Cancellation survey template: ask after the cancellation, not during

Nine questions for somebody who has already gone: why they left, when they decided it, what would have kept them, and whether they would come back. It sits at its own link, so it reaches the expired card and the quiet lapse as well as the person who clicked cancel. Answer it below, then copy the questions or open the whole thing in the editor.

9questions
About 2 minto complete
One linksent after the ending
No offersit asks, it does not bargain

Use this template Take the survey first

Cancellation Survey · answer to move through it

Use this template

Question 1 of 9 In your words

In your own words, why did you cancel?

Question 2 of 9 The reason

Which of these comes closest to your main reason for cancelling?

I stopped using it

It cost more than it was worth to me

I only needed it for one project, and that is finished

Something I needed was missing

It was harder to use than I expected

Something went wrong and it was not put right

I moved to something else

None of these

Question 3 of 9 The decision

When did you decide to cancel?

In the last week or so

A few weeks ago

A month or two ago

Longer ago than that

I cannot remember

Question 4 of 9 The decision

What was happening at the time you decided?

A renewal or a payment came up

Something stopped working the way I needed

I asked for help and it did not go well

The price or the plan changed

My own situation changed

Nothing in particular, it built up over time

Question 5 of 9 What you could have done

Before you cancelled, had you told us there was a problem?

Yes, I contacted support about it

Yes, in a survey or a review

No, I did not raise it

There was nothing to raise

Question 6 of 9 What you could have done

Is there anything that would have kept you?

A lower price or a smaller plan

The thing that was missing

Fixing the problem I ran into

Better help when I asked for it

The option to pause instead of cancel

Nothing would have kept me

Question 7 of 9 What you could have done

Where have you gone instead?

To another product or service

Back to doing it by hand

To something free

Nowhere, I stopped doing this altogether

I have not decided yet

Question 8 of 9 The close

Would you consider coming back?

Yes, if the reason I left were fixed

Maybe, later

Question 9 of 9 The close

Overall, how was the cancelling itself?

DifficultEasy

What the cancel screen never sees

Almost every cancellation survey in this category is a widget that fires inside the cancel flow. This one is not, and the difference decides which of your leavers you ever hear from.

Ending a subscription is not always a click. A card expires and nobody replaces it. A trial runs out. An app store subscription is stopped somewhere you cannot see. Somebody at the customer's end closes the account for a whole team, or a script does it through an API. A survey that fires on a cancel button never meets any of those people, and one widget vendor writes the limit into its own documentation: Memberful records that cancellations completed by an admin or through the API do not appear in its survey results.[2]

The slot is narrow as well as selective. An in-flow survey stands between a customer and the thing they are trying to do, which is leave. Memberful's is a single question, shown once the cancellation is confirmed, and the respondent may skip it.[2] That is the honest size of that moment: room to count a reason, and no room to learn anything about it.

And the company asking is the one being left. A person mid-exit, often looking at a discount or a pause offer on the same screen, is answering the organisation they are walking away from. The same nine questions opened a day later, alone, with nothing left to negotiate for, are a different conversation.

What a link gives up, plainly. It has to be opened, so it competes with everything else in an inbox, and you choose the send rather than catching the second the decision lands. Those are real costs. What you buy with them is everybody the cancel screen cannot see, and enough questions to make an answer readable.

So send it twice, from two copies. Put the link in the message that confirms the ending, and follow up about a week later to the people who did not answer. Then publish a second copy of the same nine questions for the endings your system notices rather than the customer, the failed payments and the lapsed renewals, and read the two sets separately. Same instrument, two lists, and no guessing about which kind of ending an answer came from.

Who sees the answers. They arrive in the account the survey was made in, and this template asks for no name and no email address.

A narrow funnel catching one falling stream of markers while four others fall past it, beside a wide tray joined by a link symbol that catches all five — A cancel-flow survey catches the endings that came through a cancel button. A link catches those and the endings nobody clicked.

If the ending you want to understand is not a cancelled subscription, three closer templates exist. For somebody who has quietly let a membership lapse rather than a subscriber who cancelled, the membership survey questions stay on the roster and ask about renewal. When the whole story is one bad support request rather than the product, the customer service questionnaire covers that single ticket. And if the person leaving is an employee rather than a customer, the exit interview form is a different instrument built for a different conversation.

Your reason list is part of the answer

Every page in this category prints a list of reasons to choose from. Not one of them mentions that offering the list is part of what produces the answer, and there is a measured figure for how much.

The number. Pew Research Center asked the same question two ways in a poll after the 2008 US presidential election: what one issue mattered most in deciding how you voted. Given a list with the economy on it, 58 per cent of respondents picked the economy. Asked the open version, with no list at all, 35 per cent volunteered it.[1] Same poll, same subject, twenty-three points apart.

Why that lands here. Contentsquare publishes a churn template with nine reason options and one of them is that the price is too high.[3] Put a price option in front of people and you will collect price answers. That is not evidence your price is wrong. It is evidence you offered the option, which is a different finding and a much cheaper one.

So the open box goes first. Pew asks open questions near the beginning of a questionnaire for this reason: closed questions placed before an open one make respondents much more likely to mention the concepts those earlier questions raised.[1] In this template question 1 is a blank box and question 2 is the list. Turn them round and you lose the only answer nobody wrote for the customer in advance.

Then shuffle the list. In self-administered surveys people lean towards items at the top of a list, while in telephone surveys they lean towards items heard later.[1] A link survey is self-administered, so the top of your list carries an advantage that has nothing to do with the reasons on it. The editor can randomise option order per question: turn it on for question 2, and leave question 3 alone, because that one is a timeline and its order means something.

And write the list from your own answers. Take the open box from your first fifty replies, sort them by hand, and let the six or seven groups you actually find become your options for the next quarter. A reason list inherited from a template is a list of somebody else's customers.

Two panels: a printed list of six option tiles feeding a chart with one very tall highlighted bar, and an empty open frame feeding a chart of evenly matched bars — What Pew measured: offered the economy inside a list, 58 per cent chose it; asked with no list, 35 per cent volunteered it.

The 9 cancellation survey questions

Every question, the answer scale it uses, and what to do with what comes back. Copy one block, copy all nine, or open the survey and rewrite the options to match your own product.

Plain text, one question per line, with the answer scale.

In their words, then in yours

The open box first and the reason list second. Pew states that closed questions placed before an open one on the same subject make respondents much more likely to mention the ideas those earlier questions raised,[1] which is exactly what a reason list does. Question 1 is optional so nobody is held at a blank box on the opening screen.

In your own words, why did you cancel?Open text, optionalThe only place somebody can name a reason you did not think to put on the list. Read these before you count anything else.
Which of these comes closest to your main reason for cancelling?Choose oneEight options, because that is what it takes to cover the ways a subscription actually ends, and a last one that lets somebody decline the whole frame. Switch on option-order randomisation for this question before you send it.

When the decision was made, and what made it

The click and the decision are two separate events, and the gap between them is the part of the account life where anything could still have changed. Every page in this category collapses the two into one question. These are two taps and both are countable.

When did you decide to cancel?Choose oneSort the answers by this before you read anything else. A decision taken months ago and a decision taken last week point at different parts of the product.
What was happening at the time you decided?Choose oneThe trigger, kept separate from the reason. Somebody can decide during a renewal notice and still be leaving for something that has nothing to do with the price.

What you could have done

Three taps that turn a reason into work somebody can pick up: whether you ever heard about it, what would have changed the outcome, and where the customer went. None of them offers the respondent anything, because this form asks and does not bargain.

Before you cancelled, had you told us there was a problem?Choose oneSplits your answers into two piles that need different work: problems you were told about and left unfixed, and problems you had no way of seeing.
Is there anything that would have kept you?Choose oneAsked once, after the ending, so it is a question rather than an offer. The pause option is there to be counted, not to be granted.
Where have you gone instead?Choose oneThe difference between losing a customer to a competitor and losing the whole job to nobody. Only one of those is a product problem.

The door, and the exit itself

The last two questions are the cheapest on the form. One sorts your win-back list, the other rates the only part of your product this person used most recently. Ending on a single tap keeps the closing screen from asking for one more thing.

Would you consider coming back?Choose oneThree options, kept separate from the reason, so you can write a different message to each group later.
Overall, how was the cancelling itself?5-point rating, Difficult to EasyA hard cancellation is something you can fix this quarter, and it is the last impression this customer takes away with them.

All nine are already built and wired to a live link, so a free survey builder can have them rewritten in your own words in a couple of minutes.

No score, and no recommend question. Nothing here is added up into a number. The people who answer a cancellation survey chose to answer it after leaving, so a percentage taken from them describes the people who replied and nobody else. If a comparable figure is what you were sent to get, it belongs on a survey asked of customers who are still here.

Reading a pile of leaving reasons

What comes back is a small, self-selected pile of answers from people with no reason left to be diplomatic. That makes it unusually honest and unusually easy to over-read. Here is the order to take it in.

What you are looking at	Where it comes from	What it settles	What it cannot tell you
The written answers	Question 1, given before anybody saw an option list	The words to put in front of the people who build the product	How common any of it is
The reason counts	Question 2, ranked	A short ordered list you can argue from in a meeting	Whether the same order would survive a different list of options
The decision gap	Questions 3 and 4 read together	Where in the life of an account the mind was actually made up	What happened between the decision and the click
The two piles	Question 5	Whether you were told and did nothing, or never heard at all	Why somebody chose not to raise it with you

Read the written answers before you open a chart. Question 1 was answered before the respondent saw a single option, so it is the only part of the form that is entirely theirs. Ten of them read in one sitting will change what you go looking for in the counts.

Then count question 2, and hold the count loosely. The ranking is real, but it is a ranking of the options you supplied. Report it as what people picked from the reasons you offered, put the option list next to it, and rewrite that list for next quarter from the written answers you just read.

Cross question 3 with question 2. Somebody who decided last week and somebody who decided eighteen months ago can pick the same reason and mean two different things. A missing feature named by a person who made their mind up in week two is a story about the start; the same words from last week are a story about now.

Split on question 5 before you brief anybody. Everyone who told you and left anyway is a follow-up problem with a name attached. Everyone who never told you is a product problem you had no way of seeing, and this survey is the only place it surfaces.

Two things not to do with it. Do not turn the counts into a headline churn-reason percentage unless the option list is printed beside it. And do not send these nine questions to people who are still customers: every one of them assumes an ending that has already happened.

Where the question is not an ending at all. A customer who is still weighing it up needs a different measurement, and switching risk asks the person who has not gone yet. When what you want is the whole trip rather than the exit from it, the customer experience questionnaire walks it stage by stage and points at the stage worth a survey of its own.

Cancellation survey FAQ

What is a cancellation survey?

A short set of questions asked of somebody whose subscription, membership or service has ended, about why it ended and what would have changed it. It is also called a churn survey, and when the wording is customer-side, a customer exit survey. The three names describe the same instrument pointed at the same person. What changes between them is whether you are counting reasons across a whole account base or reading one ending closely.

When should you send a cancellation survey?

In the message that confirms the ending, and again about a week later to the people who did not answer. Sending it inside the cancel flow means only the people who cancelled by clicking cancel are ever asked: Memberful documents that cancellations completed by an admin or through its API never appear in its survey results at all.[2] Leaving it a month asks somebody to reconstruct a decision they have stopped thinking about.

How many questions should a cancellation survey have?

This one has nine and the number follows from where it is asked. A survey that fires inside the cancel flow has room for one or two, because it is standing between the customer and the exit; Memberful ships a single question there.[2] A survey opened later is opened deliberately, so it can afford the follow-up questions that make a reason readable. Nine takes about two minutes. A published list of thirty-eight is a menu to choose from, not a survey to send.

Is this an in-app cancel flow, a pause offer or a deflection step?

No, and it does not turn into one. It is a survey at its own link, which you can also embed in a page of your own. It does not open when somebody clicks cancel, it does not watch for an exit, and it never offers a discount, a pause or a downgrade. Mixing an offer into these questions changes what you are measuring, because somebody weighing up a discount is no longer only answering you.

What about customers who never cancelled and just stopped paying?

They are the group an in-flow survey cannot reach at all, and they deserve their own copy of this one. Publish the nine questions a second time, send that copy when a payment fails or a renewal lapses, and read the two sets separately. Question 1 still works unchanged: somebody whose card expired and who chose not to fix it has a reason too.

Is this the same as a post-purchase survey?

No. A post-purchase survey is about one order that arrived, and the post purchase survey questions are timed to a delivery on the assumption that the relationship carries on. This one is about a subscription that stopped, asked of somebody who is no longer a customer and has nothing left to negotiate for.

Methods and sources

Michael Hodge

Survey methodology and questionnaire design, SuperSurvey

Survey methodologist and editor since 2003: wording, response scales, reliability checks, reporting that maps to a decision.

Nine questions written for this page, carrying no item from any published or licensed instrument and no 0 to 10 recommend question
Question count and completion time read off the survey running on this page, not typed from a description of it
Every source quoted here was fetched and the wording checked against the live page on 7 September 2026

Updated 8 September 2026 · How templates are reviewed

What this template is based on

The question order is the page's main design decision and it follows published guidance rather than preference: the open box precedes the reason list because closed questions asked first on the same subject shape the open answer that follows,[1] the advice to randomise the option order on question 2 follows the primacy effect reported for self-administered surveys,[1] and the scoping argument for sending a link rather than firing a widget rests on a widget vendor documenting which cancellations its own survey never sees.[2] The observation that a published reason list carries a price option is checkable against the list itself.[3]

Question licensing

Nothing here reproduces a licensed, paid or standardised instrument, and no published cancellation or churn instrument exists to reproduce. Every question was written for this page and is free to copy, edit and send anywhere, with or without attribution.

Limits and disclosure

No response-rate, completion or churn-rate benchmark is offered for this survey. The figures circulating in this category come almost entirely from companies selling cancellation-flow software and measure deflection rather than survey quality, so none of them is quoted here. The answers are also self-selected: people who leave quietly are less represented than people who leave loudly, and no weighting fixes that. Nothing on this page is legal advice about subscriptions or cancellation rules.

SuperSurvey makes the editor this template opens in. Everything on this page works with or without it.

References

Pew Research Center. Writing Survey Questions. Methods, Pew Research Center. pewresearch.org
Memberful. Cancellation survey. Memberful product documentation, metrics and analytics. memberful.com
Contentsquare. Churn survey template. Survey template library, read 7 September 2026. contentsquare.com

Send it to the people who already left

Nine questions, about two minutes, no name asked and no offer attached. Open it in the editor, rewrite the reason list to match your own product, and put the link in the message that confirms the ending.