AI SaaS Metrics Due Diligence: The Private Equity Playbook (2026)

Headline ARR is almost never the right number. PE teams should be auditing ARR composition across: new ARR, expansion ARR, contraction ARR, and churned ARR — on a monthly basis for the trailing 24 months.

Common ARR landmines AI can surface quickly: multi-year contracts recognized as ARR in year one (inflating momentum), professional services revenue bundled into SaaS ARR, beta customers on zero-cost trials counted as ARR, and auto-renewing month-to-month contracts booked as annual.

NRR is the single most important metric for SaaS valuation — and the most frequently misrepresented. The benchmark: best-in-class enterprise SaaS sits at 120–140% NRR; mid-market SMB products average 100–115%. Anything below 90% is a structural problem.

Manipulation tactics to watch for: excluding churned customers from the cohort denominator (the most common), using a rolling 3-month window instead of trailing 12, counting upsells from reactivated churned accounts as expansion, or defining "churn" as contract cancellation rather than revenue departure.

CAC payback measures how many months of gross margin it takes to recover the cost of acquiring a customer. Below 12 months is excellent for enterprise; below 18 months is acceptable; above 24 months signals a growth engine that's destroying equity value at scale.

The calculation is deceptively simple (CAC ÷ (ACV × gross margin %)) but the inputs are typically distorted. Common issues: sales compensation included in CAC but not at full burdened cost, marketing spend allocated inconsistently across quarters, or trial-to-paid conversions attributed to organic rather than paid channels.

Aggregate churn rates lie. A company can show 5% annual churn while sitting on a ticking time bomb — if that 5% is concentrated in the oldest, largest cohorts. Cohort-level retention analysis reveals the truth: is the product getting stickier over time, or is it a leaky bucket that's masking churn by growing new customer acquisition?

Request monthly revenue by customer with cohort inception date. Build a 24-month cohort table. If cohorts from 18+ months ago are retaining worse than recent cohorts, that's usually a product quality signal — the company fixed something, but the legacy base hasn't caught up.

The churn rate tells you how much is leaving. Churn diagnosis tells you why — and whether the reason is addressable with capital. There are four buckets: product-market fit churn (the product doesn't solve the problem), product quality churn (bugs, reliability, usability), go-to-market churn (wrong customers signed at unrealistic expectations), and market churn (customers' businesses failed or were acquired).

Only market churn is outside management's control. The other three are investable problems — but only if the team has diagnosed them accurately. A company claiming "market conditions" as the primary churn driver should be treated skeptically until the underlying data confirms it.

The Rule of 40 (ARR growth rate + EBITDA margin ≥ 40%) is the standard PE benchmark for SaaS financial health. But the ratio obscures more than it reveals in isolation. A company at 35% growth and -5% EBITDA is very different from one at 5% growth and 30% EBITDA — even if both score 30 on Rule of 40.

In the current market environment (2026), buyers are weighting profitability more heavily than in 2021. A company scoring 40+ via growth-only will face a valuation haircut unless it has a credible path to 15%+ EBITDA margins within 24 months of investment.