Module 4: Fighting Churn
The Churn Audit: a systematic review that finds exactly why customers leave. Save plays, win-back campaigns, and dunning — the defense that protects everything acquisition built.
The Churn Audit
The Churn Audit is a systematic review of why customers leave, ranked by impact — so you fix the biggest leak first instead of guessing. Most companies 'know' why customers churn the way people 'know' why diets fail: a confident story with no data. The audit replaces the story with evidence, and it takes one focused afternoon.
Step one: pull your last 30–50 cancellations and categorize every reason into a taxonomy — price, missing feature, never activated, support experience, business closed, failed payment, switched competitor, unknown. Step two: plot churn by cohort — which signup month, which plan, which acquisition channel churns fastest? Step three: rank. If 40% never activated, that's an onboarding problem wearing a churn costume. If one plan churns triple the others, that's a packaging problem. The pattern tells you which module to reopen: activation, product, pricing, or payments.
Step four: talk to ten churned customers. One question: 'What would have kept you?' Record answers verbatim — the exact words become your save-play copy later. Run the audit quarterly; churn reasons drift as the product and market change, and last year's audit is this year's outdated story.
A typical churn split: where the leavers actually go
Build brief
List the top five reasons you believe customers cancel. Rank them 1–5. Then write how you'd verify each one with real data.
Saved
Save plays
A save play is what a canceling customer sees before the cancellation completes. The worst version is a guilt trip ('We'll miss you!') followed by the door. The best version is a genuine off-ramp menu: pause the subscription for two months, downgrade to a cheaper tier, talk to a human, or cancel. A surprising share of 'cancellations' are really 'I need a break' or 'this tier is too much right now' — give them that option and they stay on the books instead of leaving.
Match the play to the reason. Price objection → offer the annual discount or a lighter tier, not a begging email. Missing feature → be honest about the roadmap and offer a pause until it ships. Never activated → this one you can't save at cancellation; it's an onboarding failure from month one, and the audit should have sent you back to Module 3. The save flow's real job is triage: keep the saveable, learn from the lost.
One hard rule: never train customers to threaten cancellation for discounts. If every save offer is 30% off, you've built a discount machine, not a retention system — and word spreads. Reserve discounts for genuine price pain, and make pauses and downgrades the default plays. A paused customer at $0 still counts as retained revenue potential; a discounted customer teaches the market your price is negotiable.
Build brief
Design your cancellation save flow: what does a canceling customer see, in what order — and what offer do they get for each reason?
Saved
Win-back and dunning
Churned customers are warm leads wearing a different label — they know the product, they once paid, and something specific pushed them out. A win-back sequence runs 30, 60, and 90 days after cancellation: email one acknowledges the reason they left ('you told us reporting was thin'), email two shows what changed since ('reporting 2.0 shipped last month'), email three makes a low-friction offer (a month free, a guided restart). Win-back converts at 5–15% in healthy products — found money.
Then the silent killer: dunning, the recovery of failed payments. Expired cards, insufficient funds, bank declines — involuntary churn — quietly accounts for a fifth to two-fifths of churn in card-billed SaaS. These customers didn't choose to leave; their payment just failed. The fix is unglamorous and extremely profitable: automatic retries on a smart schedule, a plain-text 'update your card' email from a real name (not a no-reply invoice), and in-app banners before the account locks. Every failed payment you recover is a customer you kept without a single feature.
Measure dunning recovery rate: recovered failed payments ÷ total failed payments. Under 50% and you're leaving money on the table — most of it fixable with better email copy and one more retry. This is the highest-ROI work in churn defense because it requires no product changes at all.
Build brief
Outline a 3-email win-back sequence for customers who left 90 days ago: one line per email — the hook and the offer.
Saved
Priya's monthly churn is 8%. Her exit data shows 40% of cancellations are failed credit card payments — not unhappy customers. What's the highest-leverage fix?
Your turn: Run a mini churn audit: list your last 20 cancellations (or your best estimates), categorize each reason, and write the #1 pattern in one sentence.
Saved.
Your next 24 hours
Email five churned customers one question: 'What would have kept you?' No pitch, no discount attached — just the question. Record their answers word for word.
Steal this template
THE CHURN AUDIT — one-page worksheet
☐ Pull last 30–50 cancellations
☐ Categorize: price / missing feature / never activated / support / closed / failed payment / competitor / unknown
☐ Plot by cohort: which month, plan, channel churns fastest?
☐ Rank: the #1 reason gets fixed first
☐ Interview 10 churned customers: "What would have kept you?"
☐ Re-run quarterly — reasons drift
Cheat sheet
- The Churn Audit: categorize 30–50 cancellations, plot by cohort, rank by impact, interview ten leavers.
- 'Never activated' churn is an onboarding problem wearing a churn costume — send it back to Module 3.
- Save plays: pause, downgrade, or talk to a human — before cancel. Match the play to the reason.
- Never discount every cancellation — you'll train the market that your price is negotiable.
- Involuntary churn (failed payments) is a fifth to two-fifths of card-billed churn. Dunning recovers it.
- Win-back at 30/60/90 days converts 5–15% — churned customers are warm leads.