Key Takeaways

• Gross revenue churn is churned MRR divided by MRR at the end of the previous month. Net revenue churn subtracts expansion MRR from the numerator first.

• Customer churn counts logos and revenue churn counts dollars, so a 40-account book can post 5.0% customer churn alongside either 0.5% or 20.0% revenue churn depending on which two accounts left.

• Stripe and Wall Street Prep both fold downgrades into gross revenue churn. On a book losing $12,000 to cancellations and $9,000 to downgrades, reporting cancellations alone understates churn by 2.25 points.

• Usage-based revenue has no cancellation event, so a customer dropping from $8,000 to $2,000 is contraction, not churn, until the contract ends or usage stays at zero across a defined window.

Why can the same month show a 5% customer churn rate and a 20% revenue churn rate?

Because the two metrics divide different things. Customer churn divides accounts lost by accounts at the start, and revenue churn divides MRR lost by MRR at the start. Every account is one logo, and no two are the same dollar amount.

Take a 40-account book holding $400,000 in MRR on September 1, spread like this:

Now run two versions of the same September, each losing exactly two accounts:

Identical customer churn, revenue churn 40x apart. One version costs you a rounding error and the other costs a fifth of the business. This is why I won't read a churn slide that shows only one of the two: the account number tells you how many relationships broke, and the dollar number tells you what breaking them cost.

Should downgrades count as revenue churn alongside cancellations?

Yes, in the standard definition. Stripe describes gross revenue churn as lost revenue "including customers who downgrade to a cheaper plan," and Wall Street Prep counts cancellations, non-renewals, and downgrades in the same numerator. A customer who halves their spend has taken half your revenue whether or not the logo stays on the slide.

The choice isn't cosmetic. Here's the same $400,000 book reported three ways, with $12,000 lost to cancellations, $9,000 to downgrades, and $16,000 gained in expansion:

Same month, same book, a 4-point spread between the highest and lowest defensible reading. What each one hides is the real decision:

• Cancellations only makes silent contraction invisible. Nobody cancelled, so nobody investigates, and the book shrinks anyway.

• Net revenue churn lets a handful of expanding accounts paper over the rest of the base, the same effect that shows up as inflated net revenue retention.

• Gross revenue churn is the version that names the leak, which is why it pairs with gross revenue retention as the harder number of the two to flatter.

Pick one, define it in writing, and don't switch definitions between board decks. Most of the disagreements I've seen about a churn number turned out to be disagreements about the numerator.

How do you measure revenue churn when a customer just used less?

Set a rule that separates a contract ending from consumption falling, because usage-based revenue produces no cancellation event to anchor the calculation to. A customer who billed $8,000 in July and $2,000 in August cancelled nothing and may bill $9,000 in September. Dropping that $6,000 into churned MRR turns ordinary volatility into a crisis on paper.

Classify the signal before you classify the dollars:

These practices keep the number honest on a consumption book:

• Anchor the denominator to a trailing three or twelve month average per account rather than a single prior month, so one heavy month doesn't manufacture churn in the next.

• Split committed revenue from overage before you calculate. Committed minimums churn on renewal dates and behave like subscriptions. Overage swings weekly and belongs in contraction, which is where the MRR movements report already puts it.

• Define the zero-usage window once and apply it to every account. Without it, accounts sit in limbo and revenue churn quietly understates itself every period.

Related terms

Churn reporting only makes sense read against the metrics on either side of it:

• Gross revenue retention is the mirror image, reporting the share of revenue kept rather than the share lost.

• Net revenue retention nets upsell against loss, the same adjustment that separates gross revenue churn from net.

• MRR movements is the report that separates contraction from churn line by line.

• Involuntary churn covers the losses that start with a failed payment rather than a decision.

• Revenue leakage covers revenue lost to billing errors, which never appears in churn at all.

• Contracted ARR covers the locked-in slice of a usage book, the part that churns on renewal dates instead of week to week.

FAQ

Is revenue churn the same as MRR churn?

Yes, on a monthly recurring revenue base the two terms describe the same calculation. MRR churn names the input explicitly, while revenue churn can also run on ARR or a quarterly base. Check which base a number uses before comparing it to anyone else's.

Can revenue churn come out negative?

Net revenue churn can, and gross revenue churn can't. Net subtracts expansion MRR from churned MRR, so a period where existing customers expand more than they cancel and downgrade produces a negative result. Gross revenue churn only counts losses, so its floor is zero.

Which revenue base should the revenue churn denominator use?

Use the recurring revenue your existing customers held at the start of the period, the same figure as MRR at the end of the previous month. Revenue from customers who signed mid-period stays out of both sides, otherwise new sales dilute the rate and hide the loss.

Should a board see customer churn or revenue churn first?

Show both, with revenue churn leading. Revenue churn quantifies the financial damage and customer churn shows how widespread the problem is. High revenue churn against low customer churn points at one large account, and the reverse points at a broken low-touch segment.