Customer retention rate
The share of the customers you had at the start of a period who were still buying at the end.
01 What it is
Why anyone looks at this number
The share of the customers you had at the start of a period who were still buying at the end.
Keeping a customer is far cheaper than winning one, and retention compounds. A few points here move lifetime value more than any acquisition campaign.
02 The formula
How it is worked out
Customer retention rate = customers_retained ÷ customers_at_start grain : cohort × year unit : percentage source : Billing history joined to the customer master
Freeze the cohort at the start of the period and follow only those customers. Adding new customers turns retention into growth and hides the leak.
03 Worked example
The same number, with real inputs
| Customers at start | 3,340 |
| Still buying at end | 2,940 |
| Calculation | 2,940 ÷ 3,340 |
| Result | 88.0% |
12% churn a year implies an average life of about eight years. If that slips to 18%, average life falls to under six and lifetime value drops by a quarter.
04 What moves it
Four things that actually change this number
Service failures
Late and incomplete orders predict churn better than surveys do.
Relationship depth
How many products, sites and people connect you to the account.
Contract structure
Renewal dates, notice periods and switching cost.
Competitor activity
Usually visible in reduced order frequency before an outright loss.
05 Where the number lives
The system, the record and the fields
| System of record | Key record | Fields you need |
|---|---|---|
| Billing history joined to the customer master | Customer cohort | customer_id, first_order_date, last_order_date, cohort_month |
Freeze the cohort at the start of the period and follow only those customers. Adding new customers turns retention into growth and hides the leak.
06 How it goes wrong
Three ways this metric misleads people
Counting revenue instead of customers
One large account leaving is invisible in a revenue-weighted number until it is not.
Fix: Report logo retention and revenue retention separately.No definition of "lost"
In a non-contractual business nobody cancels; they just stop.
Fix: Define lapsed by elapsed time since last order, per segment.Averaging across cohorts
New cohorts always churn faster, so a blended rate moves with the mix of new business.
Fix: Report by cohort.08 Questions
Frequently asked
How do we measure churn without contracts?
By lapse: a customer is lost once they have gone significantly longer than their normal interval without ordering. Set the threshold per segment from observed behaviour.
Which matters more, retention or acquisition?
Retention almost always, because it compounds and costs less. Acquisition fills a bucket; retention stops it leaking.
One definition, everywhere it is used
SCIKIQ stores this metric once and serves it to every dashboard, board pack and agent that asks.