Revenue Metrics

The metrics that describe recurring revenue

Annual recurring revenue and monthly recurring revenue measure the same quantity at different resolutions: the annualised or monthly value of contracted recurring revenue at a point in time. Both are run rates, not measures of revenue earned during a period, and treating them as recognised revenue is a category error that produces avoidable arguments in board meetings.

What belongs in them is contracted, recurring and expected to continue. One-off implementation fees, professional services and uncommitted usage overages do not qualify, however reliably they recur in practice. This is the most common inflation, and it matters beyond tidiness: every efficiency ratio computed downstream — retention, magic number, payback, lifetime value — takes recurring revenue as an input, so inflating it inflates all of them simultaneously.

Per-customer and per-deal units

Average contract value annualises so that a three-year deal does not appear three times larger than an equivalent annual one. Total contract value counts the whole term including non-recurring elements. Average revenue per account divides recurring revenue by customer count and decodes what is driving growth: rising ARPA with flat account count means expansion is working, while rising account count with falling ARPA means you are moving downmarket, deliberately or otherwise.

Average selling price is the cleanest read on discounting. Falling ASP while list price holds means discounting is increasing, which is a pricing-power signal that usually appears before it shows up in margin.

Movement, not level

Recurring revenue is most useful decomposed into new, expansion, contraction, churned and reactivated. The net change gives direction; the components give the mechanism, and only the mechanism is actionable. Two companies posting identical net new ARR can be in opposite conditions — one growing on new logos with heavy churn, the other on expansion with almost none.

The reconciliation that catches most categorisation errors: opening recurring revenue, plus new, plus expansion, minus contraction, minus churn, should equal closing exactly. Where it does not, something is miscategorised and every retention metric built on those categories is unreliable.

Segmenting before comparing

Blended metrics across self-serve and enterprise describe neither. The two motions typically differ by an order of magnitude in contract value and acquisition cost, and a blend will justify decisions that are wrong for both segments simultaneously.