CANCEL FLOWS

Cancellation offers: types, examples and how to match them to why customers leave

Cancellation offers explained: discount, pause, downgrade, help. How to match each to the reason for leaving, size a discount and measure real saves.

Published 10 min read

Cancellation offers: types, examples and how to match them to why customers leave

A cancellation offer, also called a save offer or cancel-save offer, is a proposal shown to a subscriber at the moment they try to cancel: a discount, a pause, a cheaper plan, a short extension or some help. The offers that work are matched to the reason the customer gives for leaving, not handed to everyone as a blanket discount.

This guide is for SaaS and subscription teams who want to design cancellation offers, put them in a cancel flow and tell whether they are working.

TL;DR

  • A cancellation offer is a targeted proposal made inside the cancel flow, after the customer has said they want to leave and ideally after they say why.
  • Match the offer to the reason: a pause for low usage, a downgrade for overbuying, a discount for budget, support for confusion or missing features.
  • Bigger is not automatically better: test discount levels instead of starting with the deepest cut.
  • Count a save only when the next invoice is paid, not when the customer clicks “accept.”
  • Always keep a clear, working path to complete the cancellation.

What cancellation offers are and where they sit

A subscriber who clicks “cancel” has already decided something. A save offer does not argue with that decision. It checks whether a different arrangement would solve the problem that caused it.

Where a cancellation offer sits
  1. 01 Cancel entry point A button in account settings, the billing portal or the app
  2. 02 Reason question One multiple-choice question about why they leave
  3. 03 The offer One proposal chosen from the answer, with "No thanks, continue to cancel"
  4. 04 Confirmation The cancellation completes, with an end date and a way to come back

The flow usually has four parts:

  1. The cancel entry point. A button in account settings, in the billing portal or in the app.
  2. A reason question. One multiple-choice question: too expensive, not using it enough, missing a feature, switching to another tool, temporary need, other.
  3. The offer. One proposal chosen from the answer, with a plain “No thanks, continue to cancel” option.
  4. The confirmation. The cancellation completes, with an end date and a way to come back.

This is voluntary churn: the customer chooses to leave, as opposed to a card that fails. Failed-payment churn is a separate problem with a separate fix (retries, card-update reminders), and it deserves its own process.

What are the main types of cancellation offers?

Most save offers are one of six kinds. Each answers a different objection.

Discount

A percentage off, or a fixed amount, for a set number of billing cycles. It answers “this costs too much.” It is the most common offer and the easiest to overuse, because it lowers revenue on every account that accepts it.

Pause

The subscription stops billing for a set time, often one to three months, then resumes. It answers “I am not using this right now.” The customer keeps their data and settings, and you keep the relationship.

Downgrade

A move to a cheaper plan with fewer seats, lower limits or fewer features. It answers “I am paying for more than I need.” You keep some revenue instead of none.

Plan switch or billing-term change

Moving from monthly to annual at a lower effective price, or from annual to monthly for a customer who dislikes the commitment. It answers cash-flow and commitment concerns.

Extension or free period

A free month or an extended trial of a feature the customer has not tried. It suits customers who have not reached the value yet.

Human help

A support chat, a short call or an onboarding session. It answers “I could not get it to work” or “I am missing a feature,” and it costs you time instead of margin.

For a deeper comparison of the first three, see our guide to pause vs discount vs downgrade offers.

How do you match the offer to the reason for leaving?

The single most useful design rule is to ask first, then offer. The reason a customer gives tells you which offer has a chance. Price and low usage are both common reasons, and they need different answers: a discount can help a customer for whom the price is the problem, but it does little for one who does not use the product, because paying less does not make anyone use it more.

A practical mapping:

Reason given Best first offer Why it fits Fallback
Too expensive Time-limited discount, or a downgrade Addresses price directly Annual plan at a lower effective price
Not using it enough Pause Removes the charge without losing the account Downgrade to a lower tier
Missing a feature Human help or a roadmap note A discount does not add the feature Pause until it ships
Too complicated, could not set up Onboarding session or support chat Fixes the real obstacle Free period while they finish setup
Temporary need ended Pause with a set resume date Fits a seasonal or project-based use Win-back email later
Switching to another tool Short survey follow-up, no offer A concession rarely changes this decision Win-back email later
Other or no answer Light-touch pause or downgrade Low cost to you None

Keep the number of offers per flow small. One well-chosen offer per reason beats a menu of five, and a single question keeps the flow short. Short flows also respect the customer, which matters for trust and, as covered below, for compliance. The cancellation reasons survey guide covers how to word the question.

How big should a discount be?

There is no universal number, and acceptance does not rise steadily with the size of the discount. Past a point, a larger discount costs you margin without winning more customers, so start low and test upward rather than starting with the deepest cut.

A sensible way to set the level:

  • Start with the smallest discount you would accept being asked for by every budget-driven cancel, then test upward.
  • Limit the discount to a fixed number of billing cycles and state when the full price returns.
  • Never offer the same discount again immediately after it ends, or customers learn to cancel to get it.
  • Compare the discounted price against what a downgrade would earn. A customer on a cheaper plan at full price can be worth more than a customer on the top plan at half price.

A worked example: a customer pays $99 a month and says the product is too expensive. A 40% discount for three months costs you $39.60 a month for those three months, $118.80 in total, if they stay. If they leave at the end of the discount, you have bought three months. If a downgrade to a $59 plan keeps them for a year, you earn $708 over that year. The discount only wins if the customer then stays at full price long enough to cover the difference. That is why the next section matters.

How do you measure saves?

The most common reporting mistake is to count every accepted offer as a save. Acceptance tells you the customer clicked. It does not tell you the customer paid. A save should count when the next invoice is actually paid, and a paused customer who never resumes is a delayed cancellation: a headline save rate that includes them overstates success.

A measurement routine that works for most teams:

  1. Log every cancel attempt with the reason chosen, the offer shown and the outcome (accepted, declined, abandoned).
  2. Mark a save only after the next successful charge. For a pause, that is the first charge after resume.
  3. Track retention by cohort. Look at the saved customers at 30, 60 and 90 days after the offer, and compare them with customers who cancelled without one.
  4. Split results by reason. A pause may perform well for low usage and badly for budget. Averages hide this.
  5. Watch revenue, not only count. Add up the discount you gave away and the revenue it retained. Net revenue retention and gross revenue retention are the two ratios most finance teams use for this.

If you have enough traffic, hold back a small share of cancel attempts as a control, with a plain cancel and no offer, so you can see how many of the saved customers would have stayed anyway or would have returned on their own.

Which risks should you design around?

A poorly matched discount can reduce revenue without solving the reason for leaving: it can subsidize customers whose real issue is low perceived value. Three habits keep that from happening.

Match, then measure. Use the reason-to-offer mapping above and judge it by paid retention, not acceptance.

Do not train customers to cancel for a deal. Limit how often one account can receive an offer, and vary the offer by reason instead of always showing the same coupon.

Keep the exit honest. The offer must be easy to decline, and the cancel path must lead to a completed cancellation. This is both good practice and, in the US, a legal concern. The Federal Trade Commission’s 2024 Click-to-Cancel amendments, which required cancelling to be no harder than signing up, were vacated by the Eighth Circuit in July 2025. A May 2026 Jones Day analysis reports that the agency has begun the process of reintroducing a version and continues to pursue deceptive subscription practices. State automatic-renewal laws add their own requirements, so have your counsel check your flow against the states where your customers live. Our guide to click-to-cancel rules sets out what still applies, and cancel flow best practices shows what an honest flow looks like in practice.

How do you implement save offers?

The steps are the same whether you build in-house or use a tool. Here is where each one tends to go wrong.

1. Intercept the cancel action

You need a place in your product or billing portal where the cancel request passes through your own flow before it reaches the billing system. The usual failure is a cancel path that bypasses the flow, such as a link in the payment provider’s hosted portal, so some customers never see an offer. Platform-specific walkthroughs are in our guides for adding a cancel flow to Chargebee and to Zuora. Mobile apps have their own rules; see reducing App Store cancellations while keeping the cancel button.

2. Capture the reason

Use one required multiple-choice question with an optional free-text box. Each extra question adds friction and lowers completion. Store the answer against the customer record so it can feed later win-back messages.

3. Map reasons to offers

Start from the table above and adjust to your pricing. Decide each offer’s limits before launch: how many cycles, which plans, how often one customer can use it.

4. Apply the offer in billing

This is where most home-built flows break. A discount must be created as a coupon or price change on the right subscription, a pause must set a resume date, and a downgrade must prorate correctly. A mismatch between what the screen promised and what the next invoice shows is a quick route to a support ticket and a chargeback. Test each offer against a test-mode customer in your billing system (Stripe, Chargebee, Paddle, Recurly…) and read the resulting invoice before it goes live.

5. Follow up with a win-back sequence

Some customers decline every offer. Their reason is still useful: a customer who left over a missing feature is a good candidate for a message when the feature ships. Our guide on winning back canceled Stripe customers covers the sequence.

6. Review monthly

Check paid retention by reason and offer, retire what does not hold up and test one change at a time.

How NoChurn helps

A cancellation offer works best as an answer to a specific problem, not a coupon on a timer. NoChurn runs the cancel flow as a whole on top of your billing (Stripe, Chargebee, Paddle, Recurly and others, or any billing by API): a reason question, offers chosen from the answer, the change applied on the right subscription, a “Cancel now” button visible at every step, and win-back messages for those who still leave. It takes no commission on what it saves: the price depends only on your MRR. See cancel flows.

FAQ

What is a cancel save offer?

It is a proposal shown inside the cancel flow that gives a subscriber a reason to stay: a discount, a pause, a downgrade, extra help or a free period. The best ones are chosen from the reason the customer gives. The customer can always decline and finish canceling.

Which cancellation offer works best?

It depends on the reason. Budget-driven cancels respond to a discount or a downgrade, low-usage cancels to a pause, and missing-feature or setup problems to human help. A deeper discount does not always convert better, so test levels rather than assuming.

How do I know whether my save offers actually work?

Count a save only when the next invoice is paid, then compare retention at 30, 60 and 90 days against customers who canceled without an offer. Break the results down by cancellation reason, since the averages can hide an offer that fails for one group.

Offering an alternative is allowed, but the customer must still be able to cancel without obstruction. The FTC’s 2024 rule was vacated, yet the agency is working on a new version and continues to act against deceptive subscription practices, and many states have automatic-renewal laws. Review your flow with your legal counsel.

Can I run save offers on Stripe, Chargebee or Paddle?

Yes. These billing platforms support coupons, pauses and plan changes, and NoChurn connects to them to apply the offer chosen in the cancel flow. Our platform guides above show the setup for several of them.

Sources

All accessed 16 November 2026.

PROMPT

Read https://www.nochurn.sh/llms.txt,
then follow https://docs.nochurn.sh/quickstart/
to add NoChurn to this app. Use a test key
(nc_test_…) until I say we go live.

Stop leaking revenue.

  • Free until $300 MRR
  • No card needed
  • Live in 10 minutes
  • No commission