Gross Revenue Retention Benchmarks 2026: Why the SaaS Median Just Fell to 84%

Every SaaS founder has a churn number memorized. Fewer track gross revenue retention (GRR) — the metric that strips out expansion revenue and asks a blunter question: of the revenue you had at the start of the year, how much is still there? In 2026, the answer got noticeably worse across the board, and it's worth understanding why before you assume your own numbers are just an execution problem.

What gross revenue retention actually measures

GRR is the percentage of recurring revenue you kept from existing customers over a period, counting only cancellations and downgrades against you. It ignores upgrades entirely, which is exactly what makes it useful: it can never exceed 100%, and it can't be flattered by a handful of accounts expanding while the rest of your base quietly leaks.

Compare that to net revenue retention, which adds expansion MRR back into the numerator and can climb above 100%. A company can post 110% NRR while its GRR sits at 82% — that gap is expansion revenue from a shrinking pool of accounts covering for the ones walking out the door. Boards that only look at NRR miss this. GRR is the metric that can't lie to you about how many customers you're actually keeping.

The 2026 benchmark shift: median GRR fell to 84%

The headline number, from the Aleph × Benchmarkit 2026 SaaS & AI Performance Benchmarks report published in June 2026, is a four-point drop in median GRR — from 88% to 84% — across 342 B2B SaaS and AI-native companies reporting full-year 2025 actuals. Top-quartile performance moved too: the 75th percentile slid from 95% to 91%. Nobody was spared, including the companies that were already doing this well.

Source: Aleph × Benchmarkit, 2026 SaaS & AI Performance Benchmarks (June 2026).

That "revenue impact" row isn't abstract. A four-point GRR drop on a $10M ARR base means roughly $400K a year in revenue you'd otherwise have kept just evaporates into churn and contraction. At $50M ARR, that's a $2M hole showing up every single year, recurring, with no new deals required to make it worse.

This isn't a one-year blip. It's a continuation of a slide that's been running since 2022, and the 2026 data is the sharpest single-year drop in that series. If your own GRR has softened over the past two years, you're not necessarily doing anything new wrong — the floor itself has been moving.

Why GRR is falling industry-wide

Three things showed up consistently in the 2026 data as drivers, and none of them are about product quality declining.

Shorter contracts, less lock-in

Multi-year deals have gotten harder to sell. Buyers don't want to commit to a three-year contract with a tool that might be irrelevant — or outcompeted by something AI-native — in twelve months. That shift toward annual or even monthly terms removes the structural churn suppression that long contracts provide. A subscriber on a three-year deal simply can't churn for another two years regardless of satisfaction; a subscriber on a monthly plan can churn the day their invoice hits.

Usage-based pricing cuts both ways

Usage-based and consumption pricing models are more honest about value delivered, but they also mean revenue can shrink without a formal cancellation. A customer who dials down usage 40% shows up as contraction in your GRR even though they never touched the "cancel" button. As more of the market moves to consumption pricing, more revenue erosion happens quietly through usage decline rather than an explicit churn event you can intervene on.

Cost cuts outpaced retention investment

Across the same 2026 dataset, the median company cut R&D spend by roughly 8 points of revenue and G&A by roughly 7 points, with S&M down about 2 points — a broad move toward efficiency after several years of pressure to show profitability. GRR fell 4 points over the same window. Companies got leaner faster than they protected the revenue base that leanness was supposed to preserve. Retention work — cancellation flows, dunning, proactive customer success outreach — often sits inside the budgets that got cut first.

The AI-native GRR crisis

The most striking number in the 2026 report isn't the industry median — it's the AI-native segment. AI-native software companies posted a median GRR of just 40%, against 48% NRR, compared to an 82% NRR median for traditional B2B SaaS. Budget AI tools priced under $50/month fared even worse, with median GRR around 23%.

That gap tells you something concrete about the category: AI-native products are being adopted fast and abandoned fast. Low switching costs, commoditized model access, and a wave of near-identical competitors mean customers churn the moment a marginally better tool ships. If you're building in or adjacent to AI tooling, the retention playbook that works for traditional SaaS — annual contracts, seat-based expansion, long onboarding — doesn't map cleanly onto a market this volatile. Cancellation-moment intervention matters even more here, because there's often no long contract buying you time to fix the underlying problem.

GRR benchmarks by ACV segment

Source: GrowthSpree, B2B SaaS NRR and GRR Benchmarks 2026 report.

The pattern is the same one that shows up in every retention metric: ACV and GRR move together. Higher-priced contracts come with procurement processes, integrations, and switching costs that suppress churn structurally. If you're SMB-priced, a 75% GRR isn't a red flag on its own — it's roughly where the segment sits. What matters more is your trend line and how you compare to segment peers, not the enterprise benchmark you'll never hit at a $50/month price point.

What actually moves GRR

GRR only has two levers: reduce cancellations, and reduce downgrades. Everything else is noise.

On cancellations: the highest-leverage intervention is still what happens at the moment someone tries to leave. A cancellation flow that asks why and responds with a relevant offer — pause, discount, or downgrade instead of outright cancellation — typically saves 25–40% of subscribers who would otherwise have churned outright. Every save is direct, immediate GRR.

On involuntary churn: failed payments account for 20–40% of total cancellations in most SaaS businesses, and almost all of that is recoverable with the right dunning setup — extended retry windows, pre-expiry card reminders, and in-app payment banners. Because this churn has nothing to do with product dissatisfaction, it's the cheapest GRR you'll ever recover.

On downgrades: track them separately from cancellations. A customer moving to a cheaper tier is telling you the same thing a "too expensive" cancel reason tells you — the value-to-price ratio broke somewhere. Segment your downgrades by plan and by tenure the same way you'd segment customer retention rate, and the pattern usually points at a specific tier or cohort rather than the whole business.

You can model the underlying MRR components — starting, expansion, contraction, churned — with our retention rate calculator; set expansion to zero and what comes out the other end is your GRR.

The market-wide drop in GRR this year is a useful reality check, but it doesn't excuse a below-benchmark number. The companies still posting 90%+ GRR in this environment are, almost without exception, the ones that treat the cancellation moment as a product surface worth investing in rather than a page to get customers through as fast as possible. CancelFlow adds that surface to any Stripe-based SaaS with a script tag and a function call — no backend rebuild, and the cancel-reason data alone tells you which of the three levers above is actually costing you revenue.

Frequently asked questions

Stop losing subscribers today

One script tag. One function call. A live cancellation flow in under 10 minutes.