MishaBook a demo

Aug 14, 2026

Stop Guessing on Churn

Churn rate is the percentage of customers acquired in a cohort who do not make a repeat purchase within the expected repurchase window for your category. For subscription: customers who cancel divided by active subscribers in the period. For repeat purchase (non-subscription): customers who don't buy again within 1.5x to 2x their historical average repurchase cycle.

Why Guessing Breaks Operations

Most DTC teams measure churn without defining the repurchase window first. This creates three problems: (1) churn appears to improve when it hasn't, (2) retention spend gets allocated to the wrong cohorts, (3) product decisions are made against noise instead of signal.

A brand selling skincare with a 45-day repurchase cycle cannot use a 30-day churn window. A brand with 90-day average repurchase cannot use 60-day windows. The window must match the business model, not arbitrary calendar periods.

Subscription brands have it easier - churn is explicit. Repeat-purchase brands must calculate their historical median or mean repurchase interval per customer segment, then set the window at 1.5x to 2x that number. Anything tighter creates false positives.

Define Your Repurchase Window

Start with historical data. Pull all customers acquired 12+ months ago. Calculate the time between their first and second purchase. Do this per acquisition cohort if AOV or product mix varies by channel.

  • Median repurchase interval (50th percentile) is more stable than mean for volatile cohorts
  • If median is 40 days, set churn window at 60 - 80 days (1.5x - 2x)
  • If median is 90 days, set churn window at 135 - 180 days
  • Recalculate quarterly. Seasonal products shift repurchase cycles
  • Use separate windows for different product categories if they have different repeat rates

Benchmark Against Cohort Performance

Churn is only meaningful when compared within cohorts. A customer acquired in January has a different repeat probability than one acquired in July. A customer acquired via paid search has different repeat behavior than one acquired via organic.

Build a cohort matrix: rows are acquisition months or channels, columns are time periods (30, 60, 90, 120+ days post-acquisition). Fill cells with the percentage of that cohort that has NOT repeated by that day.

This reveals patterns. If Q4 cohorts have 35% churn by day 90 and Q1 cohorts have 28%, you have a seasonal signal. If paid search cohorts have 40% churn and organic has 22%, you have a channel quality signal.

  • Healthy repeat-purchase DTC: 25% - 40% churn by day 90 (65% - 75% repeat rate)
  • Healthy subscription: 5% - 8% monthly churn (92% - 95% retention)
  • If churn is above 50% by your window, investigate product quality or fulfillment speed first
  • If churn varies >15 percentage points between cohorts, prioritize the worst-performing cohort

Common Failure Modes

Teams often measure churn in ways that hide the real problem. Recognizing these patterns prevents wasted retention spend.

  • Using calendar months instead of cohort windows: A customer acquired on Jan 28 looks like they churned by Feb 28 even if your window is 60 days. Always measure from acquisition date.
  • Mixing subscription and repeat-purchase cohorts: Subscription churn (cancellation) is not the same as repeat-purchase churn (no reorder). Track separately.
  • Ignoring fulfillment delays: If average shipping is 7 days, a customer acquired on day 1 doesn't have a fair repeat window until day 8+. Add fulfillment time to your window.
  • Treating all customers as equal: A customer who spent $500 on their first order has different repeat probability than one who spent $25. Segment by AOV.
  • Chasing churn without understanding causation: High churn often signals product-market fit issues, not retention messaging issues. Test product changes before spending on email campaigns.

Operationalizing Churn Measurement

Set up a repeatable process. This should run monthly with minimal manual work.

  • Define and document your repurchase window in writing. Include the calculation method and the date it was last reviewed.
  • Create a cohort retention table in your analytics tool (Shopify reports, GA4, or custom SQL). Rows = acquisition cohorts, columns = days post-acquisition at your window threshold.
  • Flag cohorts that exceed your churn threshold (e.g., >40% churn by day 90). Investigate the top 3 worst performers each month.
  • Separate churn into two buckets: (1) customers who have never returned, (2) customers who returned once but haven't returned again. These require different interventions.
  • Track churn by acquisition channel, product category, and AOV tier. This tells you where to focus retention budget.

When to Act on Churn Data

Not all churn is fixable. Knowing when to act separates operators from analysts.

  • If churn is 50%+ by your window and consistent across cohorts: Investigate product quality, fulfillment speed, or packaging. This is a product problem, not a retention problem.
  • If churn is 25% - 40% and stable: This is normal for most DTC. Focus on improving first-purchase AOV and repeat frequency, not reducing churn.
  • If churn spikes in a specific cohort (e.g., one paid channel or one month): Investigate that cohort's acquisition quality, product mix, or fulfillment timing.
  • If churn improves after a product change or fulfillment improvement: Measure the impact in the next cohort window. Don't declare victory until 2 - 3 cohorts confirm it.
  • If retention email campaigns reduce churn by <3 percentage points: The ROI is likely negative. Reallocate budget to product or acquisition quality.

The Operator's Checklist

Use this checklist to audit your current churn measurement.

  • [ ] Repurchase window is defined and documented based on historical median interval (not guessed)
  • [ ] Churn is measured per cohort (acquisition month, channel, or segment), not in aggregate
  • [ ] Fulfillment time is added to the window (customer can't repeat until product arrives)
  • [ ] Churn is separated from other metrics (don't conflate churn with LTV or repeat rate)
  • [ ] Thresholds are set per business model (subscription vs. repeat-purchase, not one rule for all)
  • [ ] Worst-performing cohorts are investigated monthly
  • [ ] Retention spend is tied to churn reduction impact, measured in the next cohort window

Questions

FAQ

What's the difference between churn rate and repeat rate?

Repeat rate is the percentage of customers who buy again (e.g., 70% repeat rate). Churn rate is the inverse (30% churn rate). They're the same metric, different framing. Use churn when tracking customer loss, repeat rate when tracking customer retention.

Should I measure churn from first purchase or from last purchase?

For initial churn (first repeat), measure from first purchase date. For ongoing churn (customers who've already repeated), measure from their last purchase date. Most teams focus on initial churn first because it's the biggest lever.

How often should I recalculate my repurchase window?

Quarterly at minimum. If you change product mix, pricing, or fulfillment speed, recalculate immediately. Seasonal businesses should recalculate before each season. If your window hasn't changed in a year, you're probably not measuring it right.

What churn rate should I target?

This depends on your category and business model. Most repeat-purchase DTC brands see 25% - 40% churn by day 90. Subscription brands typically see 5% - 8% monthly churn. Benchmarking against your own cohorts is more useful than industry averages because your product, price, and fulfillment are unique.

Want this on your account?

Thirty minutes. Bring the number that keeps you up.

Book a demo

More from the blog