Survey questions are the prompts you put to people to learn what they think, do or want. They come in two kinds. Closed questions give a set of answers to pick from (“How satisfied are you with support? Very, somewhat, not at all”). Open-ended questions let people answer in their own words (“What almost stopped you from renewing?”). Closed questions count; open questions explain. The main types are multiple choice, rating scales and Likert, NPS 0–10, ranking, matrix, yes/no, demographic, open-ended and follow-up. Below: each type with an example, 50 example questions by situation with the follow-up for each, and the questions to avoid.
An 8 that's really a missing summary report. The customer is happy; the export ritual is the two points.
Pick the closed types when you need to count something you already understand. Pick the open types when you need to find out what you don't. Most good surveys use both: a score, then the reason for it.
People answer in their own words. Use it for the why behind a score, and for problems you didn't think to list. Its weak spot is a vague first answer, which is what the follow-up is for.
One answer from a fixed list. Use it when you know the options and want a count. Add “Other” with a text box when the list might be incomplete. Dropdowns, sliders and image choice are the same type in different clothes.
A statement or question rated on a balanced scale, usually five points with a midpoint. Use it to measure the same thing over time. Keep the scale symmetric: as many negative options as positive.
The likelihood-to-recommend question on an eleven-point scale, grouped into detractors (0–6), passives (7–8) and promoters (9–10). Use it as a trend line, never on its own: the open follow-up is where the reason lives.
People order a short list by importance or preference. Use it when you need priorities, not ratings, and keep the list to five items or fewer. Past the top three, rankings get noisy.
Several items rated on the same scale in a grid. Use it to keep a survey short when the items share a scale. Long grids invite straight-line answering.
One fact, two answers. Use it to route people (“if yes, the next three questions”), not to measure an opinion. “Did you like it?” is a yes/no in disguise and belongs as an open question.
Who is answering: role, company size, how long they've used you. Use it to split the results. Ask these last, make them optional, and only ask what you'll use.
The question you ask after a vague answer. It's built from the person's own words and aims at the missing detail. A form can't ask it; an interviewer, or Spoke, can.
Open questions are the useful ones and the ones most likely to come back as “fine.” One specific follow-up about what the person just said is what turns it into something you can act on.
An 8 that's really a missing summary report. The customer is happy; the export ritual is the two points.
A workaround the manager doesn't know about, and a reason people stop raising things: no reply beyond the emoji.
“Small” means a sharing wall that's hit weekly and answered with screenshots. It's a seat-model problem, not a small one.
The two kinds answer different questions. A closed question tells you how many; an open one tells you why. Most surveys need one of each on every topic that matters.
Fifty open questions, five for each situation, written to copy as they are. Under every question is the follow-up to ask when the first answer comes back vague. Each set links to the full use case with more questions, a sample interview and a sample report. For 66 more open questions see open-ended survey questions; for the five best per situation see feedback questions.
For customers and clients a few times a year, or after a purchase or support conversation.
What did you come to us for in the first place?
What were you using before that?
What's the one thing we do that you'd hate to lose?
What would you have to do yourself if it went away?
Tell us about the last time you contacted support.
What happened after your first message?
Which part of working with us, if any, takes more effort than it should?
How much of your time did it take the last time?
How does what you got compare with what you were promised?
What's the biggest difference, either way?
Ask one of these right after the 0–10 score. The score is the trend; these are the reason.
What would we have to do to earn a 10 from you?
What's the first thing you'd notice if we did?
What happened in the last month that shaped your score?
Which moment stands out most?
What would make you drop your score by a point or two?
How close has that come to happening?
Who is the kind of person you'd recommend us to?
What would you tell them we're best at?
What have you already told someone else about us?
What did they ask you back?
For people using a product or a feature. Ask about the last time, not the usual time.
What did you try to do with [product] this week that didn't work?
What did you do instead?
Which feature, if any, did you find on your own?
What were you looking for when you found it?
What's the last thing you exported out of [product]?
What did you do with it once it was out?
What's the first thing you'd show a colleague who was new to [product]?
Why that one first?
What's one setting or option you've never understood?
What did you assume it did?
For staff, a few times a year. Anonymous answers are a per-survey option on paid plans; say so when it's on.
What part of your week gives you the most energy?
What's an example from the last week?
What's one thing you've stopped bothering to raise?
What happened the last time you did raise something?
When did you last feel proud of something you did here?
What made that one stand out?
What would you fix first if you were in charge for a week?
What's stopped it from being fixed so far?
How does getting things done here compare with your last job?
Which step is the slow one?
For someone leaving the company. For customers who cancel, see the customer exit survey.
When did you first start thinking about leaving?
What happened around that time?
What was the moment you decided for sure?
Who did you tell first?
What will you tell friends about working here?
What would you leave out?
What, if anything, did you never get a straight answer about?
Who did you ask?
What should the next person in your role know on day one?
What did it cost you not knowing it?
At 30, 60 and 90 days, while the outsider's view is still fresh.
Which written guide or doc did you give up on?
What did you do instead?
Which meeting in your first month was most useful?
What did you walk away with?
What's one thing you've been afraid to ask?
What's stopped you asking so far?
How clear is what's expected of you by the end of your first three months?
What's the fuzziest part?
What did you expect to be doing by now that you aren't?
What's holding it up?
For students and trainees, mid-course while there's time to change things, and at the end.
What can you do now that you couldn't before the course?
When did it click?
Which assignment taught you the most?
What did it make you work out for yourself?
Where did you fall behind, if you did?
What was going on that week?
What would you add to the course?
Where would it fit?
Which part will you have forgotten in a month?
What would have made it stick?
Within a day of the event, conference or webinar, while people still remember which session was which.
Which conversation from the event are you still thinking about?
What did you take from it?
What was the low point of the day?
What would have turned it around?
Which session would you have skipped if you'd known?
What did you expect it to cover?
What did you learn that you didn't expect to?
Which session or conversation did it come from?
Who would you bring with you next year?
What would you tell them to expect?
For members of a community, forum or group, including the quiet ones.
What did you hope to get from this community that you haven't yet?
What have you tried in order to get it?
What was the last thing that made you open the community?
What did you do once you were in?
Whose posts do you always read?
What do they do that others don't?
What stops you from posting?
When did you last almost post?
What would make you recommend the community to a friend?
Who's the friend you're thinking of?
For feedback about one person from the people around them. Also see manager feedback and peer feedback.
What's one thing this person does that you'd want them to keep doing?
When did they last do it?
What's one thing they should do more of?
What would change if they did?
Think of the last time you disagreed with them. How did it go?
What did they say that you still remember?
What would you want them to know that they might not hear otherwise?
What makes it hard to say directly?
What's one thing they've gotten better at this year?
What do you think changed?
A biased question gets you your own opinion back. Five common kinds; the full list of ten, with a fix-it table, is in biased and leading survey questions.
The wording hints at the answer you want. Cut the adjective and ask what happened.
Two questions in one sentence get half an answer to each. Split them.
The question assumes something about the person that may not be true. Check the premise first.
“Always” and “never” force a no. Ask for a count over a set time instead.
Your internal words aren't theirs. Describe the thing, not the feature name.
Seven steps, in order.
The best survey question in the world still gets “fine” and “no complaints” some of the time. What separates a useful survey from a pile of scores is the question that comes after: one specific follow-up built from what the person just said. The probing questions guide has fifty of them by type and situation, and the open-ended guide pairs every question with its own.
A form can't ask a follow-up; it shows everyone the same fixed questions. An agent survey runs the survey as a short conversation: Spoke reads each answer and asks the follow-up when it's vague, so two people can get different follow-ups depending on what they said. You set the goal, the topics, how deep to probe and what to stay away from. When you'd rather write the fixed questions yourself, classic surveys are free and unlimited on every plan.
Two ways to run the questions above: classic surveys for fixed questions you write yourself, and agent surveys, where Spoke interviews each person and asks the follow-ups. For research at depth, see AI-moderated interviews. Comparing tools? Read how ScatterSpoke stacks up against SurveyMonkey, Google Forms, Typeform and Qualtrics, and see pricing.
Closed types: multiple choice, rating scale and Likert, NPS 0–10, ranking, matrix, yes/no and demographic. Open types: open-ended questions and follow-up (probing) questions. Closed questions count; open questions explain.
Ones that are open, neutral, about one thing, and that someone will act on. “What almost stopped you from renewing?” is good; “Don't you agree our checkout is easy?” is not, because it tells people the answer you want.
Closed: “How likely are you to recommend us, 0 to 10?” Open: “What would we have to do to earn a 10?” Follow-up: “What's the first thing you'd notice if we did?” This page has 50 more, grouped by customers, employees, product, events, courses and community.
As few as the decision needs. Write down what you'll do with each answer; if no answer would change anything, cut the question. Tell people how long it takes and be honest about it.
A statement rated on a balanced scale of agreement, usually five points from strongly disagree to strongly agree with a neutral midpoint. Use it to measure the same thing over time, and keep as many negative points as positive.
A closed question gives a fixed set of answers to pick from, so you can count them. An open-ended question lets people answer in their own words, so you learn the reason. Most surveys need one of each on every topic that matters.
Ask one thing, in plain words, without hinting at the answer. Prefer “the last time” over “usually.” Use a balanced scale with a midpoint. Then plan the follow-up you'll ask when the answer comes back vague.
Leading questions (“How much easier is the new version?”), double-barreled questions (two things in one), loaded questions that assume something about the person, absolutes like “always,” jargon, and unbalanced scales. Each has a neutral rewrite on the biased and leading questions page.
Spoke asks the follow-ups for you, with every person. Free to start, no card. Your first ten agent surveys are on us.
