SaaS subscriptions · July 24, 2026

Average SaaS spend per employee in 2026: the benchmarks, and why yours arrives on a card

The most cited 2026 figure is a median of $9,455 per employee per year, from Zylo's 2026 SaaS Management Index, built on more than 40 million licenses and $75 billion in spend under management. Benchmarks aimed at smaller companies land near $250 to $350 per employee per month, which is roughly a third of that. Both are correct. They are measuring different companies, and that is the first thing to understand before you compare yourself to either one.

If you came here for a single number to take into a budget meeting, the honest answer is that a borrowed number will not survive the first question a CFO asks, which is why is ours different. What follows is the actual 2026 data with its sources, an explanation of why the published figures disagree by two to three times, a method for calculating a number that describes your company, and the part most benchmark articles skip: where the spend is coming from now, and why that makes it hard to see at all.

What is the average SaaS spend per employee?

The most widely cited 2026 figure is a median of $9,455 per employee per year, from Zylo's 2026 SaaS Management Index, which analyzed more than 40 million licenses and $75 billion in spend under management. Benchmarks aimed at smaller companies land far lower, commonly $250 to $350 per employee per month, because they measure a different population of companies.

The 2026 benchmarks, side by side

Here are the two large public datasets, what each one actually measured, and what each is useful for. Read the middle column before the number.

Swipe the table sideways to compare all columns.

Annualize the second row and you get roughly $3,000 to $4,200 per employee, against Zylo's $9,455. That is not a contradiction, it is a sampling difference. A company big enough to buy a SaaS management platform is usually big enough to carry Salesforce, Workday, an ERP, security tooling, and per-seat contracts negotiated years ago. A 40-person company on a corporate card carries none of that. If you use the wrong benchmark you will either congratulate yourself for being cheap or panic about being expensive, and both conclusions will be wrong.

How do you calculate SaaS spend per employee?

Total annual SaaS cost divided by total headcount. The work is in the numerator, not the division. Include every recurring software charge regardless of how it was bought: invoiced contracts, credit card subscriptions, app store charges, and reimbursed personal purchases. Exclude one-time implementation fees and hardware. Use the same headcount definition every quarter so the trend line means something.

In practice, four passes get you a defensible number:

1. Pull twelve months of AP. Anything paid by invoice, ACH, or wire that recurs. This is the easy half and usually the part finance already knows.
2. Pull twelve months of card transactions. Filter for merchants that charge the same amount on roughly the same day each month or year. This is where the surprises are, and it is the pass most companies skip because it is tedious.
3. Pull reimbursed expenses. Software an employee bought personally and expensed is still your SaaS spend, and it is increasingly where new tools enter the company.
4. Divide by a fixed headcount definition. Pick one, full time equivalents or total badged staff, write it down, and never quietly change it. Most per-employee metrics get ruined by the denominator, not the numerator.

Do that once and you have a number. Do it every quarter with the same rules and you have something better: a rate of change. Spend per employee rising while headcount is flat is the signal worth acting on, and it will not show up in any published benchmark.

Why is SaaS spend so hard to track?

Because most of it no longer goes through procurement. Zylo reports that business units control 81% of SaaS spend while IT directly manages 15%, and that expense-based SaaS spend rose 267% year over year, with ChatGPT now the most expensed application. Software that arrives on a corporate card or an expense report never touches a contract review.

That single shift explains most of what makes this metric painful. The old picture was a procurement pipeline: a request, a review, a contract, an invoice, a line in the ledger. The current picture is a marketing manager expensing an AI writing tool, an engineer putting a monitoring service on a team card, and a sales lead upgrading a seat tier from inside the product. Each of those is a real recurring cost. None of them generates a document anyone reviews. Cledara's dataset shows AI going from 8.8% of all SaaS purchases in April 2025 to 26.4% by March 2026, and AI tools are precisely the category that enters this way, one card charge at a time.

Consumption pricing makes it worse, because the amount is no longer fixed. Zylo found 78% of IT leaders reporting unexpected charges from consumption and AI-based pricing, and 61% saying they had to cut projects because of unplanned cost increases. A per-seat subscription you forgot about costs you the same every month. A usage-priced AI tool you forgot about can cost three times more this month than last, and the first sign is the statement.

What is a good SaaS spend per employee?

There is no universal good number, because a 40-person engineering-heavy company and a 40-person field services company are not comparable. A more useful test is direction and waste: is spend per employee rising faster than headcount, and what share of your licenses go unused? Zylo puts average unused licenses at 36%, so a third of most SaaS bills buys nothing.

Apply that to the median and the arithmetic gets uncomfortable fast. At $9,455 per employee, a 36% waste rate is roughly $3,400 per employee per year going nowhere, or about $340,000 at a hundred people. You will not recover all of it, since some idle licenses are seasonal or contractually locked until renewal. But it reframes the question. The interesting number is not whether you are above or below a benchmark, it is how much of your own spend is doing work.

What to do if your number is high

Three moves, in the order that actually pays.

First, find the subscriptions nobody owns. Every company has tools that survived the person who bought them. The fastest version of this is a focused afternoon of statement review, which we wrote up as a SaaS subscription audit. Unowned tools are the cheapest cut you will ever make, because nobody argues for them.

Second, get ahead of renewals. Price increases are invisible by design: same merchant, same card, same day of the month, a larger amount. We covered the shapes they take in how to catch a SaaS price increase. Renewal and price-change notices almost always arrive by email first, so if they land in a shared finance inbox it is worth learning to turn those notices into a structured spreadsheet of dates and amounts rather than relying on someone remembering to forward one.

Third, close the gap between the charge and the review. The two moves above are periodic, and periodic work always lags. A subscription that renews on the third of the month and gets reviewed at quarter-end has already billed twice. Subscription monitoring reads the same card and bank feeds continuously, flags a renewal before it bills, notices when a recurring amount changes, and surfaces the zombie charges that have been quietly running. Pair it with budget alerts at 80% and 100% of a software budget and the metric stops being a quarterly retrospective.

One caveat, stated plainly because it is what makes the rest credible: a monitoring layer is read-only. It never moves, holds, or blocks money, and it cannot decline a charge on a card it did not issue. What it changes is when you find out. Given that most SaaS now arrives on a card rather than through procurement, finding out on the day is most of the win, and it is the same argument as corporate card monitoring applied to the recurring half of the statement.

How much is SaaS per year for a company?

Multiply your per-employee figure by headcount. At the Zylo median of $9,455, a 100-person company spends about $945,000 a year on software. At a small-company benchmark of $300 per employee per month, the same headcount spends about $360,000. The gap between those two answers is why you should calculate your own.

Why do SaaS spend benchmarks disagree so much?

Because they measure different companies. Datasets built from enterprise SaaS management platforms skew toward large organizations with heavy per-seat contracts, while datasets built from small-business card spend skew toward younger, lighter companies. Both are accurate about their own population. Neither is a target for yours, which is why the calculation method matters more than the headline number.

It is also worth noticing that portfolios keep growing regardless of which dataset you read. Zylo reports large enterprises adding an average of 21 applications per month, and Cledara puts the median company at 25 active subscriptions with the top 10% running 49 or more. A number that goes up every month is not a number you set once a year and check at budget time. It is a feed, and it wants to be watched like one.