Feedback questionnaire examples for customers, staff and events

Searching for a feedback questionnaire sample usually turns up a list of fifty questions. The list is rarely the hard part. Anyone who has run a form for a team already knows roughly what to ask. The hard part is deciding which questions earn their place, how to word them so people finish the form, and what happens to the answers on the Monday after they arrive.

What follows is three questionnaires, written out in full, for the three situations that come up most often: a customer who has just dealt with your team, a member of staff, and someone who attended an event. Each one is short on purpose. After the examples there is a comparison of the answer scales, a list of the questions that look useful and are not, and the part that most guides skip, which is what the answers do once they are in.

Every question is paid for twice

A question costs the respondent attention and costs the team reading time. Both bills come due, and the second one is the reason good questionnaires get abandoned. A twenty question form that produces four hundred partially completed responses is worse than a six question form that produces four hundred complete ones, because nobody reads the long version twice.

There is a test that removes most questions on the first pass. For each question, name the action that changes depending on the answer. Not the insight, the action. If the answer would go in a report and nowhere else, that question belongs in a separate research exercise rather than in a form that runs every week.

The second filter is ordering. Concrete questions go first, opinions in the middle, open text last. Someone who has already answered four easy questions will write a paragraph at the end. Someone who is asked for a paragraph first will close the tab. On a phone this matters more than on a laptop, which is why one question per screen tends to beat a single long scroll for anything longer than four questions.

Required fields deserve the same scrutiny. Mark a field required only when a blank answer makes the response impossible to act on. An email address usually qualifies, because without it there is no reply. A job title almost never does.

A customer questionnaire, question by question

This is the version to use after a support conversation, a delivery, or a completed piece of work. Seven questions, two of them optional.

1. What were you trying to get done when you got in touch? (short open text)
2. Did it get sorted? (Yes / Partly / No)
3. How many times did you have to ask? (Once / Twice / Three or more)
4. How was it handled? (1 to 5)
5. How is the product or service itself? (1 to 5)
6. Would you point a colleague here? (0 to 10)
7. Was there anything the reply missed? (open text, optional)

Questions 4 and 5 look redundant and are not. Handling and substance fail separately and get fixed by different people. A run of fives on handling next to twos on substance tells you the team is doing well with a product that is not, and the reverse pattern is the one that gets misdiagnosed as a product problem for months.

Question 3 is the sleeper. A count of repeats is harder to argue with than a satisfaction score, because it is a fact about what happened rather than an opinion about how it felt. Two identical fives mean nothing next to each other if one of them took four emails.

Question 1 sits first because it is the easiest question in the form and because the answers get used more than anything else here. Read fifty of them and the gaps in your own documentation become obvious without asking anyone about documentation.

A staff questionnaire, and the anonymity decision

Staff feedback has one structural problem that customer feedback does not. Anonymous answers are more honest and impossible to follow up. Named answers can be acted on individually and come back softer. There is no wording that resolves this, so pick deliberately rather than by default.

The workable middle is an anonymous form with one optional field at the end: a checkbox saying the respondent is happy to be contacted about their answer, with a name field that appears only if it is ticked. Enough people tick it to give you the follow up conversations, and the rest stay genuinely anonymous. What does not work is anonymity that varies per question, which reads as a trick.

A six question version:

1. What part of the week takes more time than it should? (open text)
2. Do you have what you need to do the job well? (Yes / Mostly / No)
3. If something goes wrong, is it clear who decides? (Yes / Sometimes / No)
4. How manageable is the current workload? (1 to 5)
5. What would you change first if it were up to you? (open text)
6. Happy to be contacted about this answer? (checkbox, then optional name)

One rule saves more trouble than any wording choice: do not ask about things you have no intention of changing, and do not ask about things decided elsewhere, unless you say in the form where the answers will be passed and by when. Asking about pay bands you do not set produces answers you cannot act on and a round of visible inaction that makes the next form harder to run.

An event questionnaire that gets answered before people leave

Event feedback decays fast. The form sent the next morning gets a fraction of the response of the same form put on the closing slide as a link and a code, and the answers from the morning after are shorter and more polite. Five questions is the ceiling for something answered on a phone while standing up.

1. Which session was most useful? (list of the sessions)
2. Was the length about right? (Too short / About right / Too long)
3. How was the venue and the running of it? (1 to 5)
4. What should the next one cover? (open text)
5. Would you come again? (Yes / Maybe / No)

Question 4 is the one that justifies the form. It turns feedback into a decision that has to be made anyway, which means the answers get read within the week rather than filed. Question 1 needs the actual session names rather than a free text box, because the whole point is to compare sessions against each other.

One reminder, sent the following day to people who have not answered, typically adds a meaningful share of the total. A second reminder rarely does, and costs goodwill with the people who already replied. That is only possible if the form knows who has answered, which is a property of the tool rather than of the questionnaire, and one of the practical reasons registration and feedback are worth running in the same place.

Scales: pick one shape per form

Mixing scale types in one form makes the results harder to read than the extra nuance is worth. Pick one primary scale and use it consistently.

Labels matter more than the number of points. A 1 to 5 scale with the ends labelled in plain words, such as "took far too long" through "quick", produces more usable spread than the same five points labelled only with numbers, because respondents are not guessing what the middle means.

Questions that look useful and are not

"How satisfied are you overall?" on its own. A single score with nothing next to it gives you a number that moves and no reason why. Pair it with one open question or drop it.

Long matrices. Rating communication, professionalism, friendliness and expertise as four rows of the same grid produces four nearly identical columns, because respondents answer the first row and repeat themselves down the block.

"Any other comments?" as the only open field. It gets left blank. A specific open question, such as what the reply missed, gets answered by several times as many people.

Demographic fields nobody will segment by. Department, company size and role are worth collecting only if the analysis genuinely splits by them. Otherwise they are three more fields between the respondent and the end of the form.

Double questions. "Was the response fast and helpful?" cannot be answered by someone who got a fast unhelpful reply, and that person is exactly who you needed to hear from.

Leading wording. "How much did you enjoy the session?" sets a floor under the answers. "How was the session?" does not.

The answers are the cheap half

A questionnaire is a few hours of drafting. The work that follows is ongoing, and it is where feedback programmes quietly stop. Responses arrive as notification emails, get read once, and settle into an inbox with no record of which ones were answered and by whom. Three months later the form is still collecting and nobody can say what changed because of it.

What prevents that is not a better questionnaire. It is three properties of wherever the responses land. Each response needs an owner, so that a reply is somebody's job rather than everybody's. It needs a status, so that "read" and "replied" and "closed" are distinguishable at a glance. And the reply needs to be sendable from the same screen as the answers, because the friction of switching to an email client is what turns a two minute reply into a task for later.

A spreadsheet export can hold the first two of those with enough discipline. It cannot do the third, and it loses the connection between the answer and the reply as soon as two people are working from it. Tools built around response management treat owner, status and history as part of the record rather than as columns someone maintains by hand, and the situations where that pays off are the recurring ones: the same form running every month, with more than one person reading it.

The volume question is worth checking before committing to anything. Feedback forms arrive in bursts, which is exactly the shape that suits seat based pricing and fits badly with a monthly response allowance. Typeform's Basic plan is 39 USD a month with 100 responses included and one user. Its Plus plan is 79 USD a month with 1,000 responses and three users. SurveyMonkey's team plans are billed per user with a minimum of three, and responses beyond the plan allowance are charged individually. Tally offers unlimited forms and submissions on its free tier under fair usage guidelines, with a paid Pro plan at 24 USD a month. None of those is wrong, but a customer feedback form that gets quiet for a quarter and then busy after a launch is a poor match for a meter that counts responses.

What to change first

Cut the questionnaire you have to six questions, move the open text to the end, and add the repeat count question, because it is the one that produces an argument nobody can wave away. Then decide where the answers will live and who owns a reply before the next batch arrives, since that decision is the one that determines whether any of this gets read twice. If the current answer is a shared inbox, Halict shows what the alternative looks like on a single response.

Q1. How many questions should a feedback questionnaire have?

Five to seven for anything that runs regularly, and no more than five if people will answer it on a phone. The limit is not the respondent's patience so much as the team's reading time, since a long form produces answers nobody revisits. Add questions only when you can name the decision that changes based on the answer.

Q2. Should a staff feedback questionnaire be anonymous?

It depends on whether follow up matters more than candour. Anonymous answers are more honest and cannot be acted on individually, which is why the practical compromise is an anonymous form with an optional checkbox letting the respondent offer their name for a follow up conversation. Varying anonymity question by question reads as a trick and lowers response rates.

Q3. What is the best rating scale for feedback forms?

For tracking the same thing across months, a 1 to 5 rating with the two ends labelled in plain words works well. For outcomes, a three point scale such as Yes, Partly and No beats a five point one because the answer is a fact rather than a feeling. A 0 to 10 recommendation question needs roughly a hundred responses before the result means anything.

Q4. When should an event feedback form be sent?

At the end of the event, as a link and a code on the closing slide, with one reminder the following day to people who have not answered. Response rates drop sharply once attendees have left, and late answers tend to be shorter and more agreeable. Sending a second reminder rarely adds much and annoys the people who already replied.

Q5. How do you stop feedback from being read once and forgotten?

Give every response an owner and a status, and reply from the same place the answers are stored. Notification emails fail at this because there is no shared record of which responses have been handled, so two people either duplicate a reply or neither sends one. Exporting to a spreadsheet solves the record keeping and not the reply.