Churn Rate

Churn rate is the percentage of customers (or subscriptions) that stop doing business with a brand over a defined period — the inverse mechanic of Retention and the variable the whole run-103→114 unit-economics cluster turns on: high churn collapses Customer Lifetime Value (CLV), lengthens the CAC Payback Period, and drags the LTV:CAC Ratio down regardless of how cheap acquisition is. This page records how sources define and calculate churn, what benchmarks they cite, and where they disagree.

Firewall: every claim is what a source reports. See ../../CONTEXT.md Rule 1. This is a web-only run — the Reddit MCP was not connected and the YouTube transcript actor (Apify) was unavailable, so there is no practitioner counter-narrative here; every quantified benchmark traces to vendors (Eightx, MobiLoud, Shopify, Recharge, Recurly) or tier-1 consultancies (McKinsey 2021), none independently audited.

How it is calculated

For subscription businesses churn is measurable at the cancellation moment. Shopify's formula (Shopify, 2022-10-21):

Churn = (Customers at start − Customers at end + New customers acquired) / Customers at start

Subtracting new customers prevents acquisition from masking existing-customer churn. Shopify's worked example: (100 − 105 + 10) / 100 = 5%.

For non-subscription stores there is no cancellation event, so Shopify says churn must be modelled via cohorts: group customers by first-purchase date, then measure the share who did not reorder over a window of 2× the average repeat-purchase timeline (Shopify, 2022-10-21). This is the bridge to Cohort Analysis.

Eightx flags two methodology traps (Eightx, 2026-05-29):

• Subscriber churn ≠ subscription churn. Customer-level churn differs from per-line-item churn, which is inflated when one customer holds multiple SKUs.
• Monthly churn compounds, it is not ×12. Annual churn = 1 − (1 − monthly)^12, so a 6%/month churn ≈ 52% annual, not 72%.

Benchmarks (as-of 2026-06-27)

All figures are volatile and vendor-aggregated; Eightx explicitly describes its bands as point-in-time that "shift quarterly." Treat as directional.

Non-subscription / cohort churn

• For single-purchase categories (skincare, a typical fashion brand) ~75% churn per first-purchase cohort is "average," implying ~25–26% repeat (Shopify, 2022-10-21 — stale-risk, 2022).
• Stat roundups cite the average store losing ~70–77% of customers yearly; top brands retain 45–55% (Envive/Rivo via search, low confidence).

Subscription monthly churn

• Headline "good" monthly churn 5–7%; <5% excellent; top-quartile <3%/mo (~31% annual compounded) (Eightx, 2026, med).
• Recurly research (via Shopify) puts ~5% monthly churn as average for subscription businesses (date not captured).
• By billing period (Finsi 2026 cohort via Eightx): monthly subscribers ~8.3%/mo, annual plans ~2.1%, annual prepay retaining ~2.5× the monthly rate over 12 months (28% vs 11%; weekly 3%; quarterly ~18%).
• Category beats cadence (Eightx): replenishment (supplements/coffee/pet) <4%/mo vs curated boxes (apparel/lifestyle) 10–15%/mo — a 3–4× gap.
• Only Peloton publishes a hard 10-K monthly churn figure — 1.6% on hardware-locked Connected Fitness vs 7.0% on its app; BARK Inc. lost 7.0% of active subscriptions YoY in FY2025 (Eightx citing SEC EDGAR, 2026-05-29, high — primary).

Voluntary vs involuntary churn

Sources split churn into two mechanically different failure modes (DigitalApplied + Eightx citing Recurly, 2026):

Type	Cause	Typical share of total churn
Voluntary	Customer actively cancels — value, experience or product mismatch	60–75%
Involuntary	Failed/expired card, missed renewal — no decision to leave	25–40% (up to ~50% low-AOV)

Involuntary churn is framed as an engineering-led fix (Dunning — automated payment retry/recovery) rather than a marketing one; Recurly publishes a 49% baseline dunning recovery rate rising to 71% for optimised merchants (Eightx, 2026-06-05). Recharge's DTC panel reports a ~4.1% voluntary + 3.0% involuntary split (via search, low confidence).

Why it dominates the unit economics

• Eightx frames retention improvements as moving valuation 3–5× more than CAC reductions because "the core battle is the second purchase" (as-of 2026).
• Operator math (Eightx): a brand at 6%/mo churn loses ~70% of subscribers in a year; pulling churn to 4% leaves ~56% retained — "an extra 14 points of base retention without acquiring a single new customer." Every monthly churn point saved ≈ 3–5 months of paid-acquisition effort (as-of 2026-06-05).
• This links directly to the run-103→114 cluster: see Customer Lifetime Value (CLV) (churn 18→14% ≈ "cutting CAC in half"), CAC Payback Period, and Subscription Commerce.

Key terms

Term	Meaning
Churn rate	% of customers/subscriptions lost over a period
Voluntary churn	Customer actively cancels
Involuntary churn	Subscription lapses via failed/expired payment
Dunning	Automated retry/recovery of failed payments
Subscriber vs subscription churn	Customer-level vs per-line-item — the latter inflated by multi-SKU holders
Compounded churn	Annual = 1 − (1 − monthly)^12, never monthly × 12
Deferred renewal cliff	Annual plans hold ~94% to month 11, then 25–35% churn at renewal

Open questions / frontier

• Retention — the inverse, its own page now alongside this one.
• Cohort Analysis — the measurement engine non-subscription churn depends on, still dangling.
• Dunning / Involuntary Churn — the payment-recovery mechanic, dangling.
• A UK/Europe-specific churn benchmark — all hard data is US-centric (relevant to UNIQLO Europe).
• A practitioner counter-narrative — both Reddit and YouTube streams were down this run.