Logo Churn

Why count logos at all

Revenue churn is the figure that matters to the P&L, so the reasonable question is why track accounts separately. Three reasons.

First, logo churn is the honest measure of product-market fit. Revenue churn is weighted by price, so a business can hold revenue steady while steadily failing most of its customers — the small ones simply do not move the number. Logo churn treats every customer as one customer, and a rising logo churn rate against flat revenue churn is an early signal that the bottom of your base has stopped succeeding with the product.

Second, it is the input to the LTV formula. Implied average lifetime is 1 divided by customer churn, not revenue churn, and substituting one for the other silently rescales every unit-economics figure you derive.

Third, support, onboarding and success costs scale with account count rather than with contract value. A quarter that loses many small logos costs less revenue and roughly the same amount of organisational attention.

The gap is the signal

Neither rate means much alone; their ratio is a diagnosis.

• Revenue churn well above logo churn — your larger accounts are leaving. The worked example above shows the shape: a routine-looking 3.75% logo quarter concealing a 6.5% revenue quarter. This is the more urgent pattern, because large accounts are slower to replace.
• Logo churn well above revenue churn — you are shedding small accounts while retaining big ones. Sometimes healthy consolidation upmarket; sometimes an entry-level plan attracting buyers the product was never built for.
• The two roughly equal — churn is not correlated with account size, which is typical of a self-serve product on a single price point.

What counts as one logo

Harder than it sounds, and the answer changes the metric. A customer with three subscriptions who cancels one has not churned. Two subsidiaries billed separately may be one commercial relationship or two, and if they cancel together in the same month you will either report one loss or two depending on a decision nobody wrote down. A parent company consolidating five business units onto one contract will look like four cancellations and one expansion in the raw data, when nothing was lost at all.

The workable rule is to define the logo at the level you sell and renew at, and to handle consolidations as an explicit adjustment rather than letting them flow through as churn. In a self-serve product where each account is genuinely independent this problem does not arise; in enterprise it arises constantly and is the main reason two teams compute different logo churn from the same database.

Reactivations and the denominator

A customer who cancels in March and returns in June poses the same question here as elsewhere: new logo or returning one. Counting them as new inflates acquisition and quietly improves CAC, since winning back a lapsed customer is much cheaper than acquiring a stranger. Counting them as a reactivation into their original cohort is more honest about retention and is the convention most cohort tooling assumes. Either works; using both in different reports does not.