Subscription box churn rate by renewal cohort

Subscription box churn rate should count cancellations separately from skipped shipments and pauses. Start with a fixed renewal cohort, classify each subscription's outcome, and then measure retained shipment revenue on those same subscriptions. That keeps a missed box from becoming a false customer loss.

Set the cohort before counting losses

Choose one shipment cadence and one observation window. The example below follows monthly subscriptions scheduled for a September 2026 box, split by the month each subscription began. Every subscription was open at the start and had a September renewal opportunity.

Use the subscription ID as the unit. A customer with several subscriptions can cancel one and retain another; that is subscription churn, not necessarily customer churn. The general churn definition explains the denominator distinction. Here the operational question is which box relationships ended.

Export subscription ID, start cohort, cadence, scheduled renewal date, effective cancellation date, pause or skip status, and order ID. Bring a separate order export with subscription ID, shipment period, net product revenue, credits and refunds. Aggregate orders to one observation per subscription before joining them. Multiple order lines must not multiply the subscription count.

Freeze cohort membership before inspecting outcomes. Keep cancellations inside that cohort even if the current subscription export hides closed records. New September subscriptions belong in a separate acquisition cohort. If your export contains only current active subscriptions, obtain the opening snapshot and cancellation history before calculating anything.

Classify one outcome per subscription

For this example, use four mutually exclusive outcomes at the cutoff: Renewed, Skipped, Paused and Canceled. A subscription that canceled after shipping still belongs in Canceled; its shipment revenue can remain positive. The example's canceled subscriptions happened to ship nothing.

Choose a rule for a cancellation that is requested now but effective later. Use the effective end of the relationship consistently, rather than mixing request dates with service end dates. A resumed subscription is retained at the cutoff under this measure; record its intervening pause separately if you need an activity history.

Renewal, retention and churn can use different eligible populations. Keep that distinction visible beside the report, as explained in renewal rate, retention rate and churn rate.

Work the renewal cohort in a spreadsheet

Put the following export in columns A:E, with headers in row 1 and subscription rows in rows 2:9. Scheduled shipment revenue is the expected net product amount for this box, excluding tax and shipping. Realized shipment revenue uses the same basis after allocated discounts, refunds and credits. These are illustrative US business records, not a benchmark.

In Excel, calculate the April cohort's cancellation rate with:

=COUNTIFS(B2:B9,"April",D2:D9,"Canceled")/COUNTIFS(B2:B9,"April")

This counts subscriptions meeting both cohort and outcome conditions, following Microsoft's COUNTIFS documentation. Replace April with May for the other cohort. Each loses one of four subscriptions, so each cancellation rate is 25.0%.

For the combined cohort, calculate cancellations, skips and pauses separately:

=COUNTIFS(D2:D9,"Canceled")/COUNTA(A2:A9)
=COUNTIFS(D2:D9,"Skipped")/COUNTA(A2:A9)
=COUNTIFS(D2:D9,"Paused")/COUNTA(A2:A9)

The results are 25.0%, 12.5% and 12.5%. Format rates to one decimal place; calculate from unrounded counts.

Measure revenue on the same rows

Retained shipment revenue compares realized revenue with the revenue scheduled for that window. This operational measure is not a claim about when prepaid cash becomes accounting revenue.

=SUM(E2:E9)/SUM(C2:C9)
=SUMIFS(E2:E9,B2:B9,"April")/SUMIFS(C2:C9,B2:B9,"April")

The second formula applies cohort conditions to both amounts, using the structure in Microsoft's SUMIFS documentation. Covirage is not affiliated with Microsoft.

Total retained shipment revenue is 200 / 400 = 50.0%. Cancellation removes $100 of scheduled revenue, while the skip and pause remove another $100. The revenue bridge therefore reconciles: $400 scheduled minus $100 canceled minus $100 skipped or paused equals $200 realized.

The subscription bridge also closes: eight starting subscriptions equal four renewed, one skipped, one paused and two canceled. Six relationships remain open, but only four produced a box in this window. This is why an average churn rate cannot substitute for the revenue view.

Where it goes wrong

• Treating every unshipped box as cancellation. Separate relationship status from shipment status before counting either.
• Mixing monthly and quarterly cadence. Compare equivalent renewal opportunities; a subscription with no box due has not failed to renew.
• Joining order lines directly. Aggregate shipment revenue first, then confirm subscription IDs remain unique in the cohort file.
• Using cash as shipment revenue. A prepayment, refund or delayed shipment can move cash without changing the subscription outcome.
• Reclassifying history silently. Keep the original cutoff and flag late corrections, so an operational change is distinguishable from an export revision.

Use the cancellation reason and the separate skip and pause counts to choose follow-up work. The reduce churn rate workflow covers interventions after the measurement is stable.

Compare your own renewal file in Covirage

Bring your subscription snapshot and order export, with the cohort and outcome rules documented. Covirage maps the columns; its tools compute and check the subscription and revenue figures, and the external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.

See renewal watch for the supporting workflow.