Retention and churn

Past the first customers, keeping revenue matters more than winning it. Measure gross and net churn, ask every leaver why, recover failed payments, and offer a pause or a fitting save before a blanket discount.

13 steps, about 40 hours of work · 3 plays · 1 checked claims

The concept, explained in full: Retention marketing.

Do it

The steps from the launch plan that belong to this topic, in plan order.

Automate the lifecycle

Automate the trial-expiry sequence4h
Day −3, day −1, expiry day, day +3. The last one converts more than you expect.
Build proper dunning4h
Card retries, an updater for expired cards and a short email sequence. Stripe’s own billing page says it recovered $8.2B in failed payments in 2025; read any recovery figure as the vendor’s own.
Build the win-back sequence4h
Sent 30 and 90 days after cancellation, and only when you have actually shipped their reason.

Fix the leaky bucket first

Measure gross and net revenue churn monthly1h
Both numbers. Net alone hides a retention problem behind expansion revenue.
Put an exit survey on every cancellation4h
Five fixed options plus free text. The free text is where the actual reason lives.
Group the exit reasons into five buckets1h
Price, missing feature, not using it, switched, other. Five is enough to act on.
Write one response per bucket: discount, pause, help, or let go2h
Match the offer to the reason. Letting go is a legitimate answer for the wrong fit.
Build the offer screen shown after the reason is chosen8h
One screen, one offer, no maze. The cancel button stays visible.
Track saves, and the 90-day retention of saved accounts2h
A save that cancels in a month was not a save. Count only those still paying at 90 days.
Set the pause length and what stays accessible1h
Pick a fixed length and decide whether data stays readable, then say both on the screen.
Build the pause option into the cancel flow6h
Show it beside cancel, not instead of it. Billing resumes on a date the customer sees.
Email at the end of the pause with the work they left behind2h
Their saved work is the reason to come back. Put it in the subject line.
Get net revenue retention above 100%1h
Above 100%, existing customers grow revenue on their own. That is what makes $1M reachable.

Plays

Reusable patterns that explain how: backed by read sources first, then practitioner advice, then our own ideas.

Reply in person to every cancellation reason within a dayBacked · 1 source
Make the billing portal ask why a customer is cancelling, then have the founder reply personally to each answer within a day: a pause or smaller plan for price, an honest date for a missing feature, a question for a lost comparison. Some come back, and the reasons, counted monthly, tell you what to fix next.
The month-two email: what to do when the quick wins are doneOur idea
Around week five, email customers the habit or feature that keeps paying off once setup is done, with one concrete example of it working. Usage quietly drops in the second month when the first wins are behind people; an email that hands them the next level of value gives them a reason to keep the tab open.
A quarterly "here is what you got out of it" recap per customerOur idea
Once a quarter, send each customer a short recap in their own data: what the product did for them, how that compares with last quarter, and one thing they have not tried yet. People forget value they do not see, and a recap in their own numbers is the renewal argument made months early, without a sales call.

What the evidence says

Evidence, weak

Stripe’s billing page says $8.2B in failed payments were recovered in 2025 and $14 recovered for every $1 spent on its tools, and lists Smart Retries, expired card updates, recovery and retention automations and cancellation surveys. It states no share of involuntary churn recovered. Two earlier readings of a 55% average recovery rate could not be found on the page.

A vendor’s own landing-page totals with no stated method, and a figure two readers reported was not on the page when it was checked directly.

Re-checked 2026-11-10 by Claude, reading the page directly: not confirmed. The page does not state a 55% recovery rate (the only 55% is a sample tax rate). It says $8.2B in failed payments were recovered in 2025 and $14 for every $1 spent. The claim was rewritten to that.

Changed in the plan: Plan text corrected 2026-11-10, twice: the first version had 30-50% of involuntary churn with no primary source; the second had a 55% that the page does not carry. It now quotes the totals the page does state.

Stripe Billing (read in full)

True as of 2026-11-10; to be re-checked by 2027-02-10.

Sources

Stripe Billing

Where it sits in the plan

6.1Fix the leaky bucket firstin Scale to $1M ARR
5.3.3Build proper dunningin Compounding engine
5.3.4Build the win-back sequencein Compounding engine
5.3.2Automate the trial-expiry sequencein Compounding engine