Gross vs net revenue retention: the floor and the engine.

VISIBILITY — JULY 17, 2026 —

A board deck that leads with net revenue retention of 118% will pass without comment. The number reads as growth compounding inside the installed base, and it usually is. But the debate over gross vs net revenue retention exists because that single figure can stay flat while the base underneath it rots. NRR is a blend. It nets expansion against churn and contraction, and blending is how a leak disappears from view.

The distinction is not academic. The two numbers answer different questions and drive different spending. Read one without the other and you fund the wrong department.

The floor and the engine

Gross revenue retention is the floor. It measures how much of last year's recurring revenue survives — after churn and downgrades, before a single dollar of expansion. GRR is capped at 100% by definition; you cannot keep more than you started with. A GRR of 90% means one dollar in ten walked out the door and was never replaced by the same customers.

Net revenue retention is the engine. It starts from the same cohort but adds upsell, cross-sell, and seat expansion. Because expansion has no ceiling, NRR can exceed 100% — the arithmetic that lets a company grow revenue from its existing accounts without closing a new logo. Bessemer has argued for years that best-in-class SaaS runs NRR above 120%, the point at which the base alone is a growth vehicle.

The relationship is fixed: NRR is always GRR plus expansion. That is exactly why NRR flatters and GRR does not.

The trap

Consider a $10M base. Over twelve months, $1.5M churns and contracts — GRR of 85%, weak. But three large accounts triple their spend, adding $3.5M in expansion. NRR lands at 120%. The deck says the business is compounding beautifully. The reality is that broad, quiet churn across the long tail is being papered over by a handful of enterprise upsells.

This is the failure mode boards miss. NRR of 120% with GRR of 85% is a different company than NRR of 120% with GRR of 97%, even though the headline is identical. In the first, growth depends on a few accounts continuing to expand — concentration risk that inverts the moment one of them renegotiates or leaves. In the second, the base is genuinely sticky and expansion is additive rather than compensatory.

The reason ARR-based metrics require this scrutiny is the same reason ARR and recognized revenue never tie out: the top-line number averages away the composition that actually determines durability.

The diagnostics that pull them apart

Two analyses separate a healthy base from a masked one.

Cohort churn analysis. Group customers by the quarter they signed and track retention by cohort over time. Averaged retention hides whether newer cohorts are churning faster than older ones — the early signal that product-market fit or onboarding has degraded. The ChartMogul cohort methodology and Lenny Rachitsky's benchmarks both treat cohort curves, not blended rates, as the honest view.

Customer concentration. Rank expansion dollars by account. If the top five customers drive most of the year's expansion, NRR is a story about five relationships, not the base. Pull the concentration and the reported NRR can collapse toward the GRR floor — which is where the real risk lives.

Run both together and the gap between GRR and NRR stops being a mystery. You can see exactly which accounts are carrying the number and which segments are bleeding.

Benchmark bands and what each decides

For growth-stage B2B SaaS, the working bands:

• GRR: 90%+ is healthy; below 85% signals a retention problem no amount of go-to-market spend will outrun. SMB-heavy books run lower than enterprise.
• NRR: 100% is the baseline where the base sustains itself; 110–120% is strong; above 120% is elite. KeyBanc's annual SaaS survey tracks these distributions across hundreds of private companies.

The decisions diverge cleanly. GRR informs product and support investment — a weak floor is a signal to fund onboarding, reliability, and customer success before anything else. NRR informs go-to-market — a strong engine justifies leaning into expansion motions, packaging, and land-and-expand pricing.

Confuse the two and you misallocate. A company with strong NRR and weak GRR that pours money into sales is filling a leaking bucket faster. The fix is upstream, in the product and support functions that determine whether customers stay at all. These retention decisions belong in the same operating discipline as pricing and headcount planning.

Why the rollup lies

Both numbers are only trustworthy computed from live subscription data at the account level. A quarterly rollup — the one most finance teams still assemble by hand — averages the leak away. It reports a blended NRR and calls it done, precisely at the resolution where composition matters most.

The problem is timing as much as granularity. By the time a quarter closes and the spreadsheet reconciles, the cohort that started churning is a quarter old and the concentration shift is already priced into next year's plan. Account-level retention, tracked continuously against the system of record, is what lets you catch a deteriorating cohort in month two rather than month five.

This is the recurring failure of static reporting: the answer arrives after the decision window closes. The teams pulling gross and net retention from live data — rather than reconstructing them quarterly — are reading the leak while it can still be plugged. For finance leaders building this out, the mechanics of continuous account-level tracking are worth pressure-testing against your current close process.

See how live account-level retention changes the reporting cadence. →

Retention is where product, support, and go-to-market meet the P&L, which is why both figures sit on the list of metrics a finance leader should own outright. Paired with the Rule of 40 and CAC payback, gross and net retention tell you not just whether the business grows, but whether it grows because the base is sound or in spite of the fact that it isn't. Watch only NRR and you will never know which.

■