RenewalIQ Churn pattern review Churn Pattern Review · USD 59
Your churn is 4.1 percent a month and almost all of it happens in month two
Eighteen months of subscriptions, broken down by when people leave rather than how many.
You report churn as one number and it is hiding the shape. Thirty eight percent of every cohort is gone by the end of month two, and the survivors after month three churn at under one percent a month. You do not have a retention problem. You have an onboarding problem and a very good product.
Cancellations concentrate at the end of the first billing cycle and again after the fourth delivery. Beyond month three the curve is almost flat, which means the customers who get past week twelve stay for years.
Shape
When people actually leave
The curve falls 45 points in three months and 8 points in the following nine. A blended monthly figure of 4.1 percent describes neither half of that and is the reason this has looked like a slow leak rather than a cliff.
Cohorts
Has it changed over time?
Each row is everyone who signed up in that month, tracked forward. Reading down a column shows whether newer cohorts behave differently from older ones.
| M1 | M2 | M3 | M4 | M5 | M6 | ||
|---|---|---|---|---|---|---|---|
| Apr 2025 | 410 | 82% | 64% | 57% | 54% | 53% | 52% |
| May 2025 | 455 | 80% | 63% | 56% | 53% | 52% | 51% |
| Jun 2025 | 502 | 81% | 62% | 55% | 52% | 51% | 50% |
| Jul 2025 | 610 | 76% | 57% | 50% | 47% | 46% | 45% |
| Aug 2025 | 780 | 71% | 51% | 44% | 41% | 40% | 39% |
| Sep 2025 | 845 | 70% | 50% | 43% | 40% | 39% | 38% |
| Oct 2025 | 690 | 77% | 59% | 52% | 49% | 48% | 47% |
| Nov 2025 | 640 | 79% | 61% | 54% | 51% | 50% | |
| Dec 2025 | 720 | 78% | 60% | 53% | 50% | ||
| Jan 2026 | 880 | 80% | 62% | 55% |
July to September 2025 are visibly worse than every cohort either side, by around 10 points at month two. Those three months are also when you ran the half price introductory offer, and section 03 is about what that means.
Cause
The discount cohorts
| Full price cohorts | Half price cohorts | Difference | |
|---|---|---|---|
| Subscribers acquired | 3,307 | 2,235 | |
| Retained at month 2 | 61% | 52% | 9 points |
| Retained at month 6 | 49% | 39% | 10 points |
| Average deliveries before cancelling | 5.8 | 3.1 | 2.7 fewer |
| Cancelled citing price | 18% | 41% | 23 points |
| Twelve month value per subscriber | USD 312 | USD 168 | USD 144 lower |
The half price offer acquired 2,235 subscribers at a twelve month value USD 144 lower each. That is roughly USD 322,000 of value difference against an acquisition cost saving that was much smaller. It worked as an acquisition campaign and not as a subscriber campaign.
Reasons
What people say on the way out
- Test 1
Offer a smaller box at cancellation, not a discount
Twenty two percent say too much food. You currently offer 20 percent off, which answers a different objection and trains people to wait for a discount.
- Test 2
Offer a delivery day change at cancellation
Eleven percent. One click, no revenue given away.
- Test 3
Intervene at delivery four, not at cancellation
The month two cliff happens after the fourth box. Whatever you do, do it at box three.
- Do not
Repeat the half price offer as it was
Until you can run it as a clean test against a control, you cannot tell how much of the ten point gap it owns.