Weekly Churn Review Template

Q: What's the difference between churn and refund rate?

Refund is a transaction reversal (customer gets money back). Churn is customer loss (no future purchase or subscription cancellation). A customer can refund and stay (no churn) or refund and leave (churn). Track both separately. High refunds often precede high churn by 1 - 2 weeks.

Q: Should churn thresholds differ by business model?

Yes. Subscription brands typically see 2 - 5% monthly churn (0.5 - 1.2% weekly). Repeat purchase (non-subscription) should target <15% quarterly churn. High - ticket B2B can tolerate 1 - 2% annual churn. Set thresholds based on your cohort retention curve and unit economics, not industry averages.

Q: How do I separate signal from noise in weekly churn?

Use a 4 - week rolling average as baseline. Flag weeks that deviate >1.5x from that average. Single - week spikes are often noise (payment processor glitch, one customer bulk cancellation). Sustained elevation (2+ weeks) is signal. Always segment by cohort - noise in one cohort may be signal in another.

Churn Definitions and Measurement

Start with explicit definitions. Churn is the percentage of active customers who cancel or lapse in a given period. For subscription brands, this is straightforward. For repeat purchase (non-subscription), define churn as: no purchase in 90 days after last order, or explicit unsubscribe.

Measure three metrics weekly: (1) Gross churn rate - percentage of customers lost regardless of new acquisition. (2) Net churn - gross churn minus reactivations. (3) Cohort retention - percentage of customers from a specific cohort (e.g., acquired in week 1) still active at week 4, 8, 12.

Formula: Gross churn rate = (Customers lost in week / Active customers at week start) × 100. Threshold: Healthy DTC is 3 - 7% weekly gross churn. Above 10% signals acute problem. Below 2% suggests measurement error or very early cohort.

Weekly Review Checklist

Conduct review on a fixed day (e.g., Monday morning). Assign one operator ownership. Time box to 30 minutes for data pull and 30 minutes for decision.

Pull churn rate for current week and prior 4 weeks. Flag if current week > 1.5x average of prior 4 weeks.
Segment churn by cohort (acquisition week). Identify if churn spike is concentrated in one cohort or distributed.
Cross-reference with cancellation reasons (if captured). Categorize: price, product quality, delivery, competitive, other.
Check for operational failures: payment processor issues, email deliverability problems, customer service backlog.
Review refund rate and return rate. Churn often precedes or follows product quality issues.
Compare churn to prior year same week. Seasonal patterns matter.
Document decision and action owner in shared log.

Threshold - Based Decision Rules

Establish decision rules tied to specific thresholds. This removes guesswork and ensures consistent response.

If weekly churn > 10%: Immediate investigation required. Check payment processor status, email deliverability, and customer service queue. Schedule emergency call with product and marketing leads within 24 hours.

If weekly churn 7 - 10% and rising for 2+ consecutive weeks: Escalate to leadership. Likely signals product issue, market shift, or competitive pressure. Initiate root cause analysis.

If weekly churn 3 - 7% (normal range): Standard review. Document trends. No action unless cohort - specific pattern emerges.

If cohort retention drops below 40% at week 4: Flag acquisition channel or product fit issue. Review messaging and onboarding for that cohort.

If refund rate > 15% of orders: Churn will follow. Prioritize product quality audit before churn accelerates.

Common Failure Modes

Measurement drift. Definition of 'active' changes mid - year. Subscription vs. repeat purchase customers mixed in one metric. Solution: lock definitions in Q1, audit monthly.

Lag in data. Churn reported 5 days late means response is reactive, not preventive. Ensure churn data is available within 24 hours of week close.

Ignoring cohort patterns. Overall churn looks stable but one acquisition channel has 15% churn. Averaging masks the problem. Always segment.

Confusing churn with refund. A customer who refunds and leaves is different from a customer who stays but churns later. Track separately.

No action threshold. Churn reviewed but no decision rule tied to it. Result: meetings without outcomes. Use the thresholds above or define your own, but enforce them.

Seasonal blindness. Q4 churn spikes due to holiday returns, not product failure. Compare year - over - year, not week - over - week in isolation.

Cohort Retention Curves

Plot retention by cohort weekly. X - axis is weeks since acquisition. Y - axis is percentage retained. This reveals product fit and onboarding quality.

Healthy curve: 90% retained at week 1, 70% at week 4, 50% at week 12. Steep drop at week 1 - 2 suggests poor onboarding or unmet expectation. Cliff at week 4 suggests trial period ending or seasonal purchase cycle.

Compare curves across acquisition channels. If paid social cohort has 40% week 4 retention and organic has 65%, paid social messaging or audience is misaligned.

Update curves weekly. Trends emerge faster than single - week snapshots. A curve flattening at week 8 (instead of declining) signals improved retention.

Root Cause Investigation Protocol

When churn spikes, follow this sequence: (1) Check operational health - payment processor, email, SMS delivery, customer service response time. (2) Review product changes - recent updates, price increase, shipping policy change. (3) Analyze cancellation feedback - pull 20 - 30 recent cancellation surveys or support tickets. (4) Segment by channel - organic vs. paid, new vs. returning, geographic. (5) Compare to competitor activity - pricing, promotions, new product launches.

Document findings in a shared template. Include: spike date, magnitude, suspected cause, confidence level (high / medium / low), and recommended action.

Weekly Log Template

Maintain a running log. Minimal structure, maximum clarity.

Questions

FAQ

How often should churn be reviewed?

Weekly minimum for brands with >500 active customers. Daily for brands with <500 (volatility is higher). Monthly is too slow - by then, revenue is already lost. Weekly cadence allows response within 7 days of spike.

What's the difference between churn and refund rate?

Refund is a transaction reversal (customer gets money back). Churn is customer loss (no future purchase or subscription cancellation). A customer can refund and stay (no churn) or refund and leave (churn). Track both separately. High refunds often precede high churn by 1 - 2 weeks.

Should churn thresholds differ by business model?

Yes. Subscription brands typically see 2 - 5% monthly churn (0.5 - 1.2% weekly). Repeat purchase (non-subscription) should target <15% quarterly churn. High - ticket B2B can tolerate 1 - 2% annual churn. Set thresholds based on your cohort retention curve and unit economics, not industry averages.

How do I separate signal from noise in weekly churn?

Use a 4 - week rolling average as baseline. Flag weeks that deviate >1.5x from that average. Single - week spikes are often noise (payment processor glitch, one customer bulk cancellation). Sustained elevation (2+ weeks) is signal. Always segment by cohort - noise in one cohort may be signal in another.

Want this on your account?

Thirty minutes. Bring the number that keeps you up.

Book a demo