<aside> 🟢
+10% Revenue Per Visitor in 90 days. Or we refund the engagement.
</aside>
A subscription is the only retention lever that turns a decision into a default. You have two jobs: win the first subscribe, then keep them subscribed against churn that is mostly fixable billing plumbing, not unhappy customers.
For a consumable brand, subscription is the strongest retention lever you own. It turns an active choice (remember to reorder, go to the site, check out) into a default (auto-ship, nothing to do). Every decision you remove from your customer's path is a customer you keep.
The model has a tax. Subscriptions leak every month, and most operators misread where the leak comes from. They assume it is people who stopped wanting the product. A large slice of it is expired credit cards. That distinction decides where you spend your time, because one is a marketing problem and the other is a backend fix you can mostly automate.
Before any play, separate your monthly subscription churn into voluntary and involuntary. You cannot fix what you have not split.
| Segment | Typical monthly churn | Why |
|---|---|---|
| Subscription ecommerce (blended) | ~3.4% | Voluntary + involuntary combined, across categories |
| Replenishment / consumables | under ~4% monthly | Real need drives the reorder, low discretion |
| Subscription boxes (curated) | ~10–12% monthly | Discretionary, novelty-driven, easy to cancel |
That ~3.4% blended figure runs across 2,200+ merchants and 67M subscribers. In lifetime terms it works out to roughly 40% annual churn, against 70%+ annual for one-time-purchase ecommerce. Consumables sit under ~4%, curated boxes at ~10–12%.
Two things matter more than the number. First, churn gets defined a dozen ways: logo churn vs revenue churn, voluntary vs involuntary, dunning recovery netted out or not. If your tool reports "3%" without telling you which kind, that is the first thing to fix. Second, a subscription box at ~10–12% is not a broken business, it is a discretionary category. A replenishment brand at the same rate has a real problem. Benchmark against your category, never the blended average.
Now split the two churn types, because they take completely different fixes:
<aside> ✅
Steal this: Pull last month's cancelled subscribers and tag each one as "card failed" or "chose to cancel." That single split tells you whether you have a billing problem or a product problem before you spend a dollar on either.
</aside>
Most brands bolt a "Subscribe & Save" radio button onto the product page, discount it 10%, and wonder why take-rate is low. The button is not the play. The moment is.
Pitch the subscription at the point of proven repeat intent, not on a cold first visit. Three moments where that intent is real: