Subscription E-commerce

The Subscription Models

Replenishment — the customer receives the same consumable on a schedule. Coffee, razor blades, pet food, supplements. Highest retention because the need genuinely recurs.

Curation — a changing selection chosen for the customer. Beauty boxes, wine clubs, snack boxes. Driven by discovery and novelty, which decays; retention is structurally harder.

Access — a membership giving pricing, delivery or service benefits rather than products. Retention depends on whether the customer keeps using the benefit.

Retention profiles differ enormously between these. Applying curation-box benchmarks to a replenishment business, or the reverse, produces badly wrong forecasts.

The Economics

The core relationship: customer lifetime value must comfortably exceed acquisition cost, and the payback period must be short enough that growth does not exhaust cash.

LTV is driven by average order value, margin, and expected lifetime — which is a function of churn. A 5% monthly churn rate implies an average lifetime around twenty months. At 10% it halves to ten. Small churn improvements have outsized effects on LTV because the relationship is not linear.

Cash flow is the constraint that kills otherwise viable subscription businesses. Acquisition is paid upfront; revenue arrives over months. Growing fast makes this worse, not better.

Reducing Churn

Churn splits into voluntary — the customer chose to leave — and involuntary, where payment failed. Involuntary churn is frequently 20 to 40 percent of total churn and is the cheapest to fix.

Involuntary: card retry logic with sensible schedules, account updater services that refresh expired card details automatically, pre-expiry notifications, and multiple payment methods on file.

Voluntary: the first delivery experience disproportionately predicts lifetime, so over-invest there. Offer pause rather than cancel — a paused subscriber returns far more often than a cancelled one. Allow schedule and quantity changes, since the most common genuine reason for cancelling replenishment is receiving too much.

Onboarding and the First 90 Days

Most churn happens early. The first three deliveries determine whether the subscription becomes habitual.

What works: setting clear expectations on timing and contents before the first delivery arrives, a genuinely good unboxing, proactive contact after the first delivery rather than waiting for complaints, and making the second delivery feel like a decision the customer is glad they made.

Acquisition for Subscriptions

Free trials and heavy first-box discounts acquire volume and attract customers who churn immediately. The cohorts acquired this way frequently have negative unit economics.

Measure acquisition channels on retained cohort value at month six, not on signups. Channels rank very differently on that basis, and the cheapest signups are often the most expensive customers.

The Cohort Question That Decides Everything

Subscription businesses are judged on monthly revenue and determined by cohort retention, and the two can move in opposite directions for a long time.

Track retention by joining cohort, not in aggregate. Aggregate retention is flattered by growth: new subscribers dilute the churn of older ones, so a business with worsening retention can show improving headline numbers while it is acquiring.

What a cohort curve tells you that nothing else does. Where the drop is. A cliff after the first delivery is a product or expectation problem; a cliff at the end of an introductory offer is a pricing problem; a slow steady decline is a value problem. These need completely different responses and look identical in a monthly churn figure.

Whether it flattens. A curve that flattens has a durable base and the business works. One that keeps declining has no floor, and growth is only outrunning the leak.

And whether newer cohorts are better. That is the only real evidence that a retention initiative worked.

Reducing Churn Where It Actually Happens

Most churn work targets the cancellation moment, which is the last and least effective place to intervene.

The first delivery matters more than everything after it. If it arrives late, damaged, or unlike what was pictured, the subscription is already lost and no offer recovers it. Spend here first.

Give control rather than resisting cancellation. Skip a delivery, change the frequency, swap the contents, pause for a month. A pause is a retained customer; a friction wall is a complaint and a chargeback — and in several jurisdictions a cancellation flow harder than the signup flow is a regulatory problem, not just a bad experience.

Ask at cancellation and act on the answer. The reasons cluster — too much product, too expensive, no longer needed — and two of those three have a product response rather than a discount response.

And watch the payment failures. A meaningful share of churn is an expired card rather than a decision. Dunning — retry schedules, pre-expiry prompts, updater services — recovers subscribers who never intended to leave, and it is the cheapest retention work available.