Customer retention rate: formula and how to calculate it

A building supplies distributor opened its quarterly review with a single slide: customer retention rate of 112%. The room applauded. The number was wrong, and the mistake was arithmetic, not commercial.

They had divided end-of-quarter customers by start-of-quarter customers. Because the company had won a lot of new accounts, the result sailed past 100%, which retention cannot do. Calculated properly, retention was 79%. You can work with 79%. You cannot work with a number that does not exist.

The formula fits on one line. What trips most companies up sits on either side of it: who belongs in the denominator, how long the window should be, and what the number means when revenue is moving the other way.

Customer retention rate measures how many of the customers you already had at the start of a period were still with you at the end. The formula is: (customers at the end of the period minus new customers won during the period) divided by customers at the start, times 100. New customers are excluded because they have not had time to stay or leave.

What is the customer retention rate formula?

Three numbers, and each one needs a written definition before you open the spreadsheet:

• CS: active customers on the first day of the period.
• CE: active customers on the last day of the period.
• CN: customers acquired during the period.

Retention rate = ((CE − CN) ÷ CS) × 100.

Back to the distributor. It opened the quarter with 1,200 active customers, closed with 1,350 and won 400 along the way. The correct calculation: (1,350 − 400) ÷ 1,200 = 0.7917, or 79.2% retention. The wrong one, which ignored those 400, returned 112.5% and implied the company had kept customers it never lost.

The gap between the two is not a methodology footnote. It is 250 customers walking out the back door while marketing celebrated the front one.

Why are new customers left out?

Because retention answers one specific question: of the people who were already customers, how many stayed? Someone who signed up yesterday has had no chance to cancel. Putting them in the numerator counts a win that has not happened yet.

The adjustment has a useful side effect. Once new customers are removed, acquisition stops disguising the leak. Plenty of companies grow revenue and shrink their installed base at the same time, and only notice when acquisition costs rise and the treadmill stops covering the loss.

Which time window should you use?

In a subscription business the window picks itself: the billing cycle. Everywhere else, nobody cancels anything. The customer simply stops buying, and you have to decide when you consider them gone.

A rule that holds up in practice: set the window at twice the median gap between purchases. A pet supplies store with a 45-day median uses 90 days. Bought in the last 90 days means active; did not means churned. It is a convention, and the only bad convention is the one you never wrote down.

Business model	Usual window	An active customer is one who…
SaaS or monthly subscription	Month	has a live contract on the last day
Annual contract	Year, by renewal cohort	renewed on the anniversary date
Fast-repeat e-commerce	90 days	purchased in the last 90 days
High-ticket e-commerce	12 months	purchased in the last 12 months
Recurring B2B service	Month or quarter	has an active contract billed in the period
Physical retail with loyalty ID	180 days	made an identified purchase in the period

Two retention rates built on different windows are not comparable. If last year's report used 180 days and this year's uses 90, the drop you are staring at may belong entirely to the method.

Customer retention is not revenue retention

This is the expensive mistake, and it is invisible on any dashboard that only counts heads. A services firm started the quarter with 200 customers and $300,000 in recurring revenue. By the end, 184 were still there: 92% customer retention, an excellent figure.

Except some of those 184 had downgraded. Revenue from them closed at $258,000, or 86% of the starting base. Add $24,000 in upgrades from the accounts that grew, and net revenue retention lands at 94%. Three numbers describing one book of business, each telling a different story.

In contract businesses, high customer retention next to low revenue retention usually means one thing: you are keeping the logo and losing the budget. Track both, as we cover in SaaS metrics.

What is a good customer retention rate?

The honest answer: it is only good compared with your own rate last period, calculated the same way. Industry benchmarks tend to blend windows and definitions of an active customer, which makes the comparison decorative.

There is still a useful order-of-magnitude reference from Brazilian e-commerce. The E-Consumidor 2026 study by Nuvemshop and Opinion Box found that 46.5% of shoppers buy again from a brand's own store, against roughly 30% on marketplaces. That is repeat purchase rate rather than retention, but it calibrates expectations: even on your own channel, fewer than half come back.

What drives that number is less glamorous than it sounds. In Opinion Box's CX Trends 2026 survey, when two products are priced similarly, previous experience with the company is the tiebreaker for 54% of consumers. Retention is built on orders that arrive on time and support that answers, not on a points program.

Retention and churn: one calculation, two sides

Over the same window and the same base, retention and churn rate add up to 100%. Retention of 79.2% means churn of 20.8%. If your two numbers do not sum to 100, somebody changed the base midway.

Choosing which one to talk about is a communication decision. Churn raises alarms faster in an operations meeting; retention frames a growth discussion better, because it speaks directly to LTV: every point of retention stretches customer lifetime and changes how much you can afford to pay for the next one.

The average hides what matters

A retention rate of 79% might be 90% among customers who arrived through organic search and 58% among those who arrived with a discount code. The average blends them and suggests a general problem that does not exist. The problem belongs to a channel.

So aggregate retention is for the board deck, and cohort analysis is for decisions. Split by signup month, acquisition channel, first category purchased and order value band. In almost any base, two or three slices concentrate the loss.

What to do once you have the number

The number alone changes nothing. Three moves that usually pay off:

1. Cut it by acquisition channel. If a channel brings customers who do not stay, it is more expensive than its cost per lead suggests, and that is a budget decision.
2. Build an operational trigger. Define the alert before the exit: no purchase in 60 days when the median is 45, product usage falling two weeks running, a ticket reopened twice.
3. Call the ones who left. Four or five conversations a month with former customers explain more than a quarter of internal theorising, and they help you read your NPS with less guesswork.

Frequently asked questions

What does the customer retention rate tell you?

It tells you what share of your existing base kept buying or kept their contract over a period. It is a signal of fit between what you promise and what you deliver. A falling rate with an unchanged product usually points to service, delivery times or a price change.

How do you calculate monthly retention?

Same formula with a one-month window: active customers on the 1st, active customers on the last day, and customers acquired that month. In non-subscription businesses, monthly figures swing a lot; a quarter gives a steadier read.

Can retention rate exceed 100%?

Customer retention cannot: you cannot keep more customers than you had. Revenue retention can, and anything above 100% means upgrades from the accounts that stayed more than covered the ones that left. Above 100% on a headcount basis is a formula error.

What is the difference between retention and loyalty?

Retention is the metric; loyalty is the behaviour behind it. You can post high retention with no loyalty at all, through long contracts or a lack of alternatives. When that barrier falls, that kind of retention disappears quickly.

Why does customer retention rate matter?

Because it determines how long a customer keeps paying and therefore how much you can invest to win the next one. Two businesses with identical acquisition costs and different retention rates have completely different economics.

The number is easy, the discipline is the rest

Companies that measure retention properly usually find something uncomfortable in the first quarter: the base grows more slowly than acquisition suggested. That is not bad news. It is the first honest reading of the real size of the business, and it only shows up when the definition of an active customer is written down and the calculation runs the same way every time.

The tedious part is joining the sources: the CRM knows who signed, the store knows who bought, finance knows who paid, and the three disagree. That is exactly the kind of question Sherlok answers: connect the systems, ask in plain language what retention was last quarter by acquisition channel, and get the analysis back instead of a place in the data team's queue.