SaaS spend per employee is annual software subscription spend divided by average headcount over the same period. It is a cost-intensity ratio, and it is only comparable to a benchmark when both sides count the same spend, the same people, the same months, and a similar peer group.
Two published benchmarks currently sit close together near $9,450 per employee. That tells you where the market is, not whether your number is a problem.
This article gives you the benchmark figures with their sources and periods, a calculation you can repeat each quarter, a way to turn the gap into dollars, and a worksheet to normalize your own number before you compare anything.
What the published benchmarks actually say
A SaaS spend per employee benchmark is a reference figure, not a target. Each one is built from a specific customer base, a specific period, and a specific definition of spend.
The table below lists the retained figures with their owners. Do not blend them into one number, because they come from different data sets.
| Source | Measure | Figure | Period |
|---|---|---|---|
| Zylo | Median SaaS spend per employee | $9,455 | 2025 customer data, published January 29, 2026 |
| Vertice | SaaS spend per employee | $9,200 flat for three quarters through Q1 2026, $9,324 in Q2 2026, preliminary $9,450 in Q3 2026 | Quarterly, 2025-2026 |
| Zylo | Average share of SaaS licenses unused | 36% | 2025 customer data |
| Vertice | Share of apps unused / underutilized | 16% / 47% | 2026 data |
Zylo's median is built on analysis of more than 40 million SaaS licenses and $75 billion in spend under management. Vertice derives its trend from over $75 billion of global processed spend it manages in 2026.
The unused figures look contradictory but are not. Zylo counts licenses against industry-recommended utilization levels, while Vertice counts apps. Different units produce different percentages.
Zylo's median annual SaaS spend of $20.6M suggests its customer base skews toward large organizations. If you run a company at $5-100M ARR, treat these medians as a reference point, not a peer group.
How to calculate your own spend per employee
The formula is simple. The discipline is in the inputs.
Spend per employee = total annual SaaS spend / average headcount for the same 12 months
Run it in four steps:
- Define the cost scope. Decide whether you count contracted subscriptions only, or also software bought on cards and expensed by individuals. Zylo reports that business units control 81% of SaaS spend and IT directly manages 15%, so a number built only from IT-owned contracts will understate reality.
- Use average headcount, not the latest headcount. If you grew from 60 to 100 people during the year, dividing by 100 flatters the ratio. Average the month-end headcounts instead.
- Count people consistently. Decide whether contractors, offshore team members, and part-time fractional roles count as full-time equivalents, then apply the same rule every period.
- Lock the period. Use the same trailing 12 months, or the same quarter annualized, on both sides of the comparison.
Here is an illustrative example with hypothetical numbers. A company with $10M ARR, 60 average employees, and $540,000 of annual SaaS spend lands at $9,000 per employee. Its ARR per employee is about $166,667, so software costs roughly 5.4% of ARR per head.
That second ratio is often more useful than the first, because it ties software cost to what each employee produces.
How to read the gap: dollars, not percentages
Spend per employee is a ratio, not a verdict. A figure above the median can be right for a product-led team with heavy tooling, and a figure below it can hide under-investment.
Convert the gap into dollars before you react:
- Gap in dollars = (your spend per employee - benchmark) x average headcount.
- Waste exposure = your annual SaaS spend x the unused share you can verify in your own license data.
Use the published unused-license rates as a prompt to check, not as an estimate for your stack. Zylo's 36% average and Vertice's 16% unused apps both describe their own customer data.
Then place the number next to revenue productivity. The High Alpha 2025 SaaS Benchmarks Report, covering 800+ founders and operators, gives median and upper-quartile ARR per FTE by ARR band:
| ARR band | Median ARR per FTE | Upper-quartile ARR per FTE | Median headcount |
|---|---|---|---|
| Under $1M | $55,556 | $100,000 | 9 |
| $1-5M | $136,364 | $200,000 | 22 |
| $5-20M | $166,667 | $220,588 | 66 |
| $20-50M | $268,235 | $350,000 | 131 |
| Over $50M | $277,778 | $396,635 | 361 |
Source: High Alpha, 2025 benchmark data, published March 25, 2026. Median ARR per employee rises 61% from the $5-20M band to the $20-50M band.
"A company with rising software cost per head and flat ARR per head is paying for tools that are not producing output. That is the pattern to investigate first."
— Aleksandar Stojanovic, CEO & Founder at FiscallionWhy the number drifts upward
Vertice reports spend per employee rose from $9,200 to a preliminary $9,450 between Q1 and Q3 2026, a cumulative increase of 2.7%. It cites SaaS inflation of 16.4% in June 2026, the highest monthly rate on record, falling to 14.3% since. It also points to AI features moving from free trials to paid add-ons.
Zylo's data shows the same pressure from another angle:
- AI-native app spend rose 108% year over year overall, per Zylo's 2026 index.
- Expense-based SaaS spend rose 267% year over year, with ChatGPT the most expensed application.
- In a survey of 218 IT leaders, 78% reported unexpected charges tied to consumption-based or AI pricing, and 61% cut projects because of unplanned SaaS cost increases.
"The practical read: your per-employee number can rise with no new hires and no new decisions. Price increases and metered AI usage move it quietly."
— Aleksandar Stojanovic, CEO & Founder at FiscallionThat is why the figure belongs in a quarterly review, not an annual one.
What to do once you have a normalized number
Take these actions in order:
- Reconcile spend sources. Match your general ledger, card statements, and contract list so each tool has one owner and one cost.
- Flag the top 10 contracts by cost. For each, record renewal date, seat count, active users, and the person accountable for the decision.
- Check utilization on the largest lines first. A 36% unused rate across a whole stack means little. The same rate on your three largest contracts means real money.
- Separate seat-based from usage-based costs.Headcount planning controls the first. Usage caps and approval rules control the second.
- Tie renewals to the hiring plan. If you plan to add or cut people, the seat count at renewal should reflect that, not last year's roster.
This is where software spend becomes a finance decision instead of an IT housekeeping task. Fiscallion works above the ledger: it takes reliable accounting data and turns it into a spend-per-head view inside the hiring and runway model.
Common mistakes and the replacement move
| Mistake | Why it misleads | Replacement move |
|---|---|---|
| Comparing against a benchmark without checking its definition | Different data sets count different spend and different people | Write down what the benchmark counted before you compare |
| Dividing by year-end headcount | Fast growth makes the ratio look lower than it was | Divide by average monthly headcount |
| Counting only IT-managed contracts | Business units control most spend, so the gap is hidden | Include card and expense-based software |
| Treating the median as a goal | A median describes a population, not what is right for you | Set a range based on your own cohort and revenue per head |
| Cutting seats to hit a ratio | You can lose capability while the number improves | Cut unused licenses first, then review overlap |
| Reviewing once a year | Vendor price changes and AI usage compound between reviews | Review quarterly, alongside the hiring plan |
Spend-per-employee comparison worksheet
Fill this in before you compare your number to any benchmark. If a row cannot be matched, the comparison is not valid yet.
| Item | Your figure | What the benchmark used | Match? |
|---|---|---|---|
| Cost scope (contracted only, or including expensed and card spend) | |||
| Total annual SaaS spend | |||
| Headcount method (average of month-end, or period-end) | |||
| FTE definition (contractors, offshore, part-time included?) | |||
| Period (calendar year, trailing 12 months, annualized quarter) | |||
| Peer group (company size, customer base, stage) | |||
| Spend per employee (spend / average headcount) | |||
| Gap to benchmark in dollars | |||
| ARR per employee | |||
| Spend as a share of ARR per head |
Use the worksheet above to normalize cost scope, headcount, period, and peer group before comparing benchmarks. If the normalized gap raises a decision you want a second opinion on, the fractional CFO service is where that work happens, and every Fiscallion client works directly with Aleksandar Stojanovic at the CFO layer. His background is FP&A leadership experience gained through the growth of a SaaS company to €100M ARR.
FAQ
What is a healthy SaaS spend per employee for a company our size?
No retained benchmark defines "healthy" for a specific company size. The published medians, Zylo's $9,455 and Vertice's preliminary $9,450 for Q3 2026, describe what customers of those platforms spend, not what you should spend. Zylo's customer base also appears to skew large. Build your own range from your average headcount and your ARR per employee, then check it against the High Alpha ARR-per-FTE bands for your stage.
How do I normalize SaaS spend per employee when headcount changes mid-year?
Divide annual spend by the average of your month-end headcounts, not the final figure. If you grew quickly, the year-end number understates how much software each person actually consumed. Apply one FTE rule for contractors and part-time roles, and keep it the same every period. For a quarterly view, annualize the quarter's spend and divide by that quarter's average headcount.
Which SaaS spend per employee benchmark should I use if we are pre-revenue or early-stage?
The retained benchmarks do not break out spend per employee by stage, so none is a clean match. High Alpha's data shows a median headcount of 9 for companies under $1M ARR, which means one hire or one new tool can move your ratio sharply. Use your own trailing quarters as the baseline, track the ratio for direction rather than level, and add the benchmark later as a rough outer reference.
How do I compare our SaaS spend per employee against peer companies fairly?
Match four things first: what spend is included, how headcount is counted, the period, and the peer group. Then compare in dollars and as a share of ARR per employee. Treat sources separately, because Zylo and Vertice measure different customer bases and do not define unused spend the same way. If you cannot confirm the benchmark's definition, use it only as direction.
The benchmark is the start of the question
A spend-per-employee benchmark helps you ask the right question: is software cost rising faster than output per person? The answer depends on your own data, defined consistently, and reviewed on a fixed cadence.
Normalize first, convert the gap to dollars, and tie each large renewal to a named owner and the hiring plan. That turns a benchmark into a decision.
