Revenue per employee is total revenue divided by the number of full-time equivalent employees (FTEs). SaaS companies usually track the recurring version, ARR per employee: ending ARR divided by ending FTEs. In the 2026 Aleph and Benchmarkit report the median B2B SaaS company produced about $193K of ARR per employee in 2025, but the number roughly doubles between the sub-$5M and $20M to $50M ARR bands, so compare yourself against companies at your stage.
Most pages that rank for this term are HR glossaries or lists of the companies with the highest revenue per employee. Those are fine for a cross-industry definition, but they will not tell a SaaS founder whether 45 people on $6M ARR is healthy, or what next year's hiring plan does to the ratio. This guide covers the formula and its traps, stage benchmarks read from the original reports, why AI-native companies post higher numbers, and how the metric works as a planning constraint inside a forecast.
The formula
There are three versions worth knowing. They answer slightly different questions.
| Metric | Formula | Best for |
|---|---|---|
| Revenue per employee | Revenue for the period ÷ average FTEs in the period | Comparing with public filings and other industries |
| ARR per employee (ARR per FTE) | Ending ARR ÷ ending FTEs | SaaS benchmarking and board reporting |
| Gross profit per employee | Revenue × gross margin ÷ average FTEs | Comparing companies with very different margins |
Example: a company with $10M of ARR and 50 FTEs has $200K of ARR per employee. If its gross margin is 80%, gross profit per employee is $160K.
The SaaS benchmark reports use the ARR version. ICONIQ defines it as end-of-period ARR divided by end-of-period FTEs, and Aleph's own definition is ARR divided by total full-time employees, with the instruction to use FTEs consistently and not mix in contractors. If you want to compare yourself with those reports, use the same construction.
Numerator and denominator choices that move the number
Revenue per employee looks like a one-line calculation. In practice four choices decide the result, and two companies with identical businesses can report numbers 10 to 20% apart.
Revenue or ARR
Recognized revenue lags ARR in a growing company, because a year's revenue reflects the average ARR across the year, not the exit rate. It also includes services and one-time fees that ARR excludes. A fast-growing company therefore looks more efficient on ARR than on revenue. Pick one and label it.
Ending or average headcount
Pairing ending ARR with ending FTEs is the benchmark convention, and it is consistent: both are snapshots. Pairing a full year of revenue with year-end headcount is the common mistake, because it divides a flow by a snapshot and penalizes companies that hired late in the year.
Illustrative example: ARR grows from $6M to $9M in a year while FTEs grow from 45 to 55. Ending ARR over ending FTEs gives $9M ÷ 55 = $163,636. Using roughly average ARR ($7.5M) over average FTEs (50) gives $150,000. Same company, a 9% gap, purely from the construction.
Employees or FTEs
FTE means full-time equivalent. Two half-time people count as one. Founders who are not on payroll still do full-time work, and most finance teams count them.
Contractors and outsourcing
This is where comparisons break. A company with 45 employees and 8 full-time offshore contractors on a vendor's payroll has 45 FTEs on the HR system but 53 people doing the work. At $6M ARR that is $133K on the first count and $113K on the second. For external benchmarking, follow the report's definition (the Aleph guidance is employees only). For internal planning, track both, because a ratio you can improve by moving people to a contractor invoice is not measuring efficiency.
2026 benchmarks by ARR stage
Three primary datasets publish this metric for SaaS. They sample different companies, so read them as ranges, not a single truth.
Aleph and Benchmarkit, 2026 SaaS and AI Metrics Benchmarks
Published June 1, 2026, with 342 participating companies and 96 reporting ARR per employee for calendar 2025. The detailed section reports a median of $193K, up 29% from $150K the year before, a top quartile of $278,848 and a bottom quartile of $126K. The report attributes the jump to "revenue growth in the numerator and deliberate headcount rationalization in the denominator." Its executive summary quotes a lower $175K median (+17%), so the same report contains two medians; the $193K figure is the one backed by the detailed charts.
| ARR band (Benchmarkit, CY-2025) | 25th percentile | Median | 75th percentile |
|---|---|---|---|
| Under $5M | about $70K | $100,000 | $175,128 |
| $5M to $20M | $110,000 | $154,729 | $221,820 |
| $20M to $50M | $231,210 | $282,000 | $313,330 |
| $50M to $100M | $159,000 | $183,578 | $291,310 |
| Over $100M | n/a | $206K | n/a |
Two other cuts from the same report are useful. By growth rate, companies growing faster than 50% had a median of $235K, while the 31% to 50% cohort was lowest at $136K, which the authors read as hiring ahead of ARR that has not arrived yet. By pricing model, usage-based companies led at $291K, with the report's own warning that part of the gap may be definitional. Several secondary write-ups of this report list $126K as the under-$5M median and $279K as the $20M to $50M median; those are the overall quartiles, not the band medians.
SaaS Capital, 2026 private SaaS survey
SaaS Capital's 15th annual survey of more than 1,000 private SaaS companies puts the median revenue per employee at $141,125, up from $129,724 the previous year. Companies at $1M to $3M ARR had a median of $109,644. At $5M to $10M ARR, equity-backed companies ran $152,295 per employee and bootstrapped companies $177,240, and bootstrapped companies were higher at every ARR band. This sample skews smaller and more bootstrapped than the venture datasets, which is why its median sits below Benchmarkit's.
ICONIQ, venture and growth-stage companies
ICONIQ's September 2024 Topline Growth and Operational Efficiency report shows median ARR per FTE in the first half of 2024 of $130K below $25M ARR, $172K at $25M to $100M, and $249K above $100M. Two years earlier the same bands were $74K, $145K and $206K. Its September 2026 report, State of Scaling: The Great Sorting, shows top-quartile ARR per FTE in 2026 of $149K at $10M to $25M, $290K at $25M to $100M and $478K above $100M.
Summary for planning purposes: below $5M ARR, $100K to $130K is normal and a lot of companies sit lower. Between $5M and $20M, $150K to $180K is the median zone. Past $20M, $200K is the floor of healthy and $280K or more is top quartile. You can see these alongside other operating metrics in the SaaS benchmarks tool.
Why the ratio climbs with scale
Early-stage companies hire ahead of revenue by design. You need an engineering team, a first sales hire and someone to run support before ARR justifies any of them. Fixed functions like finance, legal and IT are spread over a small revenue base: Benchmarkit reports G&A falling from 22% of revenue below $5M ARR to 12% above $100M.
As the base grows, existing customers renew and expand without a matching headcount increase, so each new dollar of ARR needs fewer new people. ICONIQ frames this as the point where ARR per FTE overtakes operating expense per FTE, which in its 2024 data happens around $150M of ARR (up from about $100M in earlier years), with top-quartile companies reaching it past $50M.
The Benchmarkit dip at $50M to $100M is the other half of the story: companies at that stage often expand into new segments or regions and hire ahead of the payoff again.
Why AI-native companies report much higher numbers
Headlines about AI startups with millions of dollars of revenue per employee are mostly estimates for private companies that publish neither audited revenue nor headcount, so treat individual figures with caution. The underlying direction is visible in primary data, though.
ICONIQ's 2026 report tracks "Pacesetters," companies with top-quartile revenue growth over three years that are AI-native or AI-driven. Above $100M ARR their median ARR per FTE is $655K and the top quartile $890K, more than double the typical large SaaS company. Three mechanics explain it:
- Cost moves from people to compute. Inference is a cost of goods sold, not headcount, so an AI product can serve far more revenue per employee while spending more per customer on infrastructure.
- Usage-based revenue scales without new hires. When customers consume more, ARR rises with no sales motion attached.
- Internal AI use. ICONIQ reports that more than 70% of employees at its portfolio companies are daily active users of AI tools, and Benchmarkit attributes an 8-point fall in R&D as a share of revenue (35% to 27%) to AI-assisted engineering.
The same ICONIQ table has a detail most coverage skips: Pacesetters below $10M ARR have a median of only $75K per FTE, and those at $10M to $25M have $115K, lower than the general benchmarks. Companies growing several hundred percent a year hire hard to keep up (ICONIQ reports 146% median headcount growth in 2026 for companies growing 100% or more). High ARR per employee arrives at scale, not at seed.
Margins are the catch. Pacesetter median gross margins run 55% to 80% depending on stage, against the 80% median software gross margin in the Benchmarkit data. That is why gross profit per employee is the fairer comparison for AI products, covered in the AI COGS and gross margin guide.
Limitations of revenue per employee
Is revenue per employee a good metric? It is a good summary and a poor diagnostic. Its blind spots:
- It ignores margin. Two companies at $200K ARR per FTE look identical, but at 80% gross margin one produces $160K of gross profit per employee and at 52% the other produces $104K. See SaaS gross margin for the benchmarks.
- It ignores cost per employee. $150K of ARR per FTE is healthy with a team in a lower-cost market and a problem with an all-San Francisco team. Pair it with loaded cost per FTE.
- It can be gamed by outsourcing. Moving support to an agency raises the ratio without changing the cost base.
- It punishes hiring for future growth. A company hiring a sales team ahead of a new segment will see the ratio fall before ARR catches up. That can be the right call.
- It says nothing about growth. A flat company that cuts staff improves the ratio while shrinking its future.
How it relates to payroll, burn multiple and the Rule of 40
The most useful identity in this topic is simple:
Payroll as % of ARR = loaded cost per FTE ÷ ARR per FTE
If your average fully-loaded cost is $120K and ARR per FTE is $133K, payroll alone eats 90% of revenue. Flip it around to get a break-even target: if non-payroll costs (hosting, software, marketing programs, rent) take 30% of revenue, payroll can be at most 70%, so you need ARR per FTE of at least loaded cost ÷ 0.70. Benchmarkit makes the same point about its bottom quartile: $126K "falls below fully-loaded employee cost in most SaaS markets."
Because people are most of a SaaS company's spend, the ratio moves with the other efficiency metrics. A rising ARR per FTE usually means a falling burn multiple (less net burn per dollar of net new ARR) and a better profit margin, which lifts the Rule of 40 score. Burn multiple tells you whether this year's spending bought growth; ARR per employee tells you whether the organization you have built can ever be profitable at its current shape.
Worked example: two hiring plans from $6M to $12M ARR
All inputs are illustrative. A B2B SaaS company has $6M of ARR and 45 FTEs. Blended fully-loaded cost is $120K per FTE. The plan is to double ARR to $12M over 24 months.
Today:
- ARR per FTE = $6,000,000 ÷ 45 = $133,333
- Payroll = 45 × $120,000 = $5.4M, or 90% of ARR
That sits near the Benchmarkit median for the $5M to $20M band ($154,729) but below it, and payroll at 90% of ARR means the company is burning cash.
| Metric | Plan A: hire 40 | Plan B: hire 20 |
|---|---|---|
| Ending FTEs | 85 | 65 |
| Ending ARR | $12M | $12M |
| ARR per FTE | $12M ÷ 85 = $141,176 | $12M ÷ 65 = $184,615 |
| Payroll (loaded) | 85 × $120K = $10.2M | 65 × $120K = $7.8M |
| Payroll % of ARR | 85% | 65% |
| Marginal ARR per new hire | $6M ÷ 40 = $150,000 | $6M ÷ 20 = $300,000 |
Now apply the break-even test. If non-payroll costs are 30% of revenue, the company needs ARR per FTE of at least $120,000 ÷ 0.70 = $171,429. Plan B clears it. Plan A does not: at $12M ARR it would still be losing money on an operating basis, even though the ratio improved.
Add a realistic 4% annual raise and the loaded cost after two years is $120,000 × 1.04² = $129,792. Payroll becomes 70.3% of ARR in Plan B and 91.9% in Plan A. Salary inflation alone erodes about 5 to 7 points of margin if headcount does not get more productive.
The marginal number is the one to debate in a board meeting. Plan A assumes each new hire adds $150K of ARR, roughly the company's current average, so payroll nearly doubles to double ARR and the business never gets more efficient. Plan B assumes $300K per hire, which only works if most of the new ARR comes from expansion, a usage-based component or a sales team that is already productive. Neither plan is right by default; the point is to make the assumption explicit before the offers go out.
Using ARR per employee in your operating model
In a driver-based model, revenue per employee is an output, not an input. You forecast ARR from its drivers (new customers, expansion, churn) and headcount from a hiring plan, and the ratio falls out of the two. That is exactly why it is valuable as a check: it is the place where an optimistic revenue plan and an aggressive hiring plan collide.
A practical setup for a 36-month SaaS financial model:
- Build headcount by role and start month, with loaded cost per role, as described in SaaS headcount planning. The headcount planner gives a quick version.
- Drive ARR independently from pipeline, conversion, expansion and churn assumptions. Do not back into revenue from headcount.
- Plot ending ARR ÷ ending FTEs monthly for all 36 months, next to payroll as a percentage of ARR.
- Set guardrails: the ratio should rise over the plan, and it should cross your break-even ARR per FTE (loaded cost ÷ the payroll share you can afford) before cash runs out. If it falls for more than a couple of quarters, the hiring plan is running ahead of revenue.
- Compare scenarios. Run the hiring plan against a downside revenue case. A plan that only works at the upside ARR is a plan to cut people later.
The ratio then feeds straight into cash: payroll is most of operating spend, operating spend minus gross profit is burn, and burn determines runway. In Adlega the hiring plan and the revenue drivers live in the same model, so ARR per FTE, payroll share and runway update together when either side changes, and the AI CFO can trace any of those numbers back to the assumptions behind them.
How to improve revenue per employee
- Grow revenue from the existing base. Expansion and lower churn add ARR with little new headcount. See SaaS unit economics for the levers.
- Hire against milestones, not budgets. Tie each GTM hire to pipeline that exists, and each support hire to ticket volume.
- Automate repetitive work. Support deflection, finance close, onboarding and internal tooling are where AI tools have cut headcount needs first.
- Review pricing. Underpriced products need more customers, and more people, to reach the same ARR. Benchmarkit's usage-based cohort leads on this metric; see usage-based pricing.
- Do not improve it by layoffs alone. Benchmarkit's own conclusion is that top-quartile benchmarks are "unrecoverable through cost cuts alone."
Common mistakes
- Dividing annual revenue by year-end headcount. Use ending ARR over ending FTEs, or revenue over average FTEs. Do not mix.
- Comparing with the wrong stage. A $3M ARR company measured against an all-company median will always look bad. Use the band.
- Comparing with public giants. Lists of the highest revenue per employee companies are dominated by firms with decades of scale or very different business models.
- Ignoring contractors. Leaving out a full-time offshore team flatters the ratio and hides cost.
- Treating it as a target in isolation. Without margin and loaded cost per FTE next to it, the ratio does not tell you whether you are profitable.
- Forecasting revenue from headcount. Assuming "each rep brings $X" and multiplying produces a plan where the ratio is constant by construction and the check is useless.
FAQ
How do you calculate revenue per employee?
Divide revenue for a period by the average number of full-time equivalent employees in that period. For SaaS, the standard variant is ending ARR divided by ending FTEs. Example: $10M ARR and 50 FTEs gives $200K of ARR per employee.
What is a good revenue per employee for a SaaS company?
It depends on stage. In the 2026 Benchmarkit data the median is about $100K below $5M ARR, about $155K at $5M to $20M, and $282K at $20M to $50M. The overall median is $193K and the top quartile about $279K. SaaS Capital's private-company median is $141K.
What is the difference between revenue per employee and ARR per employee?
Revenue per employee uses recognized revenue for a period, including services and one-time fees. ARR per employee uses the annualized recurring run rate at a point in time. In a growing company ARR per employee is higher, because ARR leads recognized revenue.
Should contractors be included in revenue per employee?
For benchmarking, follow the report you compare against; Aleph's definition counts full-time employees and excludes contractors. For internal planning, also track a version that includes full-time contractors, so outsourcing does not make the company look more efficient than it is.
Is revenue per employee a good metric?
It is a good high-level check on whether the organization can become profitable at its current shape, and investors watch it closely. It is weak on its own because it ignores gross margin, cost per employee and growth. Use it with gross profit per employee, payroll as a share of revenue and burn multiple.
Why do AI companies have such high revenue per employee?
Much of their cost sits in compute rather than people, usage-based revenue grows without a matching sales effort, and teams use AI tools internally. ICONIQ's 2026 data shows a $655K median for fast-growing AI-native or AI-driven companies above $100M ARR, but much lower figures for small ones that are hiring fast. Lower gross margins offset part of the advantage.
What is the difference between revenue per employee and profit per employee?
Revenue per employee measures output per person. Profit per employee (or gross profit per employee) measures what is left after costs. A company can have high revenue per employee and still lose money if its margins are thin or its people are expensive.