.png)
Frameworks, core principles and top case studies for SaaS pricing, learnt and refined over 28+ years of SaaS-monetization experience.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
What Is Gross Retention Rate Understanding This Critical SaaS Metric
A SaaS company can post strong new bookings, grow ARR, and report net revenue retention above 100% while still losing ground with its installed base. Expansion from a few large accounts can conceal a more serious problem: customers are leaving, reducing scope, or cutting users because the product and the price no longer feel justified.
Gross retention rate, or GRR, is the metric that exposes that problem. It strips out upsell and cross-sell activity to show how much recurring revenue a company keeps from the customers it already had. Monetizely's position is clear: GRR should be treated as the hard floor beneath SaaS growth and as a guardrail for pricing decisions. A pricing change that improves bookings while weakening underlying retention is not a successful pricing change.
GRR measures the recurring revenue retained from a starting customer cohort after churn and contraction, while excluding expansion. In practical terms, it answers a direct operating question: of every dollar customers paid us at the start of the period, how many cents remain before we count new products, more seats, higher usage, or price increases?
A rigorous calculation should use a constant starting cohort and keep expansion separate. Otherwise, a customer that doubles spend on one module can hide another customer that cuts half its deployment. The basic formula is:
[ \text{GRR} = \frac{\text{Beginning cohort ARR} - \text{Churned ARR} - \text{Contracted ARR}}{\text{Beginning cohort ARR}} \times 100 ]
The following example shows why GRR and net revenue retention, or NRR, need to be read together.
| Exhibit 1: Formula and worked example | Gross Retention Rate | Net Revenue Retention Rate |
|---|---|---|
| Beginning ARR from the same customer cohort | $1,000,000 | $1,000,000 |
| ARR lost when customers leave | ($80,000) | ($80,000) |
| ARR lost through downgrades, fewer seats, or lower usage | ($40,000) | ($40,000) |
| ARR gained from upsell, cross-sell, or expansion | Excluded | $180,000 |
| Ending ARR counted in the metric | $880,000 | $1,060,000 |
| Calculated rate | 88% | 106% |
The company in this example can truthfully report 106% NRR. Yet it has also lost 12% of its starting revenue base before expansion. GRR therefore shows the durability of the core customer promise, while NRR shows the combined effect of durability and growth inside existing accounts.
GRR cannot exceed 100% when calculated correctly. Any result above 100% means that expansion, price increases, or a changing cohort has entered the numerator.
Operators often use “retention” as if it were a single measure. It is not. Logo retention, GRR, and NRR each reveal a different failure mode, and pricing teams need all three.
The table makes the management task clear: NRR tells us whether the installed base is growing, but GRR tells us whether its foundation is sound. A company with high NRR and falling GRR may still grow for a period. It is also building dependence on a narrower group of expanding accounts.
The danger is most acute when a business has several large customers with room to expand. Assume ten enterprise accounts generate $5 million of starting ARR. Two accounts add $900,000 in new modules, while smaller accounts churn or reduce spend by $500,000. NRR is 108%; strict GRR is 90%.
That result calls for two different conversations. The growth conversation asks how to repeat the $900,000 expansion. The retention conversation asks why $500,000 of the original value proposition no longer held. Treating those as one conversation is how pricing teams mistake account concentration for customer health.
The diagnostic matrix below converts the two metrics into decisions rather than dashboard decoration.
| Exhibit 3: The GRR-NRR pattern reveals the real pricing problem | High NRR | Low NRR |
|---|---|---|
| High GRR | The base is durable and account expansion is working. Test whether higher-value packages, additional products, or measured price increases can capture more value. | Customers stay, but spend is not growing. The problem is likely adoption, packaging, or limited room to expand. |
| Low GRR | Expansion is masking leakage. Freeze broad price changes and examine which segments are downgrading, churning, or negotiating down. | The offer is failing both as a retention engine and a growth engine. Revisit segment fit, package design, and the price metric before changing rates. |
A falling GRR with a strong NRR is the most dangerous pattern because it looks good in aggregate. A CFO sees account growth. A customer success leader sees more save plays. A product team sees rising usage among the largest customers. The company may still be losing relevance with the middle of its market.
Public disclosures show why definitions matter as much as the reported percentage. Vertex reported 94% GRR for the year ended December 31, 2025, and its definition explicitly removes revenue lost through departing customers, downgrades, and lower usage. Procore reported 95% GRR as of December 31, 2025, but its disclosed measure excludes both expansion and contraction, making it closer to a cancellation-focused retention measure.
No universal GRR benchmark exists because contract length, customer mix, product criticality, and company definitions differ. A payroll platform embedded in daily operations should not be judged by the same level as a discretionary analytics tool selling to small teams.
Still, public-company disclosures offer useful reference points when read with discipline. The range below is not a league table. It is a practical set of operating reference points drawn from companies that disclose their measures and methods.
| Exhibit 4: Public-company GRR reference range | Reported GRR | Measurement period | What the disclosure tells us |
|---|---|---|---|
| Blackbaud | Approximately 92% | FY ended December 31, 2025 | Contracted ARR retained relative to beginning CARR; management noted the sale of EVERFI affected the comparison. (sec.gov) |
| Vertex | 94% | FY ended December 31, 2025 | Includes lost customers, downgrades, and reduced usage; excludes expansion and price changes. (sec.gov) |
| Procore | 95% | As of December 31, 2025 | Reflects customer cancellations, but excludes contraction, so it is not directly comparable with a strict GRR measure. (sec.gov) |
| Workiva | 97.2% | As of December 31, 2025 | Measures annualized subscription and support revenue retained from prior-year active customers. (sec.gov) |
The central lesson is not that every SaaS company should target 97%. It is that a company must know what its number includes before it celebrates it. Workiva’s 97.2% measure, Vertex’s 94% measure, Procore’s 95% cancellation-focused measure, and Blackbaud’s approximately 92% measure all provide valuable signals, but they do not carry identical economic meaning.
Our view is that a company should set an internal threshold by segment rather than adopt a generic market target. Enterprise customers on three-year contracts may justify a 95% or higher strict GRR goal. A newer self-service product serving smaller accounts may begin lower, but it should still show improving retention as onboarding, product fit, and packaging mature.
Pricing influences GRR through more than a list-price increase. A company can damage retention by reducing included usage, moving a popular feature into a higher tier, changing an unlimited plan into a capped plan, or selecting a meter customers cannot forecast.
The relevant question is not whether a new model raises first-year ARR. The question is whether customers renew the same value at the new commercial terms.
| Exhibit 5: Each pricing move creates a distinct retention test | Likely GRR risk | Evidence to examine before scaling | Decision rule |
|---|---|---|---|
| Broad list-price increase | Customers renew at lower scope or seek discounts | Constant-price GRR, downgrade rate, renewal discount rate by segment | Expand only if retention holds across the affected cohort |
| Feature moved into a premium tier | Customers retain the platform but cut the feature or package | Feature adoption, package downgrade rate, usage among renewing customers | Keep the feature gated only when the target segment uses and values it |
| Unlimited plan changed to usage-based pricing | Buyers reduce activity to control spend or leave after a surprise bill | Usage distribution, overage incidence, support tickets, renewal objections | Use a clear primary meter and predictable limits before rollout |
| Per-seat pricing applied to uneven users | Customers reduce seat counts while power users retain access | Seat contraction, active-user ratio, utilization by role | Reconsider the seat as the main meter if value does not scale with headcount |
The table means that pricing should be managed as a retention experiment with a financial outcome, not as a rate card exercise. A 10% price increase that causes a 15% reduction in retained scope among a sensitive segment is not a pricing win, even if a few large renewals produce more ARR in the first quarter.
Monetizely's 5-Step Pricing Framework places the pricing metric after Goals & Segments and Positioning & Packaging, but before Rate-Setting and Operationalization. The order matters. Goals & Segments establishes which customers matter and what the company needs pricing to achieve. Positioning & Packaging creates offers those buyers can recognize and buy. Price Metric then determines the unit that makes the bill grow. Rate-Setting puts a number against that unit, and Operationalization makes the model measurable, billable, and governable in daily practice. As discussed in Monetizing Agentic AI, a rate cannot repair a package that does not fit, and a package cannot compensate for a meter customers reject.
GRR belongs most directly in the third step. It should not choose the pricing metric by itself, because a low-churn meter can still undercharge for real value. Yet it is a vital test of whether the chosen unit feels fair over time.
Consider three cases:
A collaboration product charges per seat, but only 20% of licensed users log in each month. At renewal, customers cut 1,000 seats to 650. The pricing issue is not merely adoption; the seat count is overstating the value received by the broader workforce.
A data product charges per event, but customers cannot predict how product launches will affect volume. Finance teams impose usage caps, and the vendor sees reduced activity followed by downgrades. The meter has introduced budget anxiety.
A compliance platform charges by active legal entity, and value rises as customers add entities to the system. Expansion is visible, predictable, and tied to the customer’s business footprint. In that case, the meter can support both healthy NRR and durable GRR.
A named primary meter is essential. Where availability matters and variable use has material cost, a company may add a secondary usage layer, but the primary meter must still be clear enough for customers to budget and for finance to audit. “Credits” or “consumption” are not strategies by themselves. They are billing units that must earn customer acceptance.
Weak definitions turn GRR into a comforting but useless number
Most GRR mistakes do not begin in the spreadsheet. They begin when leadership allows the number to answer more than one question.
Common operator mistakes include:
Counting only full customer churn. A customer that renews at half the prior ARR has still created a major revenue loss, even if the logo remains.
Including price increases in retained revenue. A 7% uplift can make a weak customer cohort look stable. Calculate a constant-price view beside reported revenue retention.
Allowing expansion to offset contraction. That turns GRR into NRR and removes its diagnostic value.
Mixing customer segments. Enterprise accounts, mid-market accounts, and self-service customers often have different buying cycles, support needs, and acceptable meters.
Using only annual renewal data. In-term seat reductions, lower usage, and feature disengagement are often visible months before contract end.
Comparing disclosed metrics without reading the definition. Procore’s disclosed GRR excludes contraction, while Vertex’s disclosed GRR includes downgrades and reduced usage. The percentages should not be treated as interchangeable.
The remedy is not more reporting for its own sake. It is a consistent cohort model that separates churn, contraction, expansion, and price effects. Once those flows are visible, pricing leaders can identify whether a problem comes from poor product fit, package mismatch, discounting pressure, or a pricing metric that no longer fits the way customers use the product.
The next move is to govern GRR as a strategic signal
Adopt one executive definition of strict GRR. Use the same starting cohort, include both churn and contraction, exclude expansion, and publish the definition beside every reported result.
Create a constant-price retention view for every material pricing change. This will show whether retained revenue improved because customers accepted the new value exchange or because the price itself rose.
Set separate GRR targets for each strategic segment. A single company average can hide weakness in self-service customers or dependency on a small group of large accounts.
Require a GRR bridge in major pricing reviews. Show starting ARR, churn, contraction, expansion, and price effects by package and customer segment before approving a broad rollout.
Use the gap between NRR and GRR to allocate leadership attention. A wide gap is not automatically bad, but it signals that expansion may be compensating for losses that deserve direct intervention.
Footnotes and primary sources
Monetizing Agentic AI. https://www.amazon.com/Monetizing-Agentic-AI-Handbook-Transformation/dp/B0H7Z13VKJ/
Monetizely, “Step 1: Goals and Segmentation,” “Step 2: Packaging - Designing Offers That Fit,” “Step 3: Choosing the Right Pricing Metric,” “Step 4: Finding the Right Price Points,” and “Step 5: Operationalizing Agentic AI,” accessed September 3, 2026. (getmonetizely.com)
Workiva, Form 10-K for the year ended December 31, 2025. (sec.gov)
Vertex, Form 10-K for the year ended December 31, 2025. (sec.gov)
Procore Technologies, Form 10-K for the year ended December 31, 2025. (sec.gov)
Blackbaud, Form 10-K for the year ended December 31, 2025. (sec.gov)