SaaS Metric

SaaS metrics like ARR, MRR, and churn exist only in the CFO's head and a few investor-deck slides. When the board asks 'what's our net revenue retention?', someone scrambles to define it from scratch each quarter. Different people calculate MRR differently — one includes annual contracts divided by twelve, another counts only monthly subscribers.

None — AI cannot calculate or monitor business health metrics because no formal definitions exist. Every metric request requires a human to decide what the metric means before computing it.

The finance team maintains a Google Doc called 'Metric Definitions' that describes ARR, MRR, gross churn, net churn, and NRR in narrative form. The doc says 'ARR is the annualized value of all active subscriptions' but doesn't specify whether paused accounts, free trials with credit cards, or multi-year prepaid contracts count. The VP of Sales and the CFO still argue about the number at every board meeting.

AI can reference the definitions document, but cannot compute metrics consistently because the narrative descriptions leave too many edge cases unresolved for programmatic calculation.

Each SaaS metric has a documented specification with formula, input variables, and edge-case rules. The MRR definition specifies: 'Sum of (monthly contract value) for all subscriptions where status IN (active, past_due) AND type NOT IN (internal, partner_comp). Annual contracts contribute contract_value/12.' The finance team reviews definitions quarterly and publishes a versioned 'Metrics Handbook' to the company wiki.

AI can implement metric calculations from the documented formulas, but cannot programmatically validate that the formulas match the actual computation logic in spreadsheets or BI tools without manual comparison.

SaaS metric definitions live in a metrics-as-code repository. Each metric — ARR, MRR, gross churn, net churn, NRR, LTV — is defined in a YAML file with typed input fields, a SQL or Python formula, inclusion/exclusion filters, and unit tests with expected outputs. The CI pipeline validates that every metric definition compiles and passes its test suite before merging. Analysts can query 'what changed in the NRR definition between Q3 and Q4?' from the git history.

AI can compute any SaaS metric by executing its versioned definition against live subscription and revenue tables, detect formula changes, and flag when metric values diverge from historical trends.

SaaS metrics are nodes in a formal business intelligence ontology. Each metric definition links to its source subscriptions, the contract fields it reads, the revenue recognition rules it applies, and the board reporting context it serves. An AI agent can answer 'which subscription fields feed into our NRR calculation, and what happens to NRR if we reclassify partner comps as paid?' by traversing the ontology without asking a human.

AI can autonomously perform impact analysis when metric definitions or source objects change, simulate 'what-if' scenarios across the metric dependency graph, and recommend definition refinements based on investor benchmarks.

SaaS metric definitions are self-documenting. When the subscription management system adds a new plan type, the ARR and MRR definitions auto-update their inclusion criteria. When the revenue recognition engine changes its allocation rules, affected metrics regenerate their formulas. When the CFO creates a new board metric, the system assembles its definition from existing building blocks. The metric specification is always a real-time reflection of how the business actually measures itself.

Can autonomously maintain the complete SaaS metric definition lifecycle — creating, updating, testing, and documenting metric formulas as business models and accounting rules evolve — without human specification work.

Ceiling of the CMC framework for this dimension.