Logo Churn: How to Calculate Customer Churn Without Fooling Yourself

Logo churn is the share of your customers who stopped being customers during a period. It counts relationships, not rupees. A departing account that paid you a hundred a month and a departing account that paid you a hundred thousand a month are the same single unit of loss in this metric, and that is on purpose. Logo churn is asking whether people keep choosing you, not how much money walked out of the door.

That is a genuinely useful question, because customer count is the cleanest signal of product-market fit at the small end of your base. Small customers are usually the first to feel a weak onboarding process, a missing feature or a support gap, and they leave without a negotiation. Revenue metrics can absorb that for a long time before it becomes visible, particularly if a few large accounts are growing at the same time.

The standard calculation

Logo churn rate equals customers lost during the period divided by customers at the start of the period, multiplied by one hundred. Customers lost means accounts whose recurring relationship ended and whose recurring revenue went to zero. Customers at the start means the count on the first day of the period, before any acquisition in that period. The result is a rate for that period only and carries no annual meaning on its own.

A worked example

A company starts January with 500 customers. During the month, 15 cancel and 40 are won. Logo churn for January is 15 divided by 500, or 3 per cent. Logo retention is 97 per cent. The closing count is 525, which becomes February's denominator. If instead you divided 15 by the closing count of 525, you would report 2.86 per cent, and if you divided by the average of 500 and 525 you would report 2.93 per cent. All three are defensible; only one can be used consistently, and the choice must be written down.

Turning a monthly rate into an annual one

Annual churn equals one minus the quantity one minus the monthly rate, raised to the power of twelve. Continuing the example, one minus 0.03 is 0.97, and 0.97 raised to the twelfth power is about 0.694. Annual logo churn is therefore about 30.6 per cent, and annual logo retention about 69.4 per cent. Multiplying 3 per cent by twelve would have given 36 per cent, overstating the loss by more than five percentage points and, at higher rates, by far more.

Most disputes about a churn number are really disputes about the denominator. Each of these is used in practice and each answers a slightly different question.

The renewal-cohort method deserves particular attention in businesses selling annual terms. On a twelve-month contract a customer physically cannot churn in month four, so a monthly series across a mixed base is mostly an artefact of the renewal calendar. Asking what share of this period's renewals did not renew produces a figure that is comparable from month to month.

The flattering version is the one where new customers are added to the denominator. It requires no dishonesty, only a query written against a monthly customer table rather than an opening snapshot, and it makes churn fall every time acquisition rises. A business can run a whole year on that number and conclude that a growth push improved retention, when nothing about retention changed.

The second common error is treating any plan change as churn. A customer who moves from a larger plan to a smaller one is still a customer. Recording them as churned inflates the rate, breaks the reconciliation between opening and closing customer counts, and hides the more interesting fact that the account is shrinking rather than leaving. The third is the failure to separate involuntary churn: cards expire, mandates lapse, and bank-side failures happen, and none of those are a verdict on your product.

Good logo churn is low, stable and boring, and it is concentrated in reasons you would expect: businesses closing, needs genuinely changing, budgets disappearing. It is roughly the same across acquisition channels and roughly the same across plan tiers. When you look at the cohort curve, it flattens after the first few months, which means customers who get through onboarding tend to stay.

Bad logo churn has a recognisable profile. It is front-loaded, with most departures happening before the customer ever reached routine use, which points at onboarding and at the promises made during the sale rather than at the product itself. Or it is concentrated in one channel or one segment, which usually means you are acquiring customers who were never going to fit. A blended monthly average conceals both patterns, which is why segmentation is not an optional refinement of this metric but the point of it.

Revenue churn measures the same departures in money and will disagree with logo churn whenever customer sizes vary. Gross retention is the revenue-weighted mirror of survival, capped at one hundred per cent. Net retention adds expansion and can exceed one hundred per cent, which is why it can look excellent in a business that is losing customers steadily. Contraction is a reduction in spend by a surviving customer and is not churn at all. Logo retention is simply the complement of logo churn and must be quoted from the same denominator, or the two will not add up in the same report.

Related pages on measuring and reducing customer loss:what a CRM is, CRM versus spreadsheets, lead management software, sales automation, AI CRM features, and CRM for small business.