Customer retention rate: how to calculate it and what it tells you
- Aug 18
- 3 min read
Updated: 4 days ago
Introduction
Customer retention rate is the proportion of customers you keep over a period. It sounds straightforward and is calculated wrongly often enough to be worth walking through, mostly because new customers acquired during the period get included by mistake.
It also has an unusual property: small improvements in it produce disproportionately large improvements in profit.
1. How to calculate customer retention rate
Retention rate = (customers at end − customers acquired during) ÷ customers at startThe subtraction is the part people forget. If you started with 200 customers, finished with 220, and acquired 60 along the way, you retained 160 of the original 200 — a rate of 80%, not 110%.
Without removing new acquisitions you are measuring growth, not retention, and a business losing customers steadily can appear to be retaining them perfectly.
2. Choose a period that matches your cycle
The period has to be long enough that a customer would normally have bought again within it.
Measured monthly, a business whose customers buy quarterly will show alarming churn that is not real. Measured annually, a business with a weekly cycle will miss problems for months.
Match the window to how your customers actually behave, then keep it consistent so the figures are comparable.
3. How it differs from repeat purchase rate
These are often confused. Repeat purchase rate asks what share of customers ever bought more than once. Retention rate asks what share of customers you still have at the end of a period.
The first is a property of your customer base historically. The second is a rate of loss over time, which makes it the more useful operational number — it tells you whether things are getting better or worse.
Track both if you can, but if you only track one, track this.
4. Why small changes matter so much
Retention compounds. A customer retained is retained into a period where they can be retained again, so the effect multiplies rather than adds.
The practical consequence is that a few percentage points of retention improvement can move profit substantially — far more than the same effort applied to acquisition, because retained customers cost nothing to reacquire and typically spend more.
This is also why retention is the lever worth fixing before scaling acquisition. Pouring customers into a leaking business simply raises the cost of the leak.
5. Segment it or it will mislead you
An overall rate mixes groups that behave very differently.
Segment by acquisition channel — customers from different sources frequently retain at very different rates, and a channel producing cheap customers who leave quickly may be your worst. Segment by customer type or value where you can, since losing many small customers and losing one large one are different problems requiring different responses.
6. Pair it with revenue retention
Customer retention counts people. Revenue retention counts money, and they can point in opposite directions.
Losing several small customers while keeping the large ones produces a poor customer retention rate and a healthy revenue picture. Losing one major customer does the reverse.
Calculate both where your data allows. Deciding how worried to be requires knowing which one moved.
7. Turn the number into a routine
A rate calculated once is trivia. Calculated every period and reviewed, it becomes a management tool.
Add it to the small set of numbers you look at on a regular rhythm, and write down the reason whenever it moves. After a few periods you will know what your own retention responds to, which is more useful than any general guidance.
Conclusion
Calculate retention by removing customers acquired during the period, over a window that matches your purchase cycle, and keep the method consistent.
Segment by channel and by customer value, pair it with revenue retention so you know whether a change matters financially, and review it on a rhythm with the reasons recorded. Because retention compounds, it is usually the highest-return number on your dashboard.
.png)



Comments