The problem with customer feedback is not that people lie. It is that vague questions get vague answers, and "how satisfied were you" is the vaguest question in common use.
What follows is grouped by what you are trying to find out. Take one group. A survey with all thirty gets abandoned at question nine and tells you less than three good questions would have.
If you want a number you can track
Pick one, ask it the same way every time, and resist the urge to improve the wording. The moment you reword it, this quarter stops being comparable with last quarter, and the trend was the only reason to collect a number at all.
- How likely are you to recommend us to a friend or colleague? (0 to 10)
- How satisfied were you overall? (1 to 5)
- How easy was it to get what you needed? (Very easy to very hard)
- How well did we meet your expectations?
- Would you buy from us again?
- How does this compare to what you used before?
If you want to know what to fix
These are the questions that produce a to-do list. The trick in all of them is forcing a single answer: "what should we improve" gets you three mild suggestions, and "what is the one thing we should fix first" gets you the thing that is actually wrong.
- What is the one thing we should fix first?
- What nearly stopped you from buying?
- Where did you get stuck?
- What took longer than you expected?
- What did you expect to find that was not there?
- Was there a moment you nearly gave up?
- What should we stop doing?
- What do our competitors do better?
If you want to know what to keep
Almost nobody asks this, and it is how products lose the thing that made people choose them. A redesign that improves nine metrics and quietly removes the one feature your best customers stayed for is a bad trade that no satisfaction score will catch.
- What do we do better than anyone else?
- What would you miss most if we removed it?
- What should we definitely not change?
- What made you choose us originally?
- What is the one word you would use to describe us?
- What do you tell people about us?
If you want to understand a specific experience
Ask these immediately after the thing happened. Satisfaction decays into a general impression within about a week, and after that you are measuring how somebody feels about your brand rather than how the delivery went.
- What were you trying to do?
- Were you able to do it?
- How many times did you have to contact us about this?
- Did you have to repeat yourself?
- How long did it take compared to what you expected?
- Was anything confusing or unclear?
If you want something you can quote
Testimonials written to order are useless, because "would you write us a testimonial" produces a blank page. These questions produce usable copy as a side effect of being answered honestly.
- What were you struggling with before you found us?
- What changed after you started using us?
- What result can you point to?
- How much time does this save you?
- What would you say to someone considering us?
- Can we quote you, and may we use your name?
"What should we improve" gets three mild suggestions. "What is the one thing we should fix first" gets the thing that is actually wrong.
Four wording changes that do most of the work
These apply to every question above and matter more than which questions you pick.
- Ask about behaviour, not intention. "Have you recommended us to anyone?" beats "would you recommend us?" People overestimate what they will do and report what they did reasonably accurately.
- Force a single answer. "The one thing" is the most valuable phrase in survey writing. It converts a list of mild preferences into a priority.
- Do not put the answer in the question. "How useful was our new faster checkout?" assumes two things and gets agreement to both.
- Ask about the specific, not the general. "How was your delivery last Tuesday" produces detail. "How is our delivery service" produces a shrug.
How long, and how often
Five to eight questions, under two minutes. Response rate falls sharply past that, and worse, the people who finish a long survey are systematically different from the people who abandoned it. You end up with precise data about your most patient customers.
Quarterly for a tracked score, immediately after the event for anything about a specific experience. Monthly surveys to the same list train people to ignore you, and the score moves less than the noise in your sample anyway.
What to do with the answers
Read the first fifty open-text answers yourself, in full, before any summarising. That is where the patterns come from, and no tool will hand them to you if you have not seen the raw material. After fifty, the returns drop off quickly and grouping by theme is enough.
Then close the loop. Tell people what you changed because of what they said, even if it is one thing. It is the cheapest retention work available, and almost nobody does it, which is exactly why it works.
Questions to stop asking
These appear on most feedback forms and earn their place on none of them.
- "How satisfied are you with our service?" Too vague to act on. Ask about a specific interaction instead.
- "Is there anything else you would like to tell us?" as the only open question. It is a fine closer and a terrible main question, because it gives no direction at all.
- "How did you hear about us?" on a feedback form. It belongs on a signup form, and nobody aggregates the answers anyway.
- Twenty rating scales in a grid. Efficient to build, miserable to answer, and nearly unusable on a phone.
- "Rate our website, our product, our support and our pricing." Four different departments in one question produces an average that means nothing to any of them.
Getting more responses
Response rate is mostly decided before anybody reads the first question.
- Say how long it takes, and be right. "Four questions, about a minute" outperforms no estimate substantially.
- Ask at the right moment. Right after the thing you are measuring, not on a schedule that suits your reporting cycle.
- Put the first question in the email itself if your tool supports it. One click in and people tend to finish.
- Say what you will do with it. "We read every one and it decides what we build next" is worth more than a prize draw.
- Then prove it afterwards. Telling people what changed because of the last survey is the single biggest thing you can do for the next one.