it measures. Each comes with a weak version and its rewrite.

1. Ask about one thing at a time.

Weak: “Was the support fast and helpful?”

Better: “How quickly did support respond?” and “Did the answer solve your problem?”

A yes to the weak version could mean fast, helpful, or both. Two answers you can interpret beat one you can’t. Questions that bundle two concepts are called double-barreled, and they’re one of the most common flaws in real surveys.

2. Keep the wording neutral.

Weak: “How much did you enjoy our redesigned, faster checkout?”

Better: “How was your checkout experience?”

Wording moves numbers more than most people expect. In a Pew Research experiment, support for military action in Iraq was 68% when the question mentioned only the action, and 43% when it mentioned “thousands of casualties”; their question-writing guide documents the effect. Your product adjectives do the same thing at a smaller scale.

3. Anchor questions in time and behavior.

Weak: “Do you usually export your data?”

Better: “Have you exported your data in the last 30 days?”

“Usually” means something different to every respondent. A time window converts a vague self-image question into a factual one, and facts are what you can act on.

4. Match the format to what you’ll do with the answer.

Weak: an open text box asking “What could we improve?” sent to 5,000 customers with no plan to read the answers.

Better: a multiple-choice list of your five candidate improvements, plus an optional “other”.

Format changes the results themselves. In another Pew experiment, 58% of respondents picked “the economy” as the deciding election issue when it was offered as a closed option; only 35% volunteered it when the same question was open-ended. Closed questions give you counts you can rely on; open questions surface answers you didn’t predict. Decide which job the answer has before choosing.

5. Make answer options complete and mutually exclusive.

Weak: “How many employees? 1 to 10 / 10 to 50 / 50 to 200”

Better: “1 to 10 / 11 to 50 / 51 to 200 / more than 200”

A respondent with exactly 10 employees fits two boxes in the weak version, and one with 500 fits none. Every closed question needs a home for every honest answer, which is also why “other” and “prefer not to say” exist.

6. Use the respondent’s words, not yours.

Weak: “How satisfied are you with our omnichannel engagement capabilities?”

Better: “How easy is it to reach us on the channel you prefer?”

Internal vocabulary is invisible to you and alien to them. Read your question aloud; if a customer wouldn’t say those words in a conversation, rewrite it.

7. Every question must be able to change a decision.

Weak: “What industry are you in?”, asked at the end of a CSAT survey whose results are never segmented by industry.

Better: delete it, and keep only “How satisfied are you with your most recent experience?” plus one open “why”.

The test for any question is what you’ll do differently based on the answer. No answer, no action, no question. This rule deletes more questions than it writes.

Common mistakes that ruin survey data

The rules above operate at the question level. These mistakes live at the survey level, and they can sink a survey made entirely of good questions.

Making it too long. SurveyMonkey’s analysis of about 100,000 surveys found that respondents spend around 5 minutes on a 10-question survey, and that abandonment rises once surveys take more than 7 to 8 minutes, with completion rates dropping by 5 to 20%. The same analysis found respondents speed up as surveys get longer, which suggests late answers in a long survey are also lower-quality answers.

Ignoring question order. Earlier questions change later answers. In a Pew experiment, support for legal agreements between same-sex couples came out at 45% when the question followed one about same-sex marriage, and 37% without it. Put broad questions before specific ones, and never put a question you must have after one that primes it.

Flipping the scale mid-survey. If 5 means “very satisfied” in the first block, 5 must mean the positive end until the last. Respondents settle into a pattern; a reversed item reads like noise in the data, and you can’t tell careless answers from real ones.

Promising anonymity the design doesn’t keep. An “anonymous” employee survey that asks for team, tenure, and role can identify most people in a small company. Respondents sense this and answer accordingly. Collect the minimum segmentation you’ll analyze, and say plainly what is recorded.

Surveying only the happy path. Sending the survey to active customers only, or triggering it right after a success moment, inflates every score. Include the silent and the churned if you want scores you can trust.

Skipping the pre-test. Five colleagues answering the survey before launch will find the ambiguous wording, the missing option, and the question that takes three reads. Every problem found in a pre-test is a problem your data won’t have.

From question list to live survey

Copying questions into a form builder one by one is the tedious part, and it’s now optional. In Yay! Forms, you can describe the survey in one sentence, like “customer satisfaction survey with a CSAT question, a first-contact resolution question, and one open comment field”, and the AI generates the form with the questions, answer options, and conditional logic in place. You edit from there instead of starting from a blank page. When responses arrive, the AI analysis reads the open-ended answers for you and surfaces the recurring themes, which is where surveys with good open questions usually stall. The 7-day free trial doesn’t ask for a credit card.

Frequently asked questions

What are good questions to ask in a survey?

Good questions to ask in a survey are the ones whose answers can change a decision you’re about to make. Start from the decision, pick the goal section above that matches it, and use closed questions for what you need to count and one or two open questions for what you need to understand. If no decision depends on the answer, drop the question.

How many questions should a survey have?

As few as the decision requires, and short enough to stay under about 7 to 8 minutes of answering time. SurveyMonkey’s completion-time research across roughly 100,000 surveys found a 10-question survey takes about 5 minutes and that abandonment climbs past the 7 to 8 minute mark. Transactional surveys like CSAT and NPS work at 1 to 5 questions, while research and engagement surveys can sustain 20 or more when respondents care about the outcome.

What is the difference between open-ended and closed-ended questions?

Open-ended questions collect free text in the respondent’s own words; closed-ended questions offer a fixed set of answers. Closed answers can be counted and compared at any scale, while open answers carry the reasons and the vocabulary that closed options miss. Most surveys work best leading with closed questions and reserving one or two open ones for the “why”.

What makes a survey question biased?

A question is biased when something in its wording, options, or position nudges respondents toward an answer. The frequent offenders are loaded wording (“how much do you love…”), answer options that don’t cover all honest answers, double-barreled constructions, and a preceding question that primes the topic. The rewrites in the rules section above show how to fix each pattern.

---

Fabio Borges is the founder and co-CEO of Yay! Forms. He started the company in 2022 because every form tool he had used as a customer forced a choice between price, design, and insight. He writes about surveys, forms, and the data they produce.