Churn for Multi-Channel DTC: Definitions, Thresholds, and Failure Modes

Why Channel-Specific Churn Matters

A blended churn rate across all channels obscures which acquisition sources are sustainable. A DTC brand acquiring 40% of customers via TikTok and 60% via email list will see different repeat purchase behavior in each cohort - TikTok cold audiences often churn 60-75% by day 90, while email list customers churn 20-35%. Averaging these produces a meaningless 40% figure that doesn't guide action.

Multi-channel DTC operators need separate churn tracking for: Shopify direct (organic + paid search), paid social (TikTok, Instagram, Facebook by campaign), email list (segmented by source: organic, lead magnet, purchased), SMS (if used as acquisition), and affiliate / partnership channels. Each has different customer quality profiles and repeat purchase economics.

Defining Cohort Windows and Thresholds

Churn is always cohort-based: customers grouped by acquisition date (weekly or monthly) and measured for repeat purchase activity within a fixed window. The window choice depends on product category and margin structure.

90-day churn: Standard for fast-moving consumer goods (FMCG), supplements, beauty. Captures early repeat purchase intent. Threshold: 50-65% acceptable for paid social, 25-40% for email / organic.
180-day churn: Used for higher-ticket items (apparel, home goods, electronics under $200). Threshold: 60-75% acceptable for paid social, 30-50% for email.
365-day churn: Subscription or seasonal repeat (e.g., annual skincare refresh). Threshold: 70-85% acceptable for paid social, 40-60% for email.
Day-7 and day-30 micro-churn: Early warning signals. If day-7 churn exceeds 85%, product quality or fulfillment issues are likely. If day-30 churn exceeds 75%, onboarding or email sequence is weak.

Calculation and Tracking Setup

Churn calculation is straightforward but requires clean data hygiene. For a cohort acquired in week 1 of January, measure the percentage who placed zero orders in the 90 days following their first purchase date.

Numerator: Customers in cohort with zero repeat purchases in window.
Denominator: Total customers in cohort (exclude test orders, refunded-only customers, and known fraudulent transactions).
Formula: (Customers with 0 repeats / Total cohort size) × 100 = Churn %.
Track in a simple sheet: Cohort week, channel, cohort size, repeat purchasers by day 7 / 30 / 90 / 180 / 365, churn % at each milestone.
Lag: Churn data is only final after the window closes. Day-90 churn for a January cohort is confirmed in early April. Use rolling 7-day or 14-day micro-churn as a leading indicator.

Channel-Specific Failure Modes and Diagnostics

High churn in one channel often signals a specific operational failure. Operators should follow a decision tree to isolate root cause before cutting spend or changing strategy.

Paid social churn > 75% by day 90: Audience quality issue (wrong targeting, bot traffic) or product-market fit gap for that audience. Diagnostic: Check day-7 churn. If > 90%, audience is wrong. If 60-75%, product is right but onboarding email sequence is weak.
Email list churn > 50% by day 90: Weak welcome sequence or list quality degradation. Diagnostic: Segment by list source. If purchased list churn is 70%+ and organic list churn is 35%, stop buying lists. If both are high, redesign day-1 to day-7 email sequence.
Organic / direct churn > 40% by day 90: Product quality or fulfillment issue affecting all channels equally. Diagnostic: Check refund rate and customer support tickets. If refund rate > 15%, product is the problem. If < 8%, onboarding is weak.
SMS acquisition churn > 80% by day 90: SMS list was not opted-in properly or audience expectations were misaligned. Diagnostic: Check SMS consent source and message content. If SMS was appended to email list without explicit opt-in, expect high churn.
Affiliate churn > 80% by day 90: Affiliate is driving low-intent traffic (incentive-seeking, not product-seeking). Diagnostic: Review affiliate terms. If affiliate is offering 20%+ discount stacking, audience quality is poor. Renegotiate or pause.

Cohort Decay and Natural Churn Acceptance

Not all churn is actionable. Some decay is natural and expected. The key is distinguishing between acceptable cohort decay and a signal of operational change.

A healthy cohort typically shows: 70-80% churn by day 90 for paid social (cold audiences), 30-40% for email (warm audiences), 15-25% for repeat-purchase-focused segments (loyalty program members, subscription cohorts). If a channel's churn is within historical range, no action is needed. If churn increases 10-15 percentage points month-over-month in the same channel, investigate.

Decision Rules for Churn Action

Operators need a clear decision tree to avoid over-reacting to normal variance or under-reacting to real problems.

If churn is within historical range (±5pp): No action. Monitor weekly.
If churn increases 5-10pp vs. prior 4-week average: Investigate root cause (product, email, audience). Do not cut spend immediately.
If churn increases > 10pp vs. prior 4-week average: Pause new spend in that channel within 48 hours. Investigate root cause. Resume only after fix is confirmed in a 2-week test cohort.
If day-7 churn exceeds 90%: Pause immediately. This is a fulfillment, product quality, or audience mismatch issue.
If email churn increases but paid social churn is stable: Problem is email sequence, not product. A/B test welcome sequence.
If all channels' churn increases simultaneously: Problem is product quality, fulfillment, or brand perception. Audit refunds, support tickets, and product reviews.

Reporting and Cadence

Churn should be reviewed weekly at the operator level and monthly in cross-functional meetings. Weekly reviews catch emerging issues; monthly reviews confirm trends and guide budget allocation.

Weekly operator review: Check day-7 and day-30 micro-churn for all active cohorts. Flag any channel with churn > 10pp above historical average. Investigate root cause same day.

Monthly stakeholder review: Present 90-day churn by channel for all cohorts that have completed their window. Compare to prior month. Discuss budget reallocation based on churn performance. Update channel thresholds if product or audience mix has changed.

Questions

FAQ

Should I measure churn by product category or just by channel?

Start with channel. If a channel acquires customers across multiple product categories, measure channel churn first. If churn is high, then segment by product to isolate whether the problem is audience quality (channel) or product quality (category). Most DTC operators find channel is the primary driver of churn variance.

What if my repeat purchase window is irregular (e.g., seasonal)?

Use a 365-day window and measure repeat purchase rate instead of churn. For seasonal products, a customer who buys in January and repeats in December is a keeper, even if they churn by day 90. Adjust your threshold accordingly - 365-day churn of 70-80% is acceptable for seasonal categories.

How do I account for customers who refund their first order?

Exclude them from the cohort denominator. A refunded customer is not a customer - they are a transaction reversal. Include only customers whose first order was net-positive (not fully refunded). If refund rate is > 10%, investigate product quality before analyzing churn.

Can I use email engagement (opens, clicks) as a proxy for churn?

No. Email engagement is a leading indicator of repeat purchase intent, not a replacement for churn. A customer who opens emails but never buys is still churned. Use email engagement to diagnose why churn is high (weak sequence), but measure churn only by repeat purchase behavior.

Want this on your account?

Thirty minutes. Bring the number that keeps you up.

Book a demo