Survey Questions to Ask Your Customers
For a business owner, marketer or product person writing a customer survey. The questions follow the order of a good form: overall satisfaction and likelihood to recommend first, then the product or service, buying and support, price, why customers chose you or left, and profile and closing questions last. Each has a note on wording, the answer scale to use and what a good or a worrying result looks like, and since nobody should send them all at once, the notes also say which to cut.
Want questions from the whole vault instead? Try the random question generator.
The questions
Each question, and why to ask it
Overall
Taking everything into account, how satisfied are you with us?
Why ask it
Put it first, before a question about any one detail colors the answer. Use five labeled points from very dissatisfied to very satisfied and report the share who chose the top two. When most people land on 'satisfied' and few on 'very satisfied', read it as a polite shrug and look to the open questions for what is missing.
On a scale of 0 to 10, how likely are you to recommend us to a friend or colleague?
Why ask it
This is the Net Promoter question, and it only compares with anyone else's score if you keep the 0 to 10 scale and this wording. Count 9s and 10s as promoters and 0 to 6 as detractors, each as a percentage of everyone who answered, then subtract the second from the first. A single round tells you little, so watch how the number moves between surveys.
What made you choose that number?
Why ask it
An open text box placed straight after the recommend question, and the one to keep if the survey shrinks to two questions. Sort the replies into a handful of themes and count them separately for low and high scorers. When the low scorers keep naming the same thing, you have your next piece of work.
If you needed this again tomorrow, would you come back to us?
Why ask it
Recommending and returning are different things: people recommend a wedding caterer they will never hire twice. Offer definitely, probably, not sure, probably not and definitely not, and count only 'definitely' as loyal, since 'probably' is often what people tick to be kind. For a subscription, ask how likely they are to renew.
If we could improve only one thing for you, what should it be?
Why ask it
Asking for one thing forces a choice, so you get the complaint that matters most instead of a list. Keep it optional and give it a small text box, which invites a sentence and not an essay. The same answer turning up again and again is a finding, while a scatter of unrelated wishes usually means nothing is badly broken.
How satisfied were you with your most recent purchase?
Why ask it
This one is about a single transaction, so send it within a day or two while the details are fresh, and keep its results apart from the overall score. One late delivery can sink it for a customer who still thinks well of you. In a once-a-year survey of all customers, cut it.
Compared with what you expected, was your experience with us better, worse or about the same?
Why ask it
Give three or five options with 'about what I expected' in the middle. 'Worse than expected' next to a decent satisfaction score tends to point at an ad or a sales pitch that promised too much, more than at the work itself. Add a text box that appears only for 'worse' and asks what they had expected.
How would you describe us to a friend, in your own words?
Why ask it
The replies are the words your customers use, which are rarely the words on your website. A good line can go into your marketing, but only after you ask its writer for permission. Open boxes get skipped more than scales, so this is an early one to cut when the survey runs long.
Over the past year, has your opinion of us improved, stayed the same or gotten worse?
Why ask it
Show it only to people who have been customers long enough to compare. It catches drift that a single score hides, because someone can still be satisfied and feel things are slipping. A growing share of 'gotten worse' among your longest-standing customers is an early warning, and worth a phone call to a few of them before the next round.
Product
How well does our product do the job you bought it for?
Why ask it
Swap in the name of the product or service, since 'our product' makes people guess which one you mean. Five points from not at all well to extremely well. Middling marks here beside high marks for your staff mean customers like you and are putting up with the thing you sell.
What do you like most about our product?
Why ask it
An open box, and the counterweight to every question that hunts for faults, because it tells you what not to break. When the replies praise something you treat as a side detail, such as the packaging or one member of staff, that detail is doing more of the selling than you thought. Place it ahead of the questions about problems so the answer is not colored by them.
What were you hoping to get done when you bought from us?
Why ask it
List the uses you expect and add an 'other' box, then read the other box closely. Customers often buy for a job you never advertised, and each of those is a message or a market you have been missing. Run it once and it will not need repeating every round.
How would you rate the quality of what you received?
Why ask it
Quality means one thing for a meal, another for a haircut and another for an app, so where you can, break it into two or three named parts in a small grid, such as durability, finish and fit. A single quality mark is easy to chart and hard to act on. Poor to excellent works as the scale.
How easy or difficult was it to get started with our product?
Why ask it
Naming both ends ('easy or difficult') keeps the wording from leaning toward a yes. Send it in the first couple of weeks, since the start is when people quietly give up and a survey six months on reaches only those who got through. The useful follow-up is the step where they got stuck, offered as a short list of your own setup steps so the answers can be counted.
Which parts of our product or service do you use most?
Why ask it
Make it a pick-up-to-three list, shuffled if your survey tool allows, so the first option does not win by position. Whatever nobody ticks is a candidate to explain better or to stop maintaining. Check before you cut, because a rarely used feature can still be the reason somebody stays.
Is there anything you expected our product to do that it does not?
Why ask it
Yes or no, with a text box on yes. It collects missed expectations, which are a better guide than a wish list: a wish is something they would like, and a missed expectation is something they thought they were paying for. Pass the replies to whoever writes your product descriptions as well as whoever builds the product.
How often do you use our product?
Why ask it
Give ranges tied to time (every day, a few times a week, a few times a month, less often) and avoid words like 'regularly', which each person reads differently. Its main job is to split the other answers. A complaint from a daily user weighs more than one from someone who has opened it twice.
Have you run into a problem with our product in the last three months?
Why ask it
The time frame stops people reaching back to a fault you fixed long ago. Follow a yes by asking what happened and whether they told you about it. The customers who hit a problem and said nothing are the ones no support log will show you.
Here are some things we are thinking about adding. Which one would you want first?
Why ask it
A choice among real candidates beats an open 'what should we build', which mostly returns things you cannot do. List only options you would act on, include 'none of these', and present the result as a vote on order. Avoid wording that sounds like a promise, or the losing options become complaints later.
How disappointed would you be if our product were no longer available?
Why ask it
Offer very, somewhat and not disappointed. It measures need where satisfaction measures liking, and people are often perfectly satisfied with things they would not miss. Look at who answered 'very' and what they use the product for, because that is the group to build for.
Buying and support
How easy or difficult was it to find what you were looking for on our website or in our store?
Why ask it
Ask about one place per survey: delete 'website' or 'store' so the answer points somewhere. If sales are healthy and the marks are still low, people are buying despite the layout. Ask what they were hunting for and you have a list of things to move, rename or put on the front page.
How easy or difficult was it to complete your purchase?
Why ask it
Best asked right after checkout, on a scale from very difficult to very easy. Follow anything below 'easy' with one box: what slowed you down? Typical replies are a forced account, a shipping cost that showed up late or a payment method you do not take, and each one can be fixed.
Was there anything you wanted to know before buying that you could not find?
Why ask it
The answers are your missing FAQ: sizing, delivery times, what is included, how to cancel. Put each one on the page where the doubt comes up, not in a help section a first-time visitor has no reason to open. Yes or no, then a box.
Did your order arrive, or the work get done, when we said it would?
Why ask it
Yes, no or not sure: it is a fact, so it needs no scale. Trim the wording to fit what you sell. You can check the replies against your own delivery or job records, and where the two disagree, the customer was probably given a date nobody wrote down.
How well did we keep you informed between ordering and delivery?
Why ask it
Silence after payment is its own complaint, separate from lateness. An order that arrived on time can still score low here if the customer had to chase a tracking number or a start date. The fix is often a single automatic message, which makes this a cheap question to act on.
Have you contacted our support team in the past six months?
Why ask it
A filter, not a finding. Use it to skip the support questions for everyone who answers no, because people asked to rate a service they never used either quit the survey or invent an answer. If the survey is triggered by a closed support ticket, you already know the answer and can drop it.
How satisfied were you with the help you received from our support team?
Why ask it
The headline support score, on the same five satisfaction points as the first question on this page. Send it when the case closes and tie it to the case, not to the name of the person who handled it. Report it apart from overall satisfaction, since a customer can be delighted with the help and still be leaving over the fault that made them get in touch.
How much effort did it take to get your issue sorted out?
Why ask it
The customer effort question, on five points from very little to a great deal. High effort on a solved problem still wears out patience, so do not let a good resolution rate hide it. Ask how many times they had to get in touch, and look first at the cases that took more than one contact.
Is the issue you contacted us about now fully resolved?
Why ask it
Offer yes, partly and no, since a plain yes or no pushes the half-fixed cases into yes. Route every 'no' to someone who can reopen the case, which means this question cannot be anonymous. A survey that hears about an open problem and does nothing makes the problem worse.
Did the person who helped you understand what you needed?
Why ask it
Asking about understanding gets further than asking whether staff were friendly, which nearly everyone ticks yes to. If you also want knowledge and courtesy, give each its own row, because one question about two things produces an answer you cannot read. Use the results for coaching and never rank individuals on a handful of replies.
How did you feel about the time it took us to respond?
Why ask it
Your own records hold the real wait, so ask how it felt: faster than expected, about right or too slow. 'Too slow' on a wait you thought was fine tells you what customers expect, and that is the number to staff for. Where waits vary by channel, ask which one they used.
Price
For the price you paid, how would you describe the value you got?
Why ask it
Ask about value and not price, since asking whether something is expensive invites a yes. Poor value from people who rated the product well is a pricing or packaging problem. Poor value alongside poor product marks means the price is not the thing to fix.
Compared with similar options, are our prices lower, about the same or higher?
Why ask it
'Higher' is not bad news when satisfaction is also high: those customers know they pay more and stay anyway. Add 'I have not compared' and note how many choose it. A large share there suggests price is less of a battleground than it feels from the inside.
How clear was our pricing before you bought?
Why ask it
Run it from very clear to not at all clear. Confusion here tends to resurface as billing disputes and refund requests, and it is cheaper to fix on the price page. If you quote per job, replace it with whether the final bill matched the quote.
Were there any costs you did not expect?
Why ask it
A delivery charge, a setup fee or a renewal at a higher rate is remembered long after the amount is forgotten. Each reply names a line that should appear earlier in the quote or at checkout. If one fee comes up repeatedly, weigh building it into the headline price.
Does the plan or package you have fit what you need: too little, about right, or more than you use?
Why ask it
'More than I use' is a cancellation waiting for the next renewal notice, and offering a smaller plan first can keep that customer. 'Too little' is a sales lead, so decide who follows up before you send the survey. Leave the question out if you sell one thing at one price.
If our price went up by 10 percent, what would you do?
Why ask it
What people say they would do about a price rise and what they then do are different, so treat the answers as a ranking of who is most at risk and not as a forecast. Offer four choices: stay, stay but look around, buy less, leave. Skip it unless a change is truly on the table, because the question itself tells customers one is coming.
Choosing and leaving
How did you first hear about us?
Why ask it
Multiple choice with the channels you really use, plus 'a friend or colleague told me' and 'other'. Memory is shaky here, so read it as rough shares and set it beside your analytics. A large word-of-mouth share that no ad report shows is worth knowing before you set next year's budget.
What was the main reason you chose us?
Why ask it
Allow one answer only from a list such as price, location, a recommendation, a particular feature or reputation, with an 'other' box. 'Select all that apply' gives you a tie, and a single choice gives a ranking. Lead your advertising with the winner, which is often not the one you would have bet on.
Which other companies did you look at before choosing us?
Why ask it
List the rivals you know, then add 'none, I only looked at you' and an 'other' box. The names written in are your real competition, and it can include doing the job themselves or not doing it at all. Many 'none' answers mean customers are arriving already decided, often on a recommendation or out of habit.
What nearly stopped you from buying?
Why ask it
Everyone answering got past the obstacle, and the people it stopped are not in your survey, so each reply stands in for buyers you never saw. Expect doubts about quality, unclear returns, the price or a long form. Include 'nothing' as an option so nobody has to invent a doubt, and deal with the most common real one on the page just before the buy button.
Did you switch to us from another provider, and if so, what made you move?
Why ask it
Build it as two steps, a yes or no and then a box. Whatever drove them away from someone else is the promise you have to keep: a customer who came for faster service will move on again over slow service. 'No, this is new to me' marks a first-time buyer, who needs more help at the start than a switcher does.
What, if anything, would make you stop being our customer?
Why ask it
Put to current customers, it shows the risk before it becomes a cancellation. Offer a price rise, a drop in quality, poor service, no longer needing it and a better alternative, plus an open box. Count 'no longer needing it' apart from the others, since it is the one reason you cannot fix.
What is the main reason you decided to cancel?
Why ask it
For an exit or lapsed-customer survey only, where it goes first, because someone who is leaving may answer one question and no more. Give six or seven reasons and an 'other' box. 'Too expensive' is the easy tick, so ask what they are moving to: a pricier rival means the trouble was value.
Is there anything we could have done to keep you?
Why ask it
Optional, open, and best read by somebody with the authority to act. Some replies name a small fix you could still offer, so decide beforehand whether anyone will write back. Keep the hard sell out of it, since a departing customer who is treated well sometimes returns.
Profile and closing
How long have you been a customer of ours?
Why ask it
Offer ranges: under three months, three to twelve months, one to three years, longer. If your records already hold this, attach it to the response and do not ask. Unhappy new customers point to the first weeks, and unhappy long-standing ones point to something that changed.
Which of our products or services have you bought?
Why ask it
A check-all-that-apply list that lets you read every other answer by product line. That is where an average hides trouble, with one service dragging down a score that looks fine overall. Drop it if your order system can tie the purchase to the response.
Which age group are you in?
Why ask it
Use ranges, never a birth year, and include 'prefer not to say'. Ask only if a different answer would make you do something differently. Every personal question costs some completions and adds data you must look after, and the rules on collecting and keeping it differ by place, so find out which apply where your customers are.
Which best describes your part in choosing or using our product?
Why ask it
For business customers, with four options: I chose it, I use it, both, or I approve the spending. Buyer and user are often different people with different complaints. Content buyers with unhappy users add up to a renewal at risk that the buyer's score alone will not reveal.
Which of our locations do you visit most?
Why ask it
For a business with more than one site, this turns a company-wide score into something a local manager can act on. Use a dropdown of real location names so nobody has to type. With only a few replies per site, resist ranking the sites against each other.
Would you be willing to talk with us for 15 minutes about your answers?
Why ask it
Collect an email address only on a yes. Those who agree include your angriest and your keenest customers, and a short call with a few of them teaches more than another round of the survey. State what the address is for and stick to it, and before adding anyone to a mailing list, check how marketing consent works where you operate.
What did we not ask about that you wanted to tell us?
Why ask it
Always last and always optional. The thing your questions missed turns up here, and a theme that repeats belongs in the next survey as a proper question. Thank people on the final screen and say when they can expect to hear what you did with the answers.
How to build a customer survey people finish
Practical guidance for the conversation itself
Before you write a question
Start from the decision
Write down the decision the survey is meant to inform before you write anything else: whether to change the price, what to fix in support, which product to improve next. Each question then has to earn its place by helping with that decision. A question whose answer would not change what you do is the first to go, however interesting it is.
Pick one kind of survey
A relationship survey goes to all customers once or twice a year and asks how things stand: the Overall group, a few from Price and a few from Choosing and leaving. A transactional survey follows one purchase or one support case and asks only about that. An exit survey goes to people who cancel and opens with the reason. Mixing the three produces a long form whose answers are hard to compare.
Keep it short, and know what to cut
Aim for a survey a customer can finish in two or three minutes, since every extra screen is a reason to close the tab. A workable short version is the satisfaction question, the recommend question and its reason, three or four from the group that matches your decision, and one or two profile questions. When it runs long, cut open text boxes first, then anything your own records can answer, then profile questions nobody will use.
Do not ask what you already know
Purchase date, product, plan, location and length of time as a customer usually sit in your own system. If your survey tool can attach them to each response, let it. That leaves room for the questions only a customer can answer.
Choose the moment
Send a transactional survey while the experience is fresh. Send a relationship survey in an ordinary week, not just after a price rise or an outage, unless that event is what you want to measure. Say in the invitation how long it takes and what you will do with the answers.
Wording and answer scales
One thing per question
'How satisfied are you with our speed and friendliness?' cannot be answered by someone who found you quick and rude. Split it into two rows. Look for an 'and' hiding in any question about price and quality, or delivery and packaging.
Keep the wording neutral
'How great was your experience?' and 'How much do you love the new design?' tell people which answer you want. Where it reads naturally, name both ends, as in 'how easy or difficult', and give every rating a balanced scale. Write in the words customers use, not your internal names for products and teams.
Label the scale and keep it the same way round
Five labeled points suit most satisfaction and ease questions, and the recommend question keeps its 0 to 10. Put a word on every point, not only the two ends, and balance the scale with two negative options, a middle and two positive. Keep the direction the same through the whole survey so nobody ticks the wrong end out of habit.
Give a time frame and a way out
'In the last three months' or 'on your most recent visit' tells people which memory to use. Add 'not applicable' or 'I have not used this' wherever a customer may have nothing to rate. Without it they pick the middle, and your average fills up with guesses.
Use open boxes sparingly and order the form with care
A text box after a low score is worth more than several extra scales, but three boxes in a row end surveys. Keep them optional and place each straight after the rating it explains. Put the overall questions first so the detail questions do not sway them, group the rest by subject as on this page, and leave the profile questions for the end. Before sending, try the form on a phone and on two people outside the business.
Reading the results
Report shares, not only averages
An average of 3.8 out of 5 could mean everyone is lukewarm, or that half are delighted and a quarter are furious. Show the percentage on each point, or at least the top two and the bottom two. The spread tells you whether you have one problem or two kinds of customer.
Split by group
Break the main scores down by length of time as a customer, by what people bought and by location. A flat overall number can hide a falling score in one product line. Go carefully with small groups: a handful of replies is a set of anecdotes, worth reading and unsafe to chart.
Read every comment, then count them
Sort the open answers into six to ten themes, tally each one, and keep a few customer sentences per theme word for word. The tally shows what is common, and the quotes help colleagues believe it. Read the comments from low scorers and high scorers as two separate piles.
Remember who did not answer
The people who reply are not a cross-section of your customers. The very pleased and the very annoyed tend to answer, and those who have drifted away mostly do not. Treat the results as a strong hint about direction and set them beside what you can observe: repeat orders, refunds and cancellations.
Close the loop
Reply to anyone who reported an open problem and agreed to be contacted. Then tell all customers two or three things you changed because of the survey. People who can see that the last one led somewhere have a reason to fill in the next.
Mistakes to avoid
Making every question required
Required questions feel thorough and produce abandoned surveys and made-up answers. Require only the one or two you cannot do without, usually the overall score, and let people skip the rest.
Surveying too often
A customer asked for feedback after every order learns to ignore the request. Set a rule for how often any one person can be surveyed and hold to it across teams, so marketing and support are not both asking in the same week.
Rewarding the answer you want
A thank-you for completing a survey is fine when it goes to everyone or to a random draw. A discount for a top rating, or staff who ask customers for full marks, gives you scores that describe the reward. If prizes or discounts are involved, check which rules on promotions apply where you operate.
Changing the wording every round
A score only compares with the last one if the question, the scale and the timing stayed the same. Improve the wording before the first send. After that, leave the core questions alone and change the ones around them.
Collecting answers nobody owns
Before the survey goes out, name the person who will read the results and the meeting where they will be discussed. Results that arrive with no owner get a glance and a folder, and customers who gave you their time notice that nothing changed.