SaaS Benchmarks by ARR Tier (2026)
Eight core SaaS operating metrics, segmented by ARR tier rather than funding stage, so the comparison cohort actually matches your scale. Built from KeyBanc, OpenView, Bessemer, and public SaaS filings, verified May 2026.
Why ARR tier beats stage labels
Funding stage is a financing event, not a company. The label tells you what round closed last and roughly when. It does not tell you how big the business is. A $4M ARR business sitting on a $60M Series B has a completely different cost structure, headcount, and investor expectation profile from a $4M ARR business still on its seed cheque. Yet both appear in the same Series B benchmark column.
The major SaaS surveys recognise this. The KeyBanc Capital Markets SaaS Survey segments respondents primarily by ARR bands, because operating metrics correlate far more tightly with revenue scale than with the last round raised. OpenView Expansion SaaS Benchmarks does the same. Both report distributions per ARR tier and only secondarily cut by go-to-market motion or vertical.
The reframe matters when you sit down for a board meeting or fundraise. If you anchor your peer set on funding stage, half of your "peers" are operating at a different scale and will skew the comparison. Anchor on ARR tier and the comparison cohort actually matches the operating reality of running a business your size.
The 2026 benchmark table by ARR tier
Median ranges across the major 2025-2026 surveys. Cells show the interquartile spread (roughly Q1 to Q3), not the absolute extremes. Top decile companies routinely beat the upper bound, bottom decile fall below the lower bound. Use the column closest to your current ARR and read both adjacent columns for context.
| Metric | $1M ARR | $5M ARR | $10M ARR | $25M ARR | $50M+ ARR |
|---|---|---|---|---|---|
| ARR Growth YoY (median) | 180-260% | 90-150% | 60-100% | 45-75% | 30-55% |
| NRR (median) | 98-108% | 102-114% | 104-118% | 106-122% | 108-126% |
| GRR (median) | 82-90% | 86-92% | 88-94% | 90-95% | 91-96% |
| Gross Margin (median) | 62-74% | 66-78% | 70-80% | 72-82% | 74-84% |
| Burn Multiple (median) | 1.6-2.8x | 1.1-1.9x | 0.9-1.6x | 0.7-1.3x | 0.5-1.1x |
| Magic Number (median) | 0.4-0.8 | 0.5-1.0 | 0.6-1.1 | 0.7-1.2 | 0.7-1.3 |
| CAC Payback Months (median) | 15-24 | 13-20 | 12-18 | 11-16 | 10-15 |
| Rule of 40 (median) | 30-55 | 30-55 | 32-58 | 35-60 | 38-62 |
Sources: KeyBanc Capital Markets SaaS Survey 2024 (latest published edition, cohorts at $1-5M, $5-15M, $15-50M, $50M+), OpenView Expansion SaaS Benchmarks, Bessemer State of the Cloud 2026, Iconiq Operating Metrics. Public-comp anchors for $50M+ tier drawn from Bessemer Emerging Cloud Index quarterly composite. All numbers as of May 2026.
ARR Growth YoY (median)
Growth compresses ~30-40% per tier doubling.
NRR (median)
Top decile typically +10-15 points above median.
GRR (median)
Logo retention closes the gap as ICP narrows.
Gross Margin (median)
Improves as infra cost ratios fall and services taper.
Burn Multiple (median)
Top quartile beats median by ~0.4x at every tier.
Magic Number (median)
S&M efficiency holds best at the top end.
CAC Payback Months (median)
Enterprise mix lengthens, PLG shortens.
Rule of 40 (median)
Growth falls, margin rises, score holds.
How ARR tier maps to funding stage in practice
The two axes do correlate, just not as cleanly as the stage label implies. In the 2024 KeyBanc cohort, Series A respondents clustered across $1M to $10M ARR with a median around $3M. Series B sat across $5M to $25M with a median around $9M. Series C ran $15M to $60M+. So if you are using the per-stage pages on this site as a starting point, the matching ARR tier columns above are:
- Seed / pre-A: read the $1M column with caution. Wide distributions, high variance.
- Series A: read both $1M and $5M columns and place yourself between them based on your current ARR.
- Series B: read $5M and $10M columns. Most Series B operators land in this band.
- Series C: read $25M and cross-check against the $50M public-comp anchor.
- Growth / pre-IPO: $50M tier plus the public-comp dataset is the right reference.
For the stage-specific lens, the existing seed-stage metrics, Series A, and Series B pages translate the stage milestones into the same metrics. This page inverts that: pick by ARR first, derive the stage-appropriate read second.
What happens when your tier changes
As you cross from one tier to the next, the metric that compresses first is almost always growth rate. The $1M to $5M jump is fastest; the $25M to $50M jump is slowest. Plan for the compression. A board target of holding 120 percent YoY growth from $5M to $10M is roughly a top decile outcome in the 2025 cohort, not a median plan.
NRR usually expands as you climb tiers because the ICP narrows and expansion motion matures. Top quartile NRR at $50M+ is consistently 12 to 18 points above the top quartile at $5M. Gross margin moves the same direction more slowly: every doubling of ARR shaves roughly 2 to 3 points off infrastructure cost ratio for hosted SaaS, and professional services revenue (which carries lower margin) becomes a smaller share of the mix.
Burn multiple is where post-2022 expectations bite hardest. The 2021 cohort tolerated burn multiples of 2 to 3 at $5M ARR. The 2026 cohort medians sit at 1.1 to 1.9 for the same tier, and investors price companies above 2 with a discount. If you are running a high burn multiple, plan the path to compress it within the next two quarters, not the next two years.
Reading benchmark ranges honestly
A common failure mode is picking the favourable end of every range and calling it your peer benchmark. Two rules to avoid this:
- Use median, not best-case, as the reference. The interquartile ranges above span Q1 to Q3. The Q2 (median) sits in the middle. If you are pitching yourself as average, that is the number to compare to.
- Compare across all eight metrics, not the favourite three. Cherry-picking burn multiple while ignoring NRR is the most common self-deception in fundraise prep. Investors will pull all eight within the first hour of diligence.
For a more opinionated layer on this, see SaaS metrics during a fundraise (the 14 KPIs investors stress-test first) and SaaS metrics for board meetings (how to present these benchmarks in a board pack without losing the room).
Tier-specific notes operators routinely miss
At the $1M tier, the interquartile spreads understate true variance because survival bias removes failed peers from the dataset. Roughly half of the companies that hit $1M ARR never make it to $5M. Read this tier as what is possible and not what is normal, and prioritise churn and quick ratio over efficiency metrics.
At the $5M tier, the single biggest leading indicator of crossing to $10M is NRR. The companies that stay stuck around $5M are almost always those with NRR below 100 percent. Expansion has to carry an increasing share of growth as new-logo CAC inflates.
At the $10M tier, sales productivity (magic number) becomes the tightest constraint. Most companies that hit $10M have a working motion. Whether the motion can be scaled efficiently is the test. Magic number above 0.8 separates the cohort that compounds from the cohort that plateaus.
At the $25M tier, gross margin and Rule of 40 take centre stage. By this point most investors discount growth rate alone and weight balanced scoring. Companies that under-invested in margin discipline at $10M find it very hard to claw back without a painful re-architecture or layoff cycle.
At the $50M+ tier, public-comp data is the cleanest reference. The composition of the $50M+ cohort shifts toward filed-S1 and listed companies, and those companies publish actuals quarterly. See our public SaaS company benchmarks page for the listed-cohort dashboard.
Use the per-metric calculators with this table
Every metric in the tier table has a free calculator on this site. Run your numbers, then compare to the matching tier column. The calculators carry the same 2026 benchmark distributions in their result panels.
For valuation multiples at each tier, see saasvaluationmultiple.com. For an ARR-tiered view of the financial cost of customer churn, see churncost.com.