GRR vs NRR (Gross vs Net Revenue Retention)

Use GRR to inspect gross revenue losses on a fixed starting cohort, with disjoint churn and contraction. Segment by renewal exposure, customer size, tenure, currency and pricing policy. Low retention warrants investigation; it does not identify the cause without additional evidence.

Use NRR to inspect the total recurring-revenue change in that same starting cohort. Separate quantity/usage, price, expansion and losses, and measure the cost and contribution of retaining or expanding accounts. One observed value is not a future growth guarantee.

A cohort can lose many small customers while growing revenue from a few large ones; revenue retention and logo survival answer different questions. Conversely, retained customer counts can conceal downgrades. Neither GRR nor NRR alone identifies concentration, customer outcomes or profitability. On matching period, cohort and adjustment conventions, NRR−GRR equals included expansion divided by starting revenue. If reactivation or other adjustments are included in one metric, reconcile those explicitly. Assuming persistent annual NRR of 120%, $10M would grow to about $61.9M after 10 years before new logos. This is conditional arithmetic, not an estimate of persistence or cost. Expansion and service costs may consume the revenue gain. Usage pricing can increase or decrease revenue with consumption and price; margin and predictability depend on workload costs and commitments. Its causal effect on NRR and value requires a business-specific comparison.

Revenue-retention stocks and flows remain separate from customer-count survival and contribution value. The gap between NRR and GRR measures included expansion on a common basis, not the share of growth that is free of acquisition or service cost.