Gross Revenue Retention

What Gross Revenue Retention Is

Gross Revenue Retention (GRR) is the floor. It measures the percentage of starting recurring revenue a company keeps from its existing customer base over a period, counting only losses — churn and contraction — before any expansion revenue is added back. Unlike Net Revenue Retention, GRR is mathematically capped at 100%. It cannot exceed the starting number because it excludes upsells, cross-sells, price increases, and new seats entirely.

A company with 92% GRR loses 8 cents of every dollar in its existing book each year. At $15M ARR, that is $1.2 million in annual revenue destruction before a single new deal is signed or a single upsell is closed. Expansion can cover it. Expansion covering it does not mean the hole is not there.

How Gross Revenue Retention Is Calculated

GRR Formula:

GRR = (Starting ARR − Churned ARR − Contracted ARR) ÷ Starting ARR × 100

• Starting ARR — recurring revenue from the cohort at the beginning of the period
• Churned ARR — revenue lost from customers who fully cancelled
• Contracted ARR — revenue lost from customers who downgraded (fewer seats, lower tier, reduced scope)
• Expansion ARR — excluded from both numerator and denominator

GRR is always measured on a defined cohort over a defined period, typically 12 months. The most common version is dollar-based (revenue dollars retained), though some orgs also track logo-based GRR (percentage of customers who did not cancel), which is a different and usually more flattering number.

GRR Worked Example

NRR of 112.5% reads as a healthy growth business. GRR of 92.5% reveals that $900,000 left through the back door before expansion papered over it. An investor who asks only about NRR gets a fundraising slide. An investor who asks both gets a picture of what the business looks like if growth stalls for one quarter.

Typical benchmarks by segment:

When Sales Orgs Track Gross Revenue Retention

Finance and the CFO treat GRR as the defensive health metric — the number that answers "how much of last year's revenue base is still intact?" Investors apply it as a due diligence filter: a company with 115% NRR and 72% GRR is expanding on a leaky foundation, and that foundation becomes a crisis when new bookings slow.

Customer Success leadership owns GRR operationally. AEs own it indirectly: every deal closed to a customer who churns inside 12 months is a GRR hit, and companies with chronic GRR problems often impose clawback provisions on quota attainment credit when deals cancel early. That is GRR pressure redistributed to individual rep behavior.

RevOps uses GRR to separate genuinely incremental expansion from expansion that is replacing lost revenue. A CS team generating $3M in upsells while losing $2.8M to churn is running a $200K net business, not a $3M expansion business, and GRR is the calculation that makes that visible.

Common GRR Misconceptions and Manipulation

The most expensive GRR misconception: that strong NRR makes GRR a secondary concern. It does not. Expansion requires CS headcount, AE capacity, and sales infrastructure. Churn consumes those same resources without producing net growth. A company spending $4M annually in CS motion to generate $1.5M in net NRR is destroying value even if the NRR metric looks impressive.

Definitional manipulation is the most common GRR gaming pattern, and it is almost always invisible without reading the footnotes. "Churn" can be defined narrowly: some orgs exclude seat reductions below 20% of contract value, classify product swaps as "migrations" rather than churn-and-new, or count a renewal at 60% of the original value as retained rather than contracted. Each definitional choice inflates GRR without changing the underlying economics. Asking how churn is defined is as important as reading the percentage.

Cohort selection is the second lever. Measuring GRR only on customers with 24-plus months of tenure excludes the highest-churn first-year cohort — the group most likely to have been pushed through by aggressive close tactics or SPIFF windows. The resulting number looks cleaner. The first-year churn problem stays hidden until it compounds.

Logo retention masking is a third pattern: a customer who reduces their contract from $200,000 to $80,000 did not churn. The $120,000 reduction is contraction — but whether it appears in GRR depends entirely on whether the org includes contraction in its GRR denominator, which is not universal.