GRR is different from Net Revenue Retention. NRR adds expansion back in and can exceed 100%. GRR strips expansion out and shows the pure durability of the book. Investors read both numbers side by side.
Gross Revenue Retention (GRR) is the percentage of recurring revenue retained from an existing customer cohort over 12 months, excluding expansion from upsell or cross-sell. The formula is starting ARR minus downgrades minus churn, divided by starting ARR. GRR is capped at 100% by definition, because the only movements it counts are losses. Best-in-class enterprise SaaS businesses post GRR of 90% or higher, which signals a durable book that holds its floor without leaning on expansion.
The six things to understand about Gross Revenue Retention before it goes on a board slide, into a diligence room, or against a benchmark. Each one is a place real operators lose credibility by defining the cohort loosely or quietly letting expansion creep into the numerator.
Gross Revenue Retention measures what happened to a fixed cohort of customers over a defined period, usually 12 months. It takes the recurring revenue those customers paid at the start of the window and asks how much of it is still landing at the end, after downgrades and churn, but with no credit given for upsell or cross-sell. GRR is the pure durability number.
The arithmetic is starting ARR minus downgrade ARR minus churn ARR, divided by starting ARR. If the cohort started at 100% and ended at 92%, 8% of the book eroded through downgrades and cancellations. The cohort is locked at the start of the window, so new logos acquired during the period never contribute to either the numerator or the denominator.
GRR can never exceed 100% because expansion is deliberately excluded. The formula captures only downgrades and churn, both of which are downward moves. A perfect GRR of 100% means every customer in the cohort kept every dollar of their original subscription. That is extremely rare at scale, which is why 90% to 95% is the enterprise-SaaS target.
The widely referenced benchmarks are 85% as healthy, 90% as strong, and 95% or higher as best-in-class for enterprise SaaS. SMB-focused businesses tend to post lower GRR because the segment churns harder. GRR under 80% is a durability problem that no expansion motion can mask forever. GRR under 70% signals a product fit or segmentation issue that needs addressing before growth capital is deployed.
A high GRR means the book is durable on its own, before any expansion lift. Lenders, acquirers, and conservative investors underwrite against GRR because it tests how much revenue would still be there if the expansion motion stopped tomorrow. A business with 95% GRR can borrow against the book. A business with 80% GRR cannot, no matter how flattering the NRR headline looks.
GRR is calculated against subscription records tied to accounts in the CRM, with downgrade movements captured on the account and churn flagged at the renewal. The number is reconciled against invoiced revenue in the finance system every period. A retention number produced from one system without the other agreeing is the single most common source of a GRR that falls apart under diligence scrutiny.
The GRR formula is simple arithmetic, but the inputs are where it goes wrong. The six cards below walk through each piece of the equation, the question it answers, and the common ways teams let expansion leak into a number that is supposed to exclude it.
GRR begins with the annualized recurring revenue of a defined cohort of customers on a defined start date. The cohort is locked. Any customer not in that starting set never counts toward either the numerator or the denominator. Locking the cohort is what isolates the behavior of the existing book from the noise of new acquisition.
Downgrade ARR is recurring revenue lost from customers who stayed in the cohort but reduced their spend. Fewer seats, lower tiers, dropped modules, renegotiated price. Downgrades are not churn, because the customer is still active. They erode GRR even when logo retention looks perfect. A disciplined account review process catches downgrades before they land, instead of discovering them at the renewal.
Churn ARR is recurring revenue lost from customers who left the cohort entirely during the period. The customer cancelled, did not renew, or dropped to zero spend. Honest churn accounting records the cancellation on the date the subscription actually ended, not the date the renewal process started, because dating games are how retention numbers get laundered.
Expansion from upsell, cross-sell, higher tiers, additional products, or usage-based uplifts is deliberately left out of the GRR calculation. That is the whole point. GRR shows what the book is worth without expansion propping it up. The expansion number has its own home in Net Revenue Retention. Mixing the two produces an inflated number that misleads every reader of it.
The denominator is always starting ARR of the locked cohort. The numerator is starting ARR minus downgrades minus churn. The ratio is reported as a percentage. If the result is 93%, the cohort is now worth 93% of what it was 12 months ago, with no credit for expansion. The percentage format is what makes GRR comparable across businesses of different sizes and vintages.
The standard window is 12 months, measured as period-end floor ARR of the cohort divided by period-start ARR of the same cohort. Many teams publish a trailing twelve-month GRR alongside the most recent single-period view, so a sharp recent change is not hidden by earlier smoothing. The dated snapshot and the trailing view belong next to each other on the board slide.
A single retention percentage never tells the full story. GRR and NRR describe different parts of the same cohort, and the space between them is where the real health of the business lives. The six cards below describe how the two numbers differ, when each one matters most, and how to read them together.
Gross Revenue Retention measures the cohort with no expansion counted. The formula is starting ARR minus downgrades minus churn, divided by starting ARR. GRR can never exceed 100%, because the only moves it counts are losses. It is the hardest and most defensible retention metric because no amount of upsell can cover for a leaky book.
Net Revenue Retention adds expansion back into the formula. It describes the whole motion of the existing book: what was lost, what was retained, and what was grown. NRR above 100% means the business can grow from the existing base alone. The two numbers measure the same cohort through different lenses, and both lenses are needed to read the book correctly.
If GRR is 90% and NRR is 125%, expansion contributed 35 points. That gap is the single most informative view of the expansion motion. A wide gap means upsell and cross-sell are doing the work. A narrow gap means the business is holding on but not growing the base. Both numbers side by side are what boards actually read.
Lenders, acquirers, and conservative investors underwrite against GRR because it tests the floor. A business can prop NRR up with aggressive expansion into a cohort that is churning fast, and the headline looks fine for a while. GRR cuts through that. If GRR is weak, the business has a product or segmentation problem that no amount of upsell can mask forever.
Growth-stage investors and public market analysts underwrite against NRR because it tests the compounding engine. A business with 130% NRR grows meaningfully without any new logo acquisition, and that compounding is what justifies premium valuation multiples. NRR is the single most-cited retention number in SaaS valuation benchmarks for exactly this reason.
A disciplined board report shows GRR, NRR, and the gap, segmented by cohort and segment. Reading only one hides half the story. Reading both, together, with the segments visible, is what makes a retention conversation defensible. The teams that publish only NRR are usually the ones with a GRR problem they would rather not surface.
A GRR number is only as credible as the data pipeline it was built from. The six cards below describe the pattern used by subscription businesses that reconcile retention cleanly: the CRM owns the cohort and the loss movements, the finance system owns the recognized revenue, and the two must agree every period before any retention number leaves the building.
Every active subscription lives on an account in the CRM with its start date, end date, annualized value, and status. Downgrades are captured when terms change. Churn is flagged when a renewal fails. The cohort for a given GRR window is defined by filtering accounts that had an active subscription on the chosen start date. Everything flows from the account record.
The two GRR movements each have their own workflow. Downgrades are captured at the account level with a reason, a dollar impact, and an effective date. Churn is flagged at the renewal with the cancellation date. Reporting both motions means reading the CRM, not reconstructing them from the ledger after the fact when the audit trail is already cold.
Every subscription in the CRM maps to an invoice in the finance system. Monthly, the two are reconciled. No CRM subscription without a billed contract. No billed contract without a CRM subscription. Downgrades line up with reduced billings. Cancellations line up with invoices that stop. The reconciliation is what makes GRR defensible in a diligence room.
Because every subscription is tied to an account, GRR rolls up by segment, industry, geography, product line, and cohort year. The board report can show GRR for the enterprise segment, the SMB segment, each product, and each cohort vintage from the same source data. Aggregate GRR hides the segments where the number is actually broken. Cohort slicing surfaces it.
Honest GRR comes from a renewal process that opens 90 to 120 days before the contract end date, surfaces risk early, and gives the account team time to intervene. Risk signals from product usage, support activity, and executive sponsor change land on the renewal record. Downgrades are negotiated before the renewal date, not discovered on it. The renewal pipeline is where GRR is actually made or lost.
The honest teams publish an internal policy: how downgrades are logged, how churn is dated, how mid-term renegotiations are counted, how price concessions versus product-swap downgrades are treated, and how the cohort is defined. Every quarter the GRR is calculated against the same written policy. When definitions are documented, the number survives scrutiny instead of shifting under it.
Strkr is a CRM that captures downgrades on the account, churn at the renewal, and the full subscription record in one place. Retention numbers reconcile to the system the revenue team already uses, instead of being reassembled from spreadsheets at month end and defended line by line in diligence.
The widely referenced benchmarks are 85% as healthy, 90% as strong, and 95% or higher as best-in-class for enterprise SaaS. SMB-focused businesses tend to post lower GRR because the segment churns harder. GRR under 80% signals a durability problem that no expansion motion can mask over time. GRR under 70% usually indicates a product fit or segmentation issue that needs to be addressed before more growth capital is deployed into acquisition.
The formula is starting ARR minus downgrade ARR minus churn ARR, divided by starting ARR, expressed as a percentage. The cohort is locked at the start date. New logos acquired during the window are never counted. Expansion from upsell and cross-sell is deliberately excluded. The window is almost always 12 months. If the cohort was worth 1,000,000 at the start and 920,000 at the end after downgrades and churn, GRR is 92%.
GRR excludes expansion. NRR includes it. GRR is capped at 100% because it counts only downward moves. NRR can exceed 100% because expansion is added back in. The gap between the two numbers describes the strength of the expansion motion. A business with 90% GRR and 125% NRR has real losses but a strong expansion engine that more than covers them. Both numbers belong on the board slide, read together.
No. GRR is capped at 100% by definition. The formula counts only downgrades and churn, both of which are downward movements against the starting cohort. There is no mechanism in the GRR calculation for the number to go up. A GRR reported above 100% means expansion has leaked into the numerator, which is the single most common definitional error in retention reporting and the fastest way to lose credibility in diligence.
No. GRR measures what happened to a fixed cohort of customers that existed at the start of the period. New logos acquired during the window are excluded from both the numerator and the denominator. The point of locking the cohort is to isolate how the existing book held up on its own. A business can post low GRR and still grow fast overall if new logo acquisition is strong, which is why GRR has to be read alongside bookings and NRR.
They are the same metric under two different names. Gross Revenue Retention and Gross Dollar Retention (GDR) both measure recurring revenue retained from an existing customer cohort, excluding expansion. Some firms use GDR, others use GRR, and a few publish "gross retention rate" or similar. The formula is identical: starting ARR minus downgrades minus churn, divided by starting ARR.
Monthly for internal operating cadence, quarterly for board reporting, and at a labelled point in time for any investor conversation. The standard view is a trailing twelve-month GRR, with the most recent single-period view published alongside it so a sharp recent change in downgrades or churn is not hidden by earlier smoothing. The dated snapshot plus the trailing view belong on the board slide alongside NRR and the gap between them.
Two systems that must agree. The CRM holds the account, the subscription record, the downgrade history, and the renewal. The finance system holds the invoice, the recognized revenue, and the cash. GRR is calculated from CRM subscription movements and reconciled against invoiced revenue every period. A retention number produced from one system without the other agreeing is the most common source of a GRR that falls apart under diligence.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
