Customer Churn vs Revenue Churn

Published by 4NLab · Updated · Our methods

Customer churn counts how many opening accounts leave. Revenue churn measures how much recurring revenue that opening cohort loses. A cancellation hits both. A downgrade hits only revenue churn, because the account is still there. Track both and you can tell whether the retention problem is broad or sits in a few high-value contracts.

Key differences

QuestionCustomer ChurnRevenue Churn
What is counted?Paying accounts lost from the opening customer group.Recurring revenue lost through cancellations and downgrades in the opening cohort.
What is the denominator?Paying accounts at the start of the period.Recurring revenue from those accounts at the start of the period.
Does a downgrade count?No, as long as the account stays active.Yes. The reduced recurring spend is contraction revenue.
Do larger accounts carry more weight?No. Each lost account counts once.Yes. Losing more MRR changes the rate more.
Can the rate be negative?Not under the lost-account definition here.Gross: no. Net: yes, if expansion from existing customers exceeds losses.

One scenario, two measures

Hypothetical worked scenario

A fictional SaaS business starts a month with 100 paying accounts and $100,000 MRR. Two accounts cancel: one paid $8,000 MRR and the other $200. Surviving accounts downgrade by $1,800 MRR and upgrade by $12,000 MRR. No new customers enter these churn calculations.

Customer churn = 2 ÷ 100 = 2%. Gross revenue churn = ($8,200 cancelled + $1,800 downgraded) ÷ $100,000 = 10%. Net revenue churn = ($10,000 losses − $12,000 expansion) ÷ $100,000 = −2%.

Only 2% of accounts left, but the opening cohort lost 10% of its MRR before expansion. The net rate is negative because of upgrades elsewhere. The enterprise cancellation still happened, and you still need to understand it.

Account by account: one cohort, three rates

A second fictional month, where the accounts that leave are small. Every rate below comes from this table.

AccountStarting MRREnding MRRWhat happened
Pine Analytics$3,000$3,600Upgraded
Quartz Media$1,500$1,500No change
Reed Health$1,000$600Downgraded
Sage Legal$800$800No change
Teal Travel$700$700No change
Umber Games$500$500No change
Vale Bakery$300$0Cancelled
Wren Florist$200$0Cancelled
Starting cohort$8,0008 accounts
  • Customer churn = 2 cancelled ÷ 8 starting accounts = 25%.
  • Gross revenue churn = ($500 cancelled + $400 downgraded) ÷ $8,000 = 11.3%.
  • Net revenue churn = ($900 lost − $600 expansion) ÷ $8,000 = 3.8%.

Yarrow Books, a former customer, came back this month at $400. It was not in the starting cohort, so it appears in none of these rates: report it as reactivation, and never subtract it from churn.

Here a quarter of the accounts left, yet only 11.3% of starting revenue did, because the cancellations were small. In the example above the opposite happened. You need both rates to tell the two situations apart.

Which should you use?

  • Customer churn tells you whether a broad group of accounts is leaving. Segment the lost accounts by plan, tenure, and acquisition cohort to look for a problem they shared.
  • Gross revenue churn puts a number on the recurring revenue at risk. Dig into high-value cancellations and downgrades even when only a few accounts left.
  • Net revenue churn shows whether expansion covers those losses. Read it next to the two gross measures. When lots of low-price accounts leave, customer churn can run high while revenue churn stays fairly low. No single rate tells the whole story.

Where comparisons can mislead

  • Use the same opening cohort and time window for both rates. Put new sign-ups in either denominator and the comparison stops meaning anything.
  • Say whether “revenue churn” means gross or net. Upgrades can push net revenue churn below zero while cancellations and contraction stay large.
  • A downgrade is not a lost customer. Reconcile cancellations, contraction, expansion, and new business as separate movements.

Read the full guides: Customer Churn and Revenue Churn.

Sources and methodology

These references explain the methods. The worked scenarios use hypothetical figures. They are not company results or current market benchmarks.

Found an error or a definition that differs from your reporting? Send a correction with the page URL and the method you use.