GRR is the honest denominator in retention. It tells leadership how leaky the bucket is before expansion math is allowed to patch it, which is why boards and investors read GRR alongside NRR, not instead of it.
GRR, or gross revenue retention, is the share of recurring revenue a company keeps from its existing customer base over a period, excluding any expansion revenue. The formula is starting ARR minus contraction minus churn, divided by starting ARR. GRR caps at 100 percent, which is why it shows pure stickiness while NRR can hide churn behind upsells.
The six things to know about gross revenue retention before you pick a target, defend a board number, or compare GRR against NRR in a quarterly review.
GRR measures the recurring revenue retained from the customers you started the period with, without crediting any upsell, cross-sell, or seat expansion. It answers one question: of the ARR you had on day one, how much is still there on the last day. Anything new that happened inside the base is deliberately left out.
GRR equals starting ARR minus contraction minus churn, divided by starting ARR, times one hundred. Contraction is revenue lost on downgrades. Churn is revenue lost from customers who left entirely. New logos acquired during the period are not in the numerator or the denominator, because GRR only measures the existing cohort.
GRR cannot exceed one hundred percent, no matter how many customers expanded during the period. That ceiling is the whole point. Because expansion is excluded, GRR cannot be gamed by a few large upsells, and the number falls honestly whenever anything is lost. NRR has no ceiling, which is why the two metrics tell different stories.
A healthy SaaS business usually runs GRR of ninety percent or higher. Enterprise-heavy portfolios often clear ninety-five percent because annual contracts and procurement cycles slow departures. SMB-heavy portfolios run lower. Anything below eighty-five percent is a signal to investigate, regardless of how strong the NRR number looks on top.
NRR can look great while GRR quietly erodes, because expansion from the top of the base can mask churn at the bottom. GRR strips that masking out. Investors read GRR to judge how durable the customer base is on its own, before growth tactics are allowed to compensate for retention problems.
GRR on its own tells you how sticky the base is. NRR on its own tells you how much the remaining base is growing. The gap between them tells you how much of your net retention comes from expansion versus pure stickiness. Boards and investors expect both numbers, calculated the same way, over the same window.
GRR looks simple on a slide and gets messy in a spreadsheet. The errors show up at the edges: new logos accidentally included in the denominator, expansion accidentally added to the numerator, mid-period downgrades counted twice. The six cards below walk through the full calculation, the way a finance team would check it before it ships to the board.
Take the ARR of the customers you had on day one of the period. Subtract contraction (downgrades, seat reductions, plan cuts) that happened inside that cohort. Subtract churn (full cancellations) that happened inside that cohort. The result is the retained ARR from the starting base, and nothing else.
The denominator is always the ARR of the starting cohort. It does not change when new logos arrive mid-period, because new logos are not in the cohort being measured. Keeping the denominator stable is what makes GRR comparable period over period.
Expansion revenue from the existing base is not included, which is the single rule that separates GRR from NRR. New logos acquired after day one are also not included. The cohort is frozen at the start of the period, which is why GRR is sometimes described as a cohort retention metric.
Annual GRR is the standard board number. Quarterly GRR is useful for trend watching inside the year. Monthly GRR is noisy for most businesses but can be useful for product-led companies with high transaction frequency. Pick one window, publish the methodology, and keep it consistent across every report.
Starting ARR of one hundred million. Contraction of two million in downgrades. Churn of six million in cancellations. GRR equals one hundred minus two minus six, divided by one hundred, times a hundred, which equals ninety-two percent. The eight-million-dollar gap is the real retention problem, regardless of how much the remaining base expanded.
If your GRR number is higher than your NRR number, the math is wrong. NRR is GRR plus expansion from the same cohort, so NRR must be greater than or equal to GRR by definition. A GRR above NRR means expansion got double-counted, or churn was moved out of the cohort incorrectly. Flag it and recalculate.
GRR and NRR share the same cohort and the same window. They differ on one question: is expansion from the existing base allowed in the numerator. That one difference splits them into two completely different signals, and the gap between them is one of the most useful diagnostics in revenue reporting.
Both GRR and NRR are cohort metrics. Both start with the ARR of the customers you had on day one of the period. Neither includes new logos. If a team reports a retention number that includes new business, it is not GRR or NRR, it is something else that happens to use the word retention.
NRR takes the same starting cohort, subtracts contraction and churn, and then adds expansion revenue (upsell, cross-sell, seat growth, usage overages) from the surviving customers. GRR stops before that final step. That single addition is why NRR can exceed one hundred percent and GRR cannot.
Subtract GRR from NRR and you have the expansion rate from the base, as a percentage of starting ARR. A ninety-two percent GRR and a one hundred ten percent NRR means expansion added eighteen points. A ninety-five percent GRR and a ninety-eight percent NRR means expansion is only adding three points, which is a much thinner cushion.
An NRR of one hundred twenty percent sounds healthy. If GRR is seventy-eight percent, the business is actually losing more than one in five revenue dollars from the base and papering over it with a few large expansion deals. That pattern breaks the moment a top account expands less, which is why investors scrutinize GRR first.
A GRR of ninety-six percent is excellent on its own, but if NRR is only ninety-seven percent, the product has almost no expansion motion. That often shows up in flat-rate pricing or single-seat software. The base is sticky, but every new dollar has to come from new logos, which raises CAC pressure.
GRR tells you how durable the base is before growth tactics. NRR tells you how much the surviving base grew on its own. Reading them together is how a board understands whether a strong top-line number is coming from a healthy base, a lucky expansion deal, or a sales team covering for a retention gap.
Benchmarks move with segment, deal size, and contract length. The numbers below are the ones most commonly cited by SaaS investors and public filings, with the caveat that any benchmark is only useful once the methodology behind it is clear. The last three cards cover the mistakes that quietly inflate GRR in internal reports.
A ninety percent or higher GRR is the standard definition of a healthy SaaS business. The remaining ten percent is the normal cost of doing business: customers that outgrow the product, go out of business, or discover the use case was not a fit. Below ninety percent, retention becomes a strategic problem rather than a tactical one.
Enterprise-focused companies routinely report GRR of ninety-five percent or higher. Annual or multi-year contracts, procurement friction, and deeper integrations all slow churn. Investors expect this profile from any company selling to enterprise, and a sub-ninety number in an enterprise book is a direct red flag.
A GRR below eighty-five percent is the industry threshold for a retention problem. The business is losing more than fifteen percent of its starting ARR every year, which means it has to grow new business by that much just to stand still. At that level, even a very strong NRR is a signal that expansion is masking an underlying product or fit issue.
The most common error is dropping new logos acquired during the period into the denominator or the numerator. That turns GRR into a hybrid growth metric and inflates the number. The fix is simple: freeze the cohort at the start of the period and ignore anything that joins after day one.
Teams sometimes accidentally let expansion from the surviving base leak into the numerator, which is NRR math applied to a GRR number. The result is a GRR above one hundred percent, which is mathematically impossible. If GRR exceeds one hundred, the calculation has NRR components inside it and needs to be redone.
Switching between monthly, quarterly, and annual GRR mid-report creates numbers that cannot be compared. Pick annualized GRR for the board report, keep the methodology on the slide, and resist the temptation to switch to a shorter window just because the number looks better that quarter.
GRR is only as honest as the data behind the ARR. If contracts, renewals, downgrades, and cancellations live in different tools, the number is reconstructed in a spreadsheet every quarter and no two reports agree. A CRM that owns the full revenue motion is where GRR becomes a live metric rather than a quarterly archaeology project.
Strkr tracks ARR as a first-class field on every account, calculated from active subscriptions, seat counts, and contract terms. The starting-period ARR for every cohort is captured automatically, so the GRR denominator is never a question in a renewal review.
Seat reductions, plan downgrades, and negotiated price concessions are logged as contraction events against the account, with the dollar impact and the effective date. The GRR report sums contraction automatically for any cohort window instead of asking finance to reconstruct it from Stripe exports.
When an account cancels, Strkr captures the cancellation date, the lost ARR, and a reason code (price, product fit, acquisition, no engagement). The churn total feeds GRR, and the reason codes feed the retention program, so the metric and the response are built from the same data.
The retention dashboard breaks GRR down by segment (SMB, mid-market, enterprise), by product line, by CSM or account owner, and by acquisition cohort. One number for the board, dozens of slices for the operators who have to actually move the number up.
GRR and NRR are rendered side by side on the retention dashboard, over the same window, from the same cohort. The gap between the two is calculated automatically, so the expansion contribution is visible without a side spreadsheet. Boards get both numbers, consistently, without a quarterly scramble.
Strkr flags accounts whose usage, engagement, or support signals point toward downgrade or churn, with a window long enough to actually intervene. The GRR number is the lagging indicator. The at-risk list is the leading indicator, and it is where the retention program actually earns its budget.
Strkr tracks ARR, contraction, churn, and expansion against the same account record, so retention metrics are a live dashboard instead of a quarterly archaeology project. Pricing is published. The feature pages show exactly what ships today.
GRR stands for gross revenue retention. It is the share of recurring revenue a company keeps from its existing customers over a given period, excluding any expansion. GRR is one of the two primary retention metrics investors and boards track, usually reported alongside NRR (net revenue retention).
GRR equals starting ARR minus contraction minus churn, divided by starting ARR, times one hundred. Starting ARR is the recurring revenue of the customers you had at the start of the period. Contraction is revenue lost from downgrades. Churn is revenue lost from cancellations. Expansion from the existing base is deliberately excluded.
A GRR of ninety percent or higher is considered healthy for most SaaS businesses. Enterprise-focused companies typically run ninety-five percent or higher. A GRR below eighty-five percent is a signal of a retention problem that no amount of expansion revenue can safely paper over.
GRR and NRR both measure retention of the starting cohort, over the same window. GRR excludes expansion from the existing base, so it caps at one hundred percent and shows pure stickiness. NRR includes that expansion in the numerator, so it can exceed one hundred percent. The gap between the two is the expansion rate from the base.
No. GRR is capped at one hundred percent by definition, because expansion revenue is not included in the numerator. If a GRR calculation returns a number above one hundred, expansion has leaked into the math and the formula is being applied incorrectly. In that case, the result is actually NRR, not GRR.
Because NRR can hide churn behind expansion. A company can lose fifteen percent of its base and still report an NRR above one hundred percent if a few large customers expand enough to cover it. GRR strips expansion out of the picture, which is why investors read it first to judge how durable the base is on its own.
Three mistakes recur. First, including new logos acquired during the period in the cohort, which turns GRR into a hybrid growth metric. Second, allowing expansion from the existing base into the numerator, which produces GRR numbers above one hundred percent. Third, switching the measurement window between reports so period-over-period comparisons break.
A CRM that holds ARR, contract terms, contraction events, and cancellation records against each account can calculate GRR automatically for any cohort, segment, or owner. Without a shared record, GRR gets reconstructed in a quarterly spreadsheet from exports, and no two reports agree. Strkr tracks GRR and NRR side by side on the retention dashboard, with the gap calculated for you.
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.
