Gross Revenue Retention Exposes Underlying Subscription Churn That Net Expansion Metrics Routinely Conceal

Inside a venture capital due diligence room, a software founder projects a slide showing a 120% Net Revenue Retention rate. The number suggests a compounding growth engine, where existing customers buy more software every year without the company spending a dollar on new acquisition.

But when the buyer's analysts strip away the expansion revenue to look at the underlying base, they find a Gross Revenue Retention rate of 75%. The company is not a compounding engine; it is a leaking bucket, masking a 25% annual revenue loss by aggressively upselling a shrinking pool of survivors.

This mathematical illusion happens every day in subscription software. Net Revenue Retention (NRR) captures the headline growth of a customer base, adding in the upgrades, cross-sells, and price increases. Gross Revenue Retention (GRR) strips all of that away.[3]

GRR answers one clean, unforgiving question: how much of last year's recurring revenue survived without counting a single upgrade. Because it excludes expansion, GRR can never exceed 100%. It is the purest measure of product stickiness and customer satisfaction.

The gap between the two metrics is where the risk hides. A wide spread means that expansion is covering for churn, not building on a stable foundation. If that expansion is concentrated in a few large accounts, the entire revenue base is fragile.[1]

The Mechanics of the Metrics

To calculate GRR, a finance team takes the starting Annual Recurring Revenue (ARR) of a customer cohort, subtracts the revenue lost to cancellations and downgrades, and divides by the starting ARR. If a company starts with $10 million and loses $1.6 million to churn, its GRR is 84%.

NRR takes that same $10 million starting base, subtracts the $1.6 million in churn, but then adds back any expansion revenue from the surviving customers. If the remaining accounts buy $1.8 million in additional licenses, the ending ARR is $10.2 million, yielding an NRR of 102%.

The headline number looks healthy, showing a net gain. But the underlying reality is that 16% of the original revenue walked out the door. The company had to sell nearly $2 million in upgrades just to tread water.[3]

"I know all the major growth investors heavily scrutinize retention as part of their investment thesis," Nick Mehta, former Gainsight chief executive, noted in a recent industry interview.

"Every PE sponsor is using it as a big part of their diligence, trying to understand not just the high-level metrics, but the cohorts," Mehta added. The scrutiny intensifies as market conditions tighten.

The Compressing Benchmarks

Across the software industry, retention rates have steadily compressed since the pandemic-era purchasing boom. Buyers are consolidating tools, scrutinizing seat counts, and cutting discretionary software budgets to protect their own margins.

The median B2B SaaS company posted a GRR of 84% in 2025, down four points from the previous year, according to the 2026 Aleph and Benchmarkit performance survey of 342 companies. That single-year drop represents the most alarming benchmark in the dataset.

At 84% GRR, the typical software company loses 16% of its existing ARR every year to churn and contraction before it adds a dollar of new business. Meanwhile, the median NRR for the same group sat at 102%.

That creates an 18-point spread that expansion teams must constantly fill. Different market segments carry different expectations, making blended medians dangerous for benchmarking. Enterprise software contracts, which often exceed $100,000 annually, typically see GRR above 90% due to high switching costs.

Small business software, with lower price points and higher failure rates, often sees GRR between 70% and 80%. Bessemer Venture Partners' widely cited framework puts 100% NRR at "good," 110% at "better," and 120% or higher at "best."

However, that scale was designed for growth-stage enterprise platforms. Applying it to mid-market or bootstrapped companies often warps expectations, pushing sales teams to chase expansion before the core product is fully adopted.

The Valuation Premium

In mergers and acquisitions, buyers do not ask which metric looks better; they ask which metric explains the risk. Founders tend to lead pitch decks with NRR because the number is larger and tells a compelling growth story.

Buyers start their diligence with GRR because it is harder to dress up. A 2021 analysis of 40 public B2B SaaS companies by McKinsey found that firms with an NRR of 120% or higher traded at a median enterprise value to revenue multiple of about 21x.

Companies below that threshold traded at roughly 9x. But that valuation premium only holds if the underlying GRR is solid. Two companies can both report 105% NRR, yet present entirely different risk profiles to an acquirer.

The business with 94% GRR is much easier to underwrite than the one with 78% GRR. Its growth is broad and repeatable, rather than reliant on a few massive upsells that could vanish in a single renewal cycle.

"High NRR creates valuation upside only when GRR proves the base is durable," notes a 2026 SaaS Capital research brief. The firm's survey of over 1,000 private companies found that the cleanest businesses maintain a high GRR floor alongside a credible expansion engine.

The False Positive of Expansion

When NRR exceeds 110% but GRR falls below 85%, the business is generating a false positive. This dynamic often occurs in usage-based pricing models, where a small cohort of power users dramatically increases their consumption while the majority of the customer base churns out.[1]

This concentration risk becomes glaringly obvious under cohort-level analysis. If a software provider relies on its top 5% of customers for 80% of its expansion revenue, the loss of a single major account can instantly collapse the NRR metric.[3]

Stabilizing the base retention must precede scaling the expansion motion. Pushing customer success teams to aggressively upsell accounts that are already struggling to find value often accelerates their departure, burning through the addressable market.[1]

Leading firms are increasingly tying executive compensation and customer success targets to GRR rather than NRR. This forces teams to focus on core product adoption, onboarding quality, and daily active usage, rather than just hunting for the next cross-sell opportunity.[3]

A mysterious leak in recurring revenue is rarely solved by pouring more water into the top of the bucket. Fixing GRR requires structural changes to the product, the ideal customer profile, or the implementation process.

The Shift to Durable Growth

The software market's transition from a growth-at-all-costs mindset to a focus on capital efficiency has elevated GRR to the top of the diligence checklist. Investors now demand to see a retention bridge that breaks down beginning ARR, churn, contraction, and expansion separately.

This level of transparency prevents the math from hiding the reality. A retention bridge shows exactly how many dollars were lost to outright cancellations versus partial downgrades, allowing operators to pinpoint where the product is failing to deliver value.[3]

ServiceNow, a major enterprise software provider, explicitly details its attrition rate in its financial filings. The company calculates its renewal rate based on Annual Contract Value lost during the period, excluding changes in price or added users.[2]

This level of public disclosure remains rare. Many publicly traded cloud companies blend their metrics, reporting only a "Dollar-Based Net Expansion Rate" that obscures the underlying churn and leaves investors guessing how much of the base is actually staying.[2]

When a company announces a shift to a subscription model, revenue often dips initially as large upfront payments are replaced by smaller recurring ones. Tracking GRR becomes critical to ensure that the new recurring customers are sticking around past their first renewal.[3]

As artificial intelligence tools begin to displace traditional, low-switching-cost software, maintaining high gross retention will become even more difficult. Products that merely act as workflow wrappers are already seeing their GRR erode as buyers consolidate vendors.

The most successful software companies eventually achieve a balance. They build a product sticky enough to maintain a 90% GRR, while deploying a pricing model that naturally captures more value as the customer grows, driving NRR comfortably past 110%.

Net revenue retention remains the headline indicator of a compounding business model, but it cannot function as a standalone health score. Gross revenue retention provides the structural foundation that determines whether that compounding growth can actually survive the next renewal cycle.[3]