Every other risk in an acquisition is a one-time event. Churn is the one that compounds, every month, for as long as you own the business.
What the numbers actually mean
Monthly churn | Average customer lifetime | What it implies |
|---|---|---|
1% | 100 months | Exceptional. Verify it twice. |
3% | 33 months | Healthy B2B SaaS. |
5% | 20 months | Normal for SMB tools. Manageable. |
8% | 12.5 months | You replace the entire customer base annually. |
12% | 8 months | This is a marketing business, not a software business. |
Voluntary vs involuntary
Not all churn is a verdict on the product. Involuntary churn — expired cards, failed payments — is often 20–40% of the total, and it is the cheapest thing in the entire business to fix. A dunning sequence that recovers half of it can add a point of net retention in a fortnight.
So when you see 7% churn, ask for the split. If it's 4% voluntary and 3% involuntary, you have found something worth buying: a real problem with a known, cheap solution the seller never got round to.
Look at the cohorts, not the average
An average churn rate hides the shape. Ask for retention by signup cohort. Two patterns matter:
A cliff in month one or two — an activation problem. Fixable, often quickly, and the seller has usually never looked at it.
A steady bleed across all cohorts — a value problem. Much harder, and it means the product isn't load-bearing for the customer.
Churn on revenue, not on logos
Losing 5% of customers who were on your cheapest plan is not the same as losing 5% of revenue. Ask for net revenue retention: if existing customers expand faster than others leave, NRR is above 100% and the business grows without acquiring anyone. That is a materially better asset than the logo churn alone would suggest.
The question to ask the seller
"Of the last twenty customers who cancelled, do you know why?" A seller who can characterise the answer has been paying attention and probably has a fixable, identified problem. A seller who has never asked has handed you both a risk and, if you're the sort of buyer who will ask, an edge.