Cancellation Flows
Date: 2026-09-27
The journey a subscriber takes to cancel is one of the few places a retention intervention reaches someone at the exact moment of leaving. Offering a pause, a skip or a cheaper plan saves real customers; making the exit hard only delays churn, damages trust, and in the UK is heading from bad practice to unlawful.
A cancellation flow is the sequence of screens or steps between a subscriber choosing to cancel and the cancellation taking effect, including any alternatives offered along the way.
The shape of a fair one
"Cancel subscription"
│
▼
Why are you leaving? (one question, optional, multiple choice)
│
├── too much product → offer: skip a delivery / change frequency
├── too expensive → offer: smaller plan / one-off discount
├── going away → offer: pause for 1–3 months
├── didn't like it → no offer; ask what was wrong
└── other / skip → no offer
│
▼
One offer, matched to the reason — with an equally prominent "No, cancel"
│
▼
Cancelled. Confirmation by email.
The reason question does two jobs: it routes to a relevant offer, and it’s the best churn-reason data the business will ever get — Voice of Customer Data.
Saves versus obstruction
| Retention | Obstruction |
|---|---|
| One alternative, matched to the stated reason | Several screens of offers before the button appears |
| ”Cancel” as prominent as the offer | ”Cancel” as a faint text link; the offer as a big button |
| Online, in the same place you signed up | ”Call us to cancel” — phone only, limited hours |
| Confirms the cancellation clearly | Ambiguous final state; charges continue |
The right-hand column is the roach motel pattern — easy in, hard out — Deceptive Design.
The law in the UK
The Digital Markets, Competition and Consumers Act 2024 (DMCC Act) sets up a subscription contracts regime. As of the government’s 2026 response, it’s expected to take effect in spring 2027. Among its requirements:
- Consumers must be able to end a subscription in a straightforward way, and online if they joined online. An email address alone is unlikely to count, and cancelling the direct debit at the bank doesn’t count as the trader’s exit route
- Reminder notices before a free trial or low-cost introductory period converts, and before renewals
- A 14-day cooling-off period at sign-up and at each renewal
[CHECK: whether the final secondary legislation or guidance restricts save offers shown during the exit journey, and the exact reminder timings — the detail was still being consulted on.] Consumer protection enforcement by the Competition and Markets Authority (CMA) can also reach obstructive exits already, under the Act’s unfair-practices provisions.
Measuring a save honestly
Save rate — the share of cancellation attempts that end in a retained subscription — overstates what the flow achieves, because a lot of saves only delay the cancellation.
Worked example. 1,000 cancellation attempts in a month, £20 a month subscription.
attempts 1,000
accept an offer (pause etc.) 250 ← save rate 25%
of those, cancel within 90 days 100 ← 40% of saves were just deferred
still subscribed at 90 days 150 ← real save rate 15%
monthly revenue retained: 150 × £20 = £3,000
The number to report is still-subscribed-at-90-days, not accepted-an-offer. And the comparison should be against a holdout — people who got no offer — because some of the 150 would have changed their minds anyway — Holdout Groups.
In plain terms: “25% saved” is the flattering number. Some of those just postponed leaving, and some would have stayed without the offer. Only a comparison with people who saw no offer tells you what the flow actually kept.
Designing it
- Pause and skip first. For consumables, most “cancel” reasons are really “too much product” — Replenishment Timing
- Discounts last. A discount at cancellation teaches customers that trying to cancel earns a discount — Discounting Strategy
- Separate involuntary churn. Failed payments aren’t a decision and don’t belong in this flow — Involuntary Churn and Dunning
- Winback starts here. A clean, respectful exit is the precondition for getting them back later — Winback Campaigns
- Track it as its own funnel — attempts, reasons, offers shown, accepted, still active at 30/60/90 days — Subscription Metrics, Retention and Churn