When Churn Is the Wrong Metric

Why Churn Fails as a Diagnostic Tool

Churn is a summary statistic. A 5% monthly churn rate tells an operator that 5 out of 100 customers did not buy again - but not why. It does not distinguish between:

- Customers who received a defective product and stopped engaging

- Customers who bought once, got what they needed, and had no reason to return

- Customers who intended to repurchase but forgot the brand existed

- Customers who switched to a competitor after a poor support interaction

Each failure mode requires a different intervention. Churn aggregates them into noise.

Second, churn is lagging. If a product quality issue ships in week one of a month, the cohort's churn rate does not fully surface until month two or three. By then, 500 more customers have experienced the same problem. A leading indicator - like support ticket volume, return rate, or email unsubscribe velocity - would have flagged the issue in days.

Leading Indicators to Replace Churn

Replace churn with metrics that move first and point to root cause. These are observable within 7 - 14 days, not 30 - 90.

For subscription and repeat-purchase DTC brands, track:

• Repeat purchase rate (RPR) by cohort: % of customers who buy a second time within 60 days. Threshold: > 15% for healthy CPG / beauty; > 25% for high-AOV repeat categories
• Email engagement decay: % of customers who unsubscribe or stop opening emails within 30 days of first purchase. Threshold: < 8% unsubscribe rate in first 30 days
• Support ticket volume per 100 customers: Spike in tickets (especially returns, damage, quality complaints) predicts churn 2 - 3 weeks later. Threshold: < 3 tickets per 100 customers
• Days to second purchase (D2P): Median days between first and second order. Increasing D2P signals weakening intent. Threshold: Track trend; flag if it increases > 10% month-over-month
• Product return rate: % of orders returned within 30 days. Threshold: < 5% for apparel; < 2% for food / supplements
• NPS or CSAT at purchase: Measure satisfaction immediately after first order, not months later. Threshold: > 7 / 10 CSAT

Churn Thresholds by Business Model

Acceptable churn varies by model. Use these benchmarks as a starting point, then adjust for your category and customer acquisition cost.

• Subscription (auto-renew): 3 - 5% monthly churn is healthy; > 7% signals a problem
• Repeat-purchase (non-subscription): 8 - 12% monthly churn is healthy; > 15% signals a problem
• One-time purchase (no repeat expectation): Churn is irrelevant; measure customer satisfaction and referral rate instead
• High-AOV (> $500): 2 - 4% monthly churn; lower tolerance because each customer is high-value
• Low-AOV (< $50): 10 - 20% monthly churn; higher tolerance because acquisition cost is lower

Common Churn Mistakes

Mistake 1: Measuring churn over too short a window (e.g., 14 days). This inflates churn and creates false alarms.

Mistake 2: Aggregating churn across all channels and products. Always segment by acquisition source, product line, and customer segment. A single churn number hides problems.

Mistake 3: Ignoring cohort effects. Churn from January customers may differ from churn from June customers due to seasonality, product changes, or ops improvements. Always compare like-for-like cohorts.

Mistake 4: Treating churn as the primary retention metric. Churn is a symptom, not a diagnosis. Pair it with leading indicators to understand what is actually happening.

Want this on your account?

Thirty minutes. Bring the number that keeps you up.