Net Revenue Retention: What It Is, How to Measure It, and How to Improve It

How to calculate NRR correctly

NRR measures how much revenue you retain from an existing cohort of customers over a period, including expansion (upsells, seat additions, plan upgrades) and excluding new customer revenue. The formula: NRR = (Starting MRR from a cohort + Expansion MRR − Contraction MRR − Churned MRR) / Starting MRR × 100. An NRR above 100% means you're growing revenue from existing customers even without any new customer acquisition — existing customers are expanding faster than others are churning. Above 120% is considered exceptional (Snowflake and Twilio sustained 150%+ at their peaks). Below 100% means you're shrinking even if you're acquiring new customers, which creates a leaky bucket problem that acquisition alone can't solve.

The four components of NRR and how to track them

NRR is made up of four components, each with distinct product analytics signals. Retention (no churn): tracked through L28, feature adoption breadth, and engagement frequency — the leading indicators covered in our SaaS retention piece. Expansion (upsells): tracked through product usage approaching plan limits, feature usage in higher tiers, and team growth signals (new user invitations, cross-department adoption). Contraction (downgrades): tracked through sudden drops in usage, seat reductions, or disengagement from paid features. Churn: tracked through dark period signals and feature abandonment. Build a single NRR driver dashboard that shows these four components side by side — it's one of the most useful executive-level dashboards for a SaaS company.

Which product behaviors predict expansion revenue

Expansion revenue is the most underinstrumented part of the NRR equation. The product behaviors that most reliably precede expansion events: (1) Usage volume approaching plan tier limits — track the percentage of accounts within 20% of their usage ceiling; these are your highest-priority expansion candidates. (2) Cross-department adoption — when a product that started in one team spreads to another department, multi-year expansion deals almost always follow. (3) Advanced feature adoption — accounts that use 4+ features and reach the most sophisticated features of the product have materially higher expansion rates. Instrument these signals in your analytics and pipe them into your CRM as expansion opportunity signals for your customer success team.

The relationship between NRR and CAC payback period

NRR and CAC payback period are deeply linked. With 100% NRR, a customer who costs $10,000 to acquire at $1,000/year MRR takes 12 months to pay back (ignoring margins). With 120% NRR, that same customer generates $1,200 in year 2, $1,440 in year 3 — the effective CAC payback shrinks dramatically on a lifetime value basis. This is why high-NRR SaaS companies can spend aggressively on acquisition and still produce excellent unit economics. The analytics implication: when optimizing growth spend, model out scenarios with different NRR assumptions — the relationship between acquisition CAC and retention economics is the fundamental lever of SaaS unit economics.

How to build an NRR improvement program

An NRR improvement program has three workstreams. First, retention: identify the leading indicators of churn (L28, feature breadth, dark periods) and build automated CS alerts. Second, expansion: instrument the product signals that precede upsells and build customer success workflows triggered by expansion signals. Third, onboarding quality: track whether customers who activate on more use cases in their first 90 days have materially better NRR — they almost always do. This means investing in onboarding that exposes customers to multiple product value propositions early, not just getting them to first value as quickly as possible.