Gross vs net revenue retention: the math investors actually grade

The math: what GRR and NRR actually measure

Both calculations work from a cohort of customers from twelve months ago and measure where that cohort's revenue is today. The difference is whether expansion is included.

Gross revenue retention (GRR):

GRR has a mathematical ceiling at 100 percent. The best you can do is keep every dollar of revenue from the original cohort, you cannot create new dollars within GRR. A 95 percent GRR means you lost 5 percent of your year-ago revenue base to churn and contraction.

Net revenue retention (NRR):

NRR can exceed 100 percent when expansion exceeds losses. A 115 percent NRR means the original cohort is now paying you 15 percent more than they were twelve months ago, after netting all losses against all expansion.

The gap between GRR and NRR is your expansion engine. A business with 90 percent GRR and 115 percent NRR has a 25-point expansion engine. A business with 95 percent GRR and 105 percent NRR has a 10-point expansion engine. Same NRR ceiling at the higher GRR business but a much weaker expansion contribution, which often signals a different go-to-market model rather than worse performance.

2026 GRR and NRR benchmarks by segment

Synthesised from Bessemer 2026 State of the Cloud, KeyBanc 2024 SaaS Survey, and SaaS Capital 2025 Retention Report:

Two patterns to note. First, GRR rises with average contract value across segments, reflecting the higher switching costs and more deliberate procurement at larger ACV. Second, the expansion gap is largest in consumption-based SaaS (Snowflake, Datadog, MongoDB) because the pricing model creates expansion automatically when usage scales, but the same model makes NRR more volatile when usage contracts.

The floor / ceiling framework investors apply

Sophisticated investors and underwriters apply a two-step framework when evaluating retention. First, what is the GRR floor? This sets the maximum compression risk if expansion stops. Second, what is the NRR ceiling? This sets the multiple-tier the business is positioned for.

A worked illustration. Two SaaS companies, both at $25M ARR, both reporting 115 percent NRR. Company A has 95 percent GRR; Company B has 80 percent GRR. Both look identical on the NRR slide of the investor deck. The investor model differs significantly:

• Company A (95 / 115): 20 points of expansion offsetting 5 points of structural loss. If expansion compresses by 50 percent (to 10 points), NRR drops to 105 percent. The business stays clearly above 100 percent and remains in the same multiple band.
• Company B (80 / 115): 35 points of expansion offsetting 20 points of structural loss. If expansion compresses by 50 percent (to 17.5 points), NRR drops to 97.5 percent. The business falls below 100 percent and into a meaningfully lower multiple band.

Investors model both scenarios and weight them. Company B usually trades at a 1 to 2 turn multiple discount to Company A despite identical headline NRR. This is the floor / ceiling framework in operation: same NRR, very different structural resilience, materially different valuation.

How to improve GRR vs how to improve NRR

The two metrics respond to different interventions. Knowing which one you are trying to move dictates the operating playbook.

To improve GRR: Reduce churn and contraction. The interventions are onboarding quality, support responsiveness, product reliability, addressing ICP fit on at-risk cohorts, dunning recovery for involuntary churn (see the involuntary churn page), cancellation flow optimisation, and founder-led winback at smaller scale. These work on the structural stability of the customer base.

To improve NRR: Add expansion. The interventions are upsell motion (CS-led or sales-led), cross-sell to multi-product, seat expansion mechanics (per-user pricing that grows with the customer's team), usage-based expansion (consumption pricing that scales with success), and tier upgrade prompts at usage milestones. These work on the expansion engine.

The sequence matters. Investing in expansion when GRR is weak is leaky-bucket work: you are adding water faster than you are losing it, but the bucket still leaks. Investing in GRR when expansion is weak is structurally sounder but does not produce the multiple-expansion effect that NRR drives. The right sequence is almost always: defend GRR first, then build expansion on top of a stable floor.

Frequently asked questions

Related reading on ChurnCost

• NRR benchmarks by stage, the percentile drill-down.
• Logo vs revenue churn, the other key two-rate distinction.
• Revenue churn deep-dive, the calculation underlying GRR.
• Valuation impact, how NRR moves the multiple.
• B2B SaaS benchmarks 2026, the full segment table.
• Churn cost at $10M ARR, where this framework starts to dominate.