Gross Revenue Retention (GRR) is the share of recurring revenue retained from an existing customer cohort over a defined period, before any expansion revenue is counted. The formula is (starting MRR − churned MRR − contracted MRR) / starting MRR, where churned MRR is revenue lost to cancellations and contracted MRR is revenue lost to downgrades on accounts that stayed. Because upgrades, add-ons, and cross-sells are excluded by construction, GRR is capped at 100% — anything reported above that ceiling is mislabeled NRR.
Net Revenue Retention (NRR) adds expansion MRR to the numerator and can exceed 100%, which is what makes the contrast useful. A subscription brand running 105% NRR on 88% GRR is masking 12 points of base churn with 17 points of expansion. GRR shows the leak; NRR shows whether refills outpace it. Operators who only watch NRR can miss churn until expansion slows and the base surfaces unmasked.
Read GRR by acquisition cohort, not in aggregate. A 95% vintage averaged with a 75% vintage looks like 85% blended and tells the operator nothing — sign-up months, channels, and onboarding versions retain differently. Cohort analysis by sign-up month or quarter is the honest read; blended GRR is a reporting number, not a decision number.
Split the churn component into voluntary and involuntary. Voluntary churn is active cancellation; involuntary churn is payment failure — expired cards, declined transactions, closed accounts. The playbooks do not overlap: dunning flows and card-updater enrollment address the involuntary slice; product, pricing, and onboarding changes address the voluntary slice.
GRR is the base-health number for subscription DTC and SaaS-style commerce models. It caps the LTV the brand can model against existing customers and tells the operator whether the base is healthy enough to acquire against.