How to Calculate Customer Retention Rate: Formula, Example and Benchmarks

Customer retention rate = ((customers at the end of the period, minus new customers acquired during the period) / customers at the start of the period) x 100. If you began the quarter with 500 customers, added 80 and ended with 540, the calculation is ((540 - 80) / 500) x 100 = 92%. Your churn rate is the remainder, 8%. Subtracting new acquisitions is the step people skip, and skipping it lets growth hide the customers you lost.

Retention rate is the most quoted number in subscription businesses and one of the easiest to calculate misleadingly. This covers the formula, a worked example, the difference between counting logos and counting revenue, what a good rate looks like by model, and the part that actually changes outcomes: finding out why the number is what it is.

The customer retention rate formula

CRR = ((E - N) / S) x 100

• S = customers at the start of the period
• N = new customers acquired during the period
• E = customers at the end of the period

Subtracting new customers is what makes this a retention measure rather than a growth measure. Without that subtraction, a company losing a quarter of its base while acquiring aggressively can report a number above 100% and feel fine about it. The formula asks a narrower question: of the customers you already had, how many stayed?

A worked example

A B2B SaaS company starts Q1 with 500 customers. During the quarter it signs 80 new ones and 40 existing customers cancel. It ends the quarter with 540.

CRR = ((540 - 80) / 500) x 100 = (460 / 500) x 100 = 92%

Churn rate is 100% minus retention rate, so 8% for the quarter. Note that quarterly and annual figures are not interchangeable and do not simply multiply: 8% quarterly churn compounds to roughly 28% annually, not 32%, because each quarter churns a smaller remaining base. Always state the period alongside the number, since a rate quoted without one is uninterpretable.

Logo retention versus revenue retention

Counting customers treats every account as equal, which is rarely true. A company that keeps 95% of its logos but loses its three largest accounts is in more trouble than the headline suggests. That is why subscription businesses track revenue retention alongside it.

Gross revenue retention is the honest one, because it cannot be flattered by upsells. Net revenue retention above 100% means your existing base grows on its own, which is the strongest signal a subscription business can produce. Track both: an NRR of 105% built on 82% gross retention is a company frantically expanding a leaking bucket.

Measure by cohort, not just in aggregate

A blended retention rate mixes customers who signed up three years ago with those who signed up last month, and the old, loyal ones will prop the number up long after new-customer quality has declined. Cohort analysis fixes that: group customers by the month they signed up and track each group over its own lifetime. If you have not built one before, our guide to how to do a cohort analysis walks through the five steps and the three ways to read the chart.

What you are looking for is whether recent cohorts retain better or worse than older ones at the same age. If the January cohort was at 88% by month six and the May cohort is at 76% by month six, something changed in acquisition, onboarding or the product, and the aggregate number will not show it for another year. Cohorts give you that year back.

What is a good customer retention rate?

It varies enormously by business model, and cross-industry averages are close to useless. Established B2B SaaS companies commonly target annual logo retention in the 85% to 95% range with net revenue retention above 100%, since contracts are long, switching costs are high and the product is usually embedded in daily work. Self-serve consumer subscriptions run substantially lower and are considered healthy at levels that would alarm an enterprise vendor, because the purchase is individual, cheap and easy to abandon.

Price point, contract length and how essential the product is to someone's job move the benchmark more than industry does. The comparison worth making is your own trend, segmented. A rate that is flat overall while your newest cohorts and your smallest plan decline is telling you something a single number never will.

Why customers actually leave

Across most subscription products the causes cluster into a short list. The customer never reached the value they signed up for, which is really an onboarding failure showing up months later. A persistent friction or reliability problem wore them down. The champion who bought it changed jobs and nobody else understood why the tool was there. Price stopped matching perceived value, often after a usage drop nobody noticed. Or a competitor solved the specific job better.

There is also a category teams consistently underestimate: involuntary churn, where a customer intended to stay but a card expired, a payment failed, or an invoice went unpaid and unchased. It routinely accounts for a meaningful share of cancellations in self-serve businesses, and it is the cheapest churn to fix, since it needs no product change at all. Making sure billing failures and unpaid invoices are chased automatically instead of manually often recovers more retention in a month than a quarter of feature work.

The rest is diagnosis, and the signals almost always exist before the cancellation. Login frequency falls. The number of features an account touches narrows. Support tickets rise and the tone sharpens. A satisfaction score slips. Those shifts usually appear months ahead of a cancellation, which is precisely the window in which the account can still be saved.

How to improve retention rate

Fix the first two weeks. A large share of eventual churn is decided during onboarding, when a customer either reaches real value or forms the habit of not opening your product. Measure time to first value and cut it.

Watch the early signals, not the renewal date. A quarterly health review finds problems that started in week three. Falling usage and rising ticket volume in the same account should trigger attention immediately, not at renewal.

Rank the causes by revenue, not by frequency. The most-mentioned complaint is often not the most expensive one. Cluster the reasons customers give, attach the accounts and revenue behind each, and fix in that order.

Close the loop when you fix something. Customers who reported a problem and later hear it was solved are measurably more likely to stay and to report the next one, which keeps your diagnostic signal alive.

From the number to the reason

Calculating retention rate takes ten minutes. Knowing why it moved is the part that takes teams weeks, usually because the behavioral data sits in an analytics tool while the reasons sit unread in support tickets, survey responses and reviews that nobody has time to theme.

Joining those two is the entire job. When a cohort's retention dips and the feedback from that same cohort is already clustered into ranked causes with the revenue attached, a retention review becomes a decision rather than an investigation. That is what customer retention software is for at the diagnostic layer, and it is the same evidence that churn analysis depends on. For the tactical side of acting on what you find, see our guide to reducing churn, and if you want the leading indicator rather than the lagging one, what counts as a good NPS score covers the sentiment signal that tends to move first.