How retention rate is calculated
Pick a period — a month, a quarter, a year — and count the customers you had on the first day. On the last day, count how many of that same group are still customers. The second number divided by the first is your retention rate. Anyone who joined during the period is deliberately left out of both counts, because new customers would flatter the result and hide whether the existing ones actually stayed.
What counts as still being a customer depends on the business. For a subscription it is obvious: the payment either renewed or it did not. For a shop, a clinic or a trekking agency there is no cancellation event, so you have to set a window — someone who has not bought or booked within a gap that is normal for your category is treated as lapsed. Choose that window once, write it down, and then leave it alone, or your trend line will be measuring your own definition rather than your customers.
Why retention rate matters
Retention decides what an acquired customer is actually worth. Every plan built on customer lifetime value rests on an assumption about how long people stay, and retention rate is that assumption measured rather than hoped for. Lift it and every channel you run becomes more affordable, because the same acquisition cost now buys a longer relationship.
It also exposes problems that acquisition reporting hides. A business can post rising revenue while quietly losing customers, as long as it replaces them faster than it loses them. That holds up until the ad auction gets dearer or the market thins out, and then the leak shows itself all at once.
Where retention rate goes wrong
The usual error is publishing one number for the whole customer base. People acquired through a discount offer, through search and through a personal referral behave very differently, and a single blended figure flattens them into something you cannot act on. Split by cohort — by the month people joined, and by the channel that brought them — and the picture usually changes.
The second error is filing retention under customer service. A great deal of it is settled before the first purchase, by the promise the advert made and whether the product matched it. Retention is very often an acquisition quality problem wearing a different name.
How to act on it
Start with when people leave rather than how many. If most losses land in the first weeks, the problem is onboarding or expectation, and the fix sits in your welcome messages, your delivery, or the claims in your ads. If losses are spread evenly across the life of the relationship, it is ongoing value, and that is a product and service question rather than a marketing one.
Then change one thing for one cohort, hold everything else steady, and compare that cohort against the ones before it. A planned email programme for existing customers is a sensible first move: it is cheap, it is measurable, and it reaches people who have already paid you once. Watch the result next to churn rate so you are seeing the same movement from both directions.
