C4 · Economics & metrics · identity

Gross Revenue Retention

Also: GRR

Definition. GRR is beginning recurring revenue of a fixed cohort less gross churn and contraction, divided by that beginning base. It lies in [0,100%] when losses are disjoint and bounded by the starting revenue base. It is revenue weighted, not the fraction of customers surviving or a causal PMF measure.

established Last updated 2026-06-18 Source: Definition is accounting on a declared cohort. Benchmark bands are illustrative historical practitioner ranges, not current universal PMF thresholds.

Formula

Gross Revenue Retention identity

GR R_{[a, b]} = \frac{R _{a} - gross churn _{[a, b]} - contraction _{[a, b]}}{R _{a}}

Plain English: GRR = (Beginning ARR - Contraction - Churn) / Beginning ARR

Notation: GRR_[a,b] = (beginning cohort recurring revenue - gross losses over [a,b]) / beginning cohort recurring revenue

Benchmark by stage

Source: Definition is accounting on a declared cohort. Benchmark bands are illustrative historical practitioner ranges, not current universal PMF thresholds.

Stage	Gross Revenue Retention	Notes
Enterprise SaaS (best-in-class)	> 90%	Illustrative historical range; investigate causes using matched cohorts and renewal terms rather than infer PMF from the number alone.
Enterprise SaaS (good)	85–90%	Illustrative historical range; investigate causes using matched cohorts and renewal terms rather than infer PMF from the number alone.
Mid-market SaaS	80–90%	Illustrative historical range; investigate causes using matched cohorts and renewal terms rather than infer PMF from the number alone.
SMB-focused SaaS	75–85%	Illustrative historical range; investigate causes using matched cohorts and renewal terms rather than infer PMF from the number alone.
Concerning (any segment)	< 75%	Illustrative historical range; investigate causes using matched cohorts and renewal terms rather than infer PMF from the number alone.

Naive vs corrected

Version	Formula
Naive	Using NRR as a proxy for GRR (conflates expansion with retention, masking gross churn behind expansion — a company with 80% GRR and 120% NRR looks healthy but has serious retention problems)
Corrected	Report GRR and NRR with the same starting cohort, window, currency and loss classifications. Their difference equals the included expansion rate only under the same reactivation and adjustment policy.

Common errors

Treating GRR as customer survival or a direct PMF test
Counting a downgrade to free as retained paid revenue
Combining annual beginning balances with one month of losses
Comparing segments or cohort ages without adjustment
Double-counting churn and contraction or losses on revenue added after the starting snapshot

Where this sits

Part of the Economics & metrics (C4) cluster in the GTM World Model. Related to the model's "On a common cohort and convention with no reactivation/other adjustments: NRR=GRR+Expansion/Beginning_R. If GRR=.85 and NRR=1.15, the expansion increment is .30 of the starting revenue base, not 30% of surviving customers." equation.

nrr →mrr →ltv →burn multiple →

How to cite this

@misc{shalvi_gtm_metric_grr_2026,
  author = {Singh, Shalvi},
  title  = {Gross Revenue Retention — GTM World Model Metrics},
  year   = {2026},
  url    = {https://shalvisingh.com/gtm/metrics/grr}
}

Singh, Shalvi. "Gross Revenue Retention — GTM World Model Metrics." shalvisingh.com, 2026. https://shalvisingh.com/gtm/metrics/grr