Ask one question when customers cancel

Most companies learn why customers leave by guessing. The cancel flow is the one moment every churning customer passes through, and one well-placed question turns it into your best source of retention research.

Why it works

Customers who are leaving have already decided, so they have little reason to hold back. One question at that moment gets an honest answer from people you would struggle to reach later. Because nearly everyone who cancels passes through the same screen, the answers cover your whole churn, including customers who never reply to survey emails.

The answer also tells you which customers you can still help. Someone leaving because of a temporary need or a tight budget may take a pause or a cheaper plan if it appears right then.

How to do it

1. Ask one multiple-choice question: “What’s the main reason you’re cancelling?” Options such as: too expensive, missing a feature, switched to another tool, not using it enough, only needed it temporarily, other.
2. Add an optional free-text box after the choice. A sentence in the customer’s own words is worth more than the category.
3. Respond to the reason. For “too expensive”, offer a downgrade or pause. For “only needed it temporarily”, offer a pause. For “missing a feature”, link to the roadmap or a workaround. Never block the cancellation.
4. Send the answers somewhere visible: a shared channel or a weekly digest, not a spreadsheet nobody opens.
5. Review monthly and size each reason. The biggest category is usually your next retention project.

Example

A subscription tool finds that a large share of cancellations say “only needed it for a project.” It adds a “pause for 1–3 months” option in the cancel flow. Some of those customers pause instead of leaving and return when their next project starts, and the team now has a clear signal that a pay-per-project plan is worth exploring.

Metrics to track

• Response rate to the cancel-flow question. A low rate suggests the question is confusing or easy to miss.
• Share of cancellations by reason, tracked month by month to see whether a fix shrinks its category.
• Save rate per offer: the share of customers shown a pause or downgrade who accept it.
• Return rate of paused customers once their pause ends.
• Retention of saved customers over the following months, to confirm saves are real and churn was not simply delayed.

When it backfires

• Dark patterns. Hiding the cancel button or chaining multiple offers damages trust and may break consumer protection rules.
• Asking without acting. If nothing changes based on the answers, the survey is just friction.
• Too many options. Long lists make “other” the default answer. Six options is plenty.

Questions people ask

Should the cancellation survey be mandatory?

Keep it to one quick, skippable question, or make it mandatory only if it is a single click. Long forms between a customer and the cancel button create resentment and support tickets.

Is it legal to make cancellation harder?

Rules vary by country, and several jurisdictions require cancelling to be as easy as signing up. A survey should never block or delay the cancellation itself. Check the rules where your customers are.

How many responses do I need before acting on the results?

Enough that one reason clearly stands out over a month or two. For a small company that may be only a few dozen answers, so read the free-text replies closely as well as counting categories.