Customer retention rate: what good looks like
The honest answer to "what is a good retention rate" is that nobody can tell you, and the businesses that improve fastest are the ones that stop asking and start measuring their own consistently.
What is a good business plan retention rate?
There is no single defensible figure, because published benchmarks mix twelve-month contracts with rolling monthly plans, count dead records as active, and rarely say which they did. A more useful target is your own rate, calculated the same way every month on a database you have cleaned - and a trend that moves in the right direction over a quarter.
Why published benchmarks will mislead you
Search for a business retention benchmark and you will find numbers between about sixty and ninety per cent, quoted confidently and almost never with a definition attached. They are not comparable with each other, let alone with your business.
Three things vary underneath those figures, and each one moves the result by more than any operational change you could make this year.
The first is contract length. A business selling twelve-month contracts with a cancellation fee will report far better annual retention than one selling rolling monthly plans, without being any better at keeping people. The second is what counts as a customer: some operators count anyone with an unexpired contract, others count anyone who attended in the last ninety days, and the two differ by a third in the same building. The third is whether the database has ever been cleaned, which is where most of the distortion lives.
So the benchmark is not a target. Your own figure, calculated the same way every month, is the only version that tells you whether last quarter’s changes did anything.
How to calculate it without flattering yourself
Customers at the end of the period, minus those who joined during it, divided by customers at the start. Joiners are excluded so growth cannot hide churn.
| Month | Active at start | Joined | Left | Retention |
|---|---|---|---|---|
| April | 420 | 38 | 31 | 92.6% |
| May | 427 | 29 | 44 | 89.7% |
| June | 412 | 22 | 52 | 87.4% |
| July | 382 | 19 | 35 | 90.8% |
Read the trend, not the month. A single month tells you almost nothing - June looks alarming until you notice it is the start of the summer travel period, which lands in the same weeks every year. Compare each month with the same month last year, and each quarter with the quarter before it.
Three rules that make the number honest
Almost every business we look at is reporting a retention rate several points better than reality. The cause is always the denominator.
Write off contacts who have gone quiet for six months
Hundreds of records for people who left years ago sit in most business databases, inflating the customer count and flattering every percentage calculated from it. Archiving them is unglamorous and slightly painful, and it is the single change that makes every other number mean something. Keep the records, mark them lapsed, take them out of the active count.
Pick one definition of active and never change it mid-year
Either an unexpired plan, or an attendance in the last ninety days. Both are defensible. Switching between them, or letting different reports use different ones, produces a number that moves for reasons nobody can explain. Write the definition down where the person building the report will see it.
Separate the segments before you draw conclusions
One-off clients, package buyers and subscription customers churn on completely different clocks, and blending them produces an average that describes nobody. Report them separately even when the totals are small - a pack buyer who has not rebooked in three weeks is already gone, while a customer who missed three weeks is not.
The three levers that actually move retention
Once the number is honest, the interventions that move it are unglamorous and mostly happen in the first ninety days of a plan.
Attendance in the first six weeks
The strongest predictor of whether somebody is still a customer next year is how often they came in the first six weeks. A customer who attends eight times in that window behaves very differently from one who attends twice, and the second group is both larger and easier to fix, because nothing has gone wrong yet - they simply have not built the habit.
Whether anyone notices a gap
A customer who stops coming and hears nothing for eleven weeks has been told, accurately, that their absence is not noticeable. This is the mechanism behind most quiet churn, and it is covered in detail in why customers stop coming.
How the renewal is handled
A renewal that arrives as a single automated message three days before expiry converts far worse than a ladder that starts a month out and escalates to a human phone call. The customer has not changed. The amount of notice they were given has.
The three numbers worth on a weekly report
Retention rate itself is a quarterly number - it moves too slowly to steer by weekly. What is worth looking at every week is the count of customers who have not attended in thirty days, the count of plans expiring in the next thirty days, and the number of new joiners who have attended fewer than three times. All three are leading indicators, and all three are actionable the same day.
How Ragnos AI handles this
Most businesses can produce this analysis exactly once, by hand, in a spreadsheet, and then never repeat it consistently enough for the trend to be readable.
Ragnos AI keeps the definitions fixed and the calculation the same every month: one active-customer definition, lapsed contacts archived out of the denominator automatically at six months, and one-off work, packs and plans reported separately rather than blended.
The weekly view is the operational one - dormant customers, expiring plans, and new joiners who have not built a habit yet - so the number on the quarterly report is a consequence of work already done rather than a surprise.
Frequently asked questions
What is a good business retention rate?
No published figure is comparable enough to use as a target, because the definitions underneath them vary more than the performance does. Calculate your own the same way every month, on a cleaned database, and judge it by whether the trend improves over a quarter.
What is the difference between retention rate and churn rate?
They are two views of the same number: churn is the proportion of customers lost in a period, retention is the proportion kept. If 8 per cent of customers leave in a month, monthly churn is 8 per cent and retention is 92 per cent. Pick one and report it consistently rather than switching between them.
Should new joiners be included in the retention calculation?
Not in the numerator. Customers at the end of the period minus those who joined during it, divided by customers at the start, is the standard form. Including joiners lets a good sales month disguise a bad retention month, which defeats the purpose of measuring it.
How often should a business calculate retention?
Report it monthly and read it quarterly. The underlying behaviour moves too slowly for a monthly figure to be meaningful on its own, and reacting to single months is how businesses end up changing things that were working.
Why is my retention rate higher than it feels on the business floor?
Almost always the denominator. Old contacts who left years ago are still counted as customers, so the percentage is calculated against a base that includes hundreds of people nobody expects to see again. Archive anyone unresponsive for six months and the figure will drop, and start being useful.
Keep reading
All featuresSee your own numbers, honestly
Thirty minutes on a working account, including what your retention rate looks like once the dead contacts come out of the denominator.
No card needed. No obligation. English or Arabic.