Choosing customer success software key metrics without vanity numbers

What to take away

Start from a familiar UK scenario

Picture a Manchester software firm with 300 business accounts on annual contracts. Asked for a retention number at the Monday leadership meeting, the customer success lead produces three different figures from three dashboards. Nobody in the room trusts any of them.

That is the normal starting position, and the fix is not a new platform. Agree what each metric means, who owns it and where the source data lives.

Platforms such as Gainsight, Totango, ChurnZero, Planhat and Vitally each ship their own default definitions for renewal, retention and adoption. Which one you run matters less than whether those definitions are written down and owned.

Which metrics earn their place

Four metrics answer most questions a UK board will ask.

Gross renewal rate shows how much recurring revenue renewed before expansion is counted. Net revenue retention adds expansion and subtracts churn, so it shows whether the base is growing on its own.

Product adoption depth measures how many licensed seats or modules are genuinely used each month. Time to first value measures the days between contract signature and the first outcome the customer recognises.

Everything else, from health scores to ticket sentiment, is supporting evidence. Keep that set small and label it diagnostic.

Which numbers are vanity metrics

Vanity numbers rise without telling you whether revenue is safe. Page views, session counts, logins, seats provisioned, tickets closed and a bare health score with no trend or segment are the usual offenders.

A raw NPS score is the most misused of the lot: one figure with no segment, sample size or trend gives a board nothing to act on, and the same applies to an average ticket sentiment score that ignores severity and account value.

Apply one test. If a number can improve while gross renewal and net revenue retention fall, it is diagnostic at best and vanity at worst, and it does not belong on the headline slide.

Define each metric in writing

A metric dictionary is one page per measure. It states the calculation, the data source, the reporting period and the named owner.

Agree whether renewal rate is counted on contract anniversary or on the calendar quarter. Both are defensible; mixing them is not. Write down how you treat mid-term upgrades, credits and multi-year deals.

The CIM data and insights training material is a reasonable starting point for teams that need analysts to interpret customer data consistently rather than rebuild definitions each quarter.

Handle telemetry and consent properly

Product analytics usually involves personal data: user identifiers, session records, email addresses. Under UK GDPR you need a lawful basis, a retention period and a route for access requests.

Practical steps include minimising event payloads and excluding free-text fields from telemetry. If your tool sends data outside the UK, the transfer mechanism belongs in the vendor contract.

Work through a labelled example

Suppose a Leeds firm holds £1,200,000 of annual recurring revenue across 200 accounts. Over a year, £1,020,000 renews, £90,000 churns and £150,000 comes from expansion.

Gross renewal is 85 per cent. Net revenue retention is 97.5 per cent. Both figures are illustrative, but the arithmetic shows why reporting only one of them flatters the picture.

Now add adoption. If 60 per cent of licensed seats are active monthly, the retention figure is fragile even where the contract renewed.

Connect metrics to attribution

Retention rarely has one cause. A renewal may follow a training session, an executive review and a support fix in the same quarter. Deciding which gets credit is a method choice: first-touch, last-touch, linear split and time-decay each hand the same renewal to a different team. Choose between the standard attribution methods and stay with one.

Choose one method, apply it consistently and state its weakness in the report. Consistency matters more than sophistication when the audience is a finance director.

Report on a fixed rhythm

Weekly operational reviews need adoption and ticket data. Monthly reviews need renewals in flight. Quarterly and annual reviews need the four headline metrics with written commentary.

Keep the reporting layers apart. Weekly operational reviews need adoption and ticket data; monthly reviews need renewals in flight; quarterly and annual reviews need the four headline metrics with written commentary. One dashboard trying to serve all three is how the same question ends up with three answers.

Common questions

How many key metrics should a team track?

Four headline metrics plus a small diagnostic set is enough for most UK teams. Anything beyond that tends to go unread.

Does UK GDPR apply to product telemetry?

Yes, where the data identifies or could identify a person. That includes user-level event data and support records, so document your basis and retention.

Should health scores replace renewal rate?

No. Health scores are predictive and vendor-specific. Renewal and net revenue retention are the outcomes the business is judged on.

How often should definitions change?

Rarely, and only with a dated note in the dictionary. Silent changes to a metric definition destroy trust faster than an imperfect calculation.