Customer retention metrics explained: churn rate, retention rate and repeat rate
Customer retention metrics made plain: how to calculate churn rate, retention rate, revenue churn and repeat purchase rate, what a good number looks like, and a simple monthly routine you can run with a spreadsheet.

Most small business owners can tell you last month's sales without looking. Far fewer can tell you how many customers from last year are still buying. That second number is where the health of the business shows up, because a company that keeps its customers grows on a stable base, while one that keeps losing them has to refill a leaking bucket every month.
This guide explains the handful of retention numbers worth tracking: churn rate, retention rate, revenue churn and repeat purchase rate. You will get the formulas, a worked example, a sense of what "good" looks like, and a simple routine for checking them without a data team.
Why retention numbers matter
Keeping a customer is usually cheaper than finding a new one. A well-known Harvard Business Review piece puts the cost of acquiring a customer at anywhere from five to 25 times the cost of keeping one, and it is honest that the multiple depends on the study and the industry. Bain's Frederick Reichheld is often cited for the finding that a small gain in retention can lift profit by a large percentage. Treat the exact figures as direction, not a promise; the lesson is that retention is worth measuring.
Retention numbers also act as an early warning. Sales can look fine for months while loyal customers quietly drift away, because new customers cover the gap. By the time revenue dips, the problem is old. A churn figure you check every month shows the leak while it is still small.
Retention is a result, not a task
You cannot "do" retention directly. It is the outcome of service, onboarding, product quality and follow-up. The metrics only tell you whether those things are working. Use them as a scoreboard, then go and fix the causes, which the other guides in this series cover.
The core metrics, one by one
Churn rate
Churn rate is the share of customers you lost over a period. Stripe's explainer gives the formula as customers lost during the period divided by customers at the start, times 100. Start January with 200 customers, lose 10, and your monthly churn is 5 percent.
Two details trip people up. First, only count customers who were there at the start of the period; people who joined and left within the same month belong in a separate "early churn" check. Second, decide what "lost" means for your business before you count. For a subscription it is a cancellation. For a shop or a trades business it might be "has not bought in 12 months".
Retention rate
Retention is the mirror image. Stripe's version is customers at the end of the period, minus new customers acquired, divided by customers at the start. Start with 1,000, add 100 new, finish with 1,050, and you kept 950 of the original 1,000, which is 95 percent retention. Churn plus retention for the same group always adds to 100 percent, so track whichever one your team finds easier to talk about.
Revenue churn
Customer counts hide the dollars. Revenue churn measures money lost to cancellations and downgrades, and Stripe notes it matters most when spending varies or pricing is tiered, since losing a few large accounts can hurt more than losing many small ones. If you have ten clients and one of them is a third of your revenue, a 10 percent customer churn can be a 33 percent revenue event.
Repeat purchase rate
For a shop, café or online store with no subscription, churn is awkward because nobody formally cancels. Use repeat purchase rate instead: the share of customers who bought at least twice in a period. Stripe's article notes that for ecommerce, a repeat-customer rate of 20 to 30 percent is a common proxy for retention. It is a rough guide, so compare against your own history first.
A worked example
Here is how the numbers look for an imaginary design studio over one quarter.
| Measure | Value | How it is worked out | |---|---|---| | Clients at start of quarter | 40 | From your client list | | Clients lost | 6 | No work for 90 days, no open project | | New clients won | 8 | Signed in the quarter | | Clients at end | 42 | 40 - 6 + 8 | | Quarterly churn | 15% | 6 / 40 | | Quarterly retention | 85% | 100 - 15 | | Revenue at start (monthly) | $80,000 | Recurring plus average project spend | | Revenue lost from the 6 | $14,000 | Two of the six were large accounts | | Revenue churn | 17.5% | 14,000 / 80,000 |
Notice that the studio grew from 40 to 42 clients and still lost nearly a fifth of its revenue from existing clients. Looking only at the headline count would have hidden that.
Turning a short period into a yearly picture
A monthly churn rate sounds small and compounds fast. Losing 5 percent a month means keeping 95 percent each month, and 0.95 multiplied by itself twelve times is about 0.54, so you would lose roughly 46 percent of a starting group within a year. A common shortcut also says the average customer lasts about one divided by the monthly churn rate, so 5 percent monthly churn implies roughly 20 months. Both are estimates, but they make a "small" number feel as serious as it is.
What is a good number?
There is no universal target. Stripe's summary lists 3 to 7 percent monthly churn for small-business software, around 1 percent for enterprise software, and notes that results should be judged against your own past performance and direct competitors rather than a single figure. A neighborhood restaurant, an accounting firm and a software subscription will all look different.
Practical ways to set a target:
- Measure your own last four quarters first and aim to beat that trend.
- Compare against businesses of the same type and size, not against famous consumer brands.
- Set separate targets for new customers (first 90 days) and established ones, because the first group leaves more often.
Slicing the numbers so they tell you something
A single blended churn figure is a smoke alarm: it tells you something is wrong but not where. Break it down:
- By cohort. Group customers by the month they joined and compare how many are still active after 30, 90 and 365 days. Stripe recommends checking retention across several windows, such as 30 days, 90 days and one year.
- By source. Customers from referrals, ads and search often behave differently. Knowing which channel brings people who stay lets you spend more there.
- By plan, size or service. Small accounts may leave faster than large ones, or the reverse.
- By reason. Add a "why did you leave?" field to the record whenever someone goes. A short list of reasons beats guessing.
The US Chamber of Commerce's churn overview makes a related point: it says onboarding, relationship building and customer service account for a large share of churn, which is a reminder that the causes are usually in your own processes, not in bad luck. The onboarding guide in this series, Customer onboarding that reduces churn, goes deeper on the first weeks.
A simple monthly routine
You do not need a dashboard tool to start. A spreadsheet or the reports in your CRM is enough.
- Define "active" and "lost" in writing. One sentence each, so the number means the same thing every month.
- Pull the starting list. Everyone active on the first day of the month.
- Mark who left. Add a reason for each, even a rough one.
- Calculate churn, retention and revenue churn. Use the formulas above.
- Check one cohort. Look at customers who joined three months ago and see how many remain.
- Write one action. For example, "call the three accounts that stopped ordering".
- File the result next to last month's so the trend is visible.
Common mistakes
- Counting new customers as retained. Retention only looks at the group you started with.
- Changing the definition of "lost" mid-year. The trend becomes meaningless.
- Watching only the count. Revenue churn exposes the big-account problem.
- Treating the average as the story. A healthy blended number can hide a bad cohort.
- Measuring without acting. If a number never changes a decision, it is decoration.
- Comparing to the wrong benchmark. Borrow ideas from others, but judge yourself against your own history.
Tracking this in WorkBOS
If your customers live in a CRM, retention is mostly a reporting question. In WorkBOS you can keep contacts, deals and pipeline stages in one place, record invoices against each customer, and use surveys and forms to ask departing customers why they left, so the data sits beside the relationship rather than in a separate spreadsheet. The docs explain how records connect, and the Academy has short lessons on setting up a CRM. Our CRM setup guide covers the fields worth capturing, and the pricing page lists what each plan includes; the Free plan is enough to try this on a small customer list. Whatever tool you use, the habit matters more than the software.
Frequently asked questions
Q: What is the difference between churn rate and retention rate? A: They describe the same group from opposite sides. Churn is the percentage who left; retention is the percentage who stayed. For the same starting group and period, the two add up to 100 percent.
Q: How often should I calculate churn? A: Monthly is a good default for most small businesses, with a quarterly cohort review. If you sell annual contracts, a monthly figure will be mostly zero, so also check renewal rate at each renewal date.
Q: My business has no subscriptions. Which number should I use? A: Use repeat purchase rate, plus a "lapsed customer" count, meaning people who have not bought within the window that is normal for your trade, such as 6 or 12 months.
Q: Is it worth chasing a few percentage points of retention? A: Usually yes. Small improvements compound over the year, and keeping customers tends to cost far less than winning new ones. Check the effect against your own margins before assuming the large figures quoted in articles apply to you.
Q: Do these metrics need special software? A: No. A spreadsheet with a start-of-period list, a column for lost customers and a reason is enough. Software helps once you have hundreds of customers or several products.
Sources are linked in the text above: Stripe, Harvard Business Review, US Chamber of Commerce, WorkBOS guides.
Run it all from one place
Projects, clients, finance, people and AI employees in one system. Start free, no card needed.
Start free →