Net Revenue Retention Benchmarks 2026: What good Actually Looks Like

Net revenue retention benchmarks have quietly become the single number every SaaS board, investor, and operator argues about — and the median has been drifting down. Across private B2B SaaS, median NRR fell from roughly 105% in 2021 to about 101% in 2024, according to Benchmarkit and Maxio survey data. That compression is real, but the median is also the least useful figure in the entire dataset.

The reason: a single blended NRR number averages together two fundamentally different businesses. Enterprise SaaS selling at over $100K average contract value retains near 118% on a median basis, while SMB SaaS under $25K ACV sits at 97% — a 21-point spread that no headline benchmark exposes. A team targeting small businesses and measuring itself against "110% is good" is chasing a standard built for a different company.

This guide does three things. It defines NRR and GRR precisely and shows the expansion-vs-churn math. It breaks the benchmarks down by ACV segment, ARR stage, and pricing model in a single reference matrix. And it turns the data into a lever playbook plus a target-setting framework so you can decide what good means for your business, not someone else's. Every figure below is sourced and dated; where a number is vendor-stated or volatile, we say so.

Net revenue retention measures how much recurring revenue you keep and grow from your existing customer base over a period, ignoring any new logos. The formula is straightforward: NRR = (Beginning ARR − Churn ARR − Contraction ARR + Expansion ARR) / Beginning ARR. Because expansion is added back in, NRR is uncapped — a company whose customers upgrade and grow can exceed 100% with zero new sales.

Gross revenue retention is the same calculation without the expansion term: GRR = (Beginning ARR − Churn ARR − Contraction ARR) / Beginning ARR. GRR is capped at 100% and measures pure leakage — it tells you how much of last year's revenue survives before any upsell. The gap between NRR and GRR equals expansion ARR divided by beginning ARR, so reading both together separates two distinct questions: are customers leaving (GRR), and are the ones who stay growing (the gap)?

The headline trend in private SaaS retention is mild but real and consistent across surveys. Benchmarkit and Maxio data show median NRR sliding from roughly 105% in 2021 to about 101% in 2024. Gross retention has moved the same direction: median GRR fell to about 88% in 2024 from 90% in 2022. KeyBanc Capital Markets' 16th annual private SaaS survey similarly reported GRR dipping to around 86% in 2023, with NRR staying above 100% through 2025.

Two forces explain most of it. The first is post-pandemic SaaS rationalization — buyers consolidating tools and scrutinizing seat counts after the 2020–2021 over-purchasing boom. The second, newer pressure is AI substitution: low-switching-cost software is increasingly displaced by AI-native alternatives or absorbed into broader platforms. Benchmarkit's read on the GRR slide is blunt.

Here is the original-analysis caveat, though: a four-point drop in the median over four years is not a crisis, it is a recalibration. What changed more than the average is the dispersion. The top of the market kept compounding — public companies above 120% NRR held their premium — while the bottom decile slid further. The danger for operators is not the median moving; it is mistaking a borrowed median for a target when your segment's reality is 20 points away in either direction.

This is the part most NRR roundups skip. Average contract value predicts go-to-market structure and expansion ceiling better than ARR or company age, and once you split the benchmark by ACV the single median falls apart. SaaS Capital's retention research, cross-referenced with aggregator data covering roughly 939 B2B SaaS companies, shows enterprise (over $100K ACV) at a median 118% NRR, mid-market ($25K–$100K ACV) at about 108%, and SMB (under $25K ACV) at 97%.

That last figure is the one that reframes everything. An SMB-focused SaaS at 97% median NRR is not failing — it is exactly at benchmark for its segment. SMB customers churn more, expand less, and run on self-serve motions with thin account coverage. Holding 100%+ at SMB scale is genuinely strong; 120% there is exceptional. At enterprise scale, by contrast, 120% is merely "best" and the top performers reach 130–135%.

To find your real benchmark, answer two questions: what is your ACV tier, and what is your pricing model? The matrix below combines ACV segmentation, gross-retention floors, the Good / Better / Best thresholds, and the dominant expansion lever for each band into a single view. No published roundup we found combines these axes in one table — this is the reference an operator can locate themselves in and instantly see what good looks like for their cell.

The pricing-model rows carry their own lesson. m3ter's benchmarks show usage-based and hybrid models landing in the 115–130%+ range versus 95–105% for flat-rate subscriptions — a structural advantage of roughly 20–25 percentage points, because consumption revenue grows automatically as a customer derives more value. If you are weighing a pricing change, our deeper breakdown of usage-based pricing models walks through when the switch pays off.

The most strategically important shift in the data is not the NRR median at all — it is where growth comes from. Maxio's 2025 benchmark report found median expansion ARR rising from about 25% of total new ARR in 2022 to 40% in 2024. Benchmarkit's data takes it further at scale: at the $50–$100M ARR band expansion accounts for roughly 58% of new ARR, and above $100M ARR it reaches about 67%.

Read that progression forward and the implication is structural. Below roughly $20M ARR, new-logo acquisition is still the primary engine and expansion is a supplement. Somewhere around that threshold the two cross over, and at scale expansion becomes the dominant growth motion. High Alpha's 2025 analysis found companies scaling from $1M to $20M ARR lifted NRR by about 12 percentage points along the way — partly product maturation, partly a customer mix that shifts toward stickier mid-market accounts. The companies that under-invest in customer success and expansion tooling during that climb hit a ceiling exactly when expansion is supposed to take over.

The compounding math makes the stakes concrete. A company holding 120% NRR with zero new customer acquisition grows a $10M ARR base to roughly $24.9M in five years on expansion alone — $12M, $14.4M, $17.3M, $20.7M, $24.9M, year by year. That is the entire argument for treating retention as a growth strategy rather than a defensive one. ChartMogul's data echoes it from the other direction: companies at 100%+ NRR grow at a median 48% year over year, roughly double the ~24% of those below 100%.

The most alarming and underreported data point of late 2025 is what is happening to AI-native software. ChartMogul's data put the median NRR for AI-native SaaS products near 48%, with median GRR around 40% — figures dramatically below conventional SaaS. The number had improved through the year (GRR rose from about 27% in January 2025 to 40% by September), but the gap versus traditional SaaS remains enormous.

The diagnosis is not "AI software is bad" — it is a price and lock-in problem. ChartMogul's price-tier breakdown is stark: AI-native plans above $250/month retain at about 85% NRR, $50–$249/month plans at 61%, and sub-$50/month plans at just 32%. Cheap, novelty-driven AI tools are highly substitutable, and price-sensitive users churn the moment a better or free alternative appears. Higher-priced AI tools woven into enterprise workflows retain far better because they accumulate switching costs.

The fix for an AI-native product is the same lever playbook applied to a harder problem: raise the price floor, build genuine workflow lock-in, and move customers onto annual billing. The data should also be read with care — these figures come from ChartMogul's own subscriber base, not a random sample of all AI companies, so they describe a real but specific population. Either way, the directional signal is clear, and identifying at-risk accounts early matters more here than anywhere; modern churn prediction models can surface the warning signs 60–90 days before a renewal lapses.

Benchmarks tell you where you stand; levers tell you what to pull. The table below quantifies six of the highest-leverage moves in percentage points of NRR improvement, with the evidence base, implementation effort, best-fit segment, and rough time-to-impact so you can prioritize. Some figures are vendor-stated — noted in the table — and should be treated as directional rather than guaranteed.

*The CS/AM role-separation figure is vendor-stated (Gainsight 2023 State of Customer Success) and reflects a customer-success platform vendor's own data — treat as directional. Billing-cadence and pricing-model deltas are from ChartMogul and m3ter benchmark data; involuntary-churn recovery is from m3ter.

The sequencing matters as much as the levers. The two lowest-effort, fastest-impact moves — fixing involuntary churn through smart dunning and nudging monthly customers onto annual plans — are where most teams should start, because they recover revenue already on the books without re-architecting the product. The pricing-model switch has the largest ceiling but the longest runway and the most execution risk. For the leaky end of the funnel, pairing these with disciplined customer retention automation — automated health scoring, renewal triggers, and at-risk alerts — is what makes the gains stick.

NRR is not only an operating metric; it is one of the cleanest predictors of enterprise value. Software Equity Group's analysis of public SaaS companies found that the roughly 16.5% of index companies with NRR above 120% traded at a median 9.3x EV/TTM revenue — about a 63% premium to the index median. Companies below 100% NRR carried just 3.1x, a 46% discount. The same study found that 72% of index companies reporting NRR exceeded 100%, and that more than 80% of the above-120% group traded above the broader index median.

The mechanism is intuitive. Investors pay for predictable, compounding revenue, and a high NRR is the purest signal that the existing base will grow on its own. The same logic shows up in the private market: SaaS Capital's research found that moving NRR from the 90–100% band into 100–110% improves growth rate by about 5 percentage points, and that companies at the highest NRR levels grow roughly 83% faster than the population median. Consumption-heavy and land-and-expand public names — the kind operating in DevOps, security, and data infrastructure — routinely report NRR in the 113–125% range, though those specific figures move every quarter and should be checked against the latest earnings before being cited.

OpenView and High Alpha's benchmark teams have been explicit that these two variables — not growth rate in isolation — are the real leading indicators.

With the data in hand, the target-setting move is mechanical. Start from your ACV tier's median in the matrix above, not the blended benchmark. Adjust for your pricing model — add roughly 20 points of headroom if you run usage-based, subtract if you are flat-rate monthly. Then sanity-check against your GRR floor: if your gross retention is below the segment floor, your NRR target is a fantasy until the leak is fixed, because expansion cannot durably outrun churn.

For a venture-backed B2B company, the relevant comparator is the roughly 106% venture-backed median, not the much lower 82% figure sometimes quoted — that broader number folds in early-stage products without product-market fit and is not a meaningful benchmark for a funded business. The same care applies to the Good / Better / Best scale: 120% is exceptional at SMB scale and merely "best" at enterprise scale. A realistic target is your segment median plus roughly 8–12 points of achievable lift from the two or three levers you can actually pull this year.

None of this lives in a spreadsheet alone. NRR is a downstream signal of how well your marketing, onboarding, and customer programs compound value over time — which is why we treat it alongside measuring marketing ROI rather than as a finance-only number. If you want a retention and expansion program built on this data rather than borrowed benchmarks, our CRM and lifecycle automation engagements start with exactly this kind of segmentation and lever analysis.