Most customer success dashboards measure activity: touchpoints logged, QBRs completed, tickets closed, accounts contacted. All easy to count, all easy to hit, and none of them predict whether a customer renews.
The four that matter
1. Net dollar retention. The outcome measure. It combines churn, contraction, and expansion on the existing base and is the closest thing to a single grade for the function.
2. Gross revenue churn. Because NDR alone hides it. Strong expansion can mask heavy churn and the netted figure looks stable while the base erodes.
3. Retention cost per customer. Retention bought at any price is not a result. This is what stops the first two being gamed by throwing people at accounts.
4. A back-tested health score. The only leading indicator in the set. Worthless unless it has been checked against customers who actually churned.
Four numbers. The first three are lagging and honest; the fourth is leading and must be validated to be worth anything.
Why activity metrics persist
They are available on day one, they respond immediately to effort, and they make a team look busy in a quarter where retention has not resolved yet.
The cost is that they are gameable without lying: a CSM can hit every touchpoint target on the accounts easiest to reach while the at-risk ones go untouched, and the dashboard will be green.
If activity is tracked, track it as a diagnostic for why an outcome moved, never as the target.
Segment before judging
Self-serve and enterprise customer success are different jobs with different economics. A blended NDR moves with mix, so a good enterprise quarter can hide a self-serve problem entirely.
Report the four numbers per segment, and per cohort where tenure varies widely: first-year customers churn at rates that make a blended figure meaningless.
The boundary with sales
Expansion sits in both functions and the split is a design decision. Whichever you choose, be consistent about it in retention cost, or the cost of expansion will land in retention and make CRC look worse than it is.
Where this sits
Customer success metrics are Measurement in the Tenbound Pipeline Architecture Standard. The recurring finding across all four is that they surface Market problems late: accounts acquired outside the ICP cost more to keep, expand less, and churn more, and no customer success motion fixes a targeting decision made a year earlier.