A data analysis of how many tools companies actually run, how much of that spend is wasted, and why the fix is fewer, better-chosen tools — not another management dashboard.
The average company runs roughly 300 SaaS applications and wastes $21 million a year on licenses nobody opens, according to Zylo’s latest SaaS Management Index. Half of provisioned licenses go unused, more than a third of apps are bought outside IT, and AI-native tools are the fastest-growing line on the invoice. The numbers below show what SaaS sprawl costs in 2026 — and why the 80/20 answer is to run fewer tools, chosen well, rather than more tools, managed harder.
Key takeaways
- The average portfolio holds ~305 apps (240 median) and has stopped growing, per Zylo’s latest index — sprawl is now about cost, not count
- Companies waste $21M a year on unused licenses, up 14.2% year over year, per Zylo’s 2025 index
- Only 49% of provisioned licenses are actually used, per Zylo’s 2024 index — you pay for two seats to use one
- Lines of business control 70% of SaaS spend; IT manages just 26.1% (Zylo 2025)
- The average company runs 15 duplicate training apps, 11 project-management tools, and 10 collaboration apps (Zylo 2024)
- AI-native app spend grew 75.2% in a year, and 66.5% of IT leaders were hit by surprise charges (Zylo 2025)
- SaaS spend now averages $4,830 per employee, a 21.9% year-over-year jump (Zylo 2025)
How many SaaS apps does the average company use?
The average company runs about 305 SaaS applications, with a median of 240, per Zylo’s latest SaaS Management Index. Portfolio size has flattened — app counts dipped 0.07% year over year — so the sprawl problem has shifted from adding tools to paying for the ones already installed.
The average hides a wide range by company size. Zylo’s 2025 index put large enterprises at 660 apps against 152 for small businesses. Different measurement methods land differently, too: Productiv’s State of SaaS Sprawl counted 254 apps on average (364 for enterprises) back in 2021.
The gap between counts is the point. No buyer needs to evaluate 300 tools in a category — they need the one or two worth running. That is the entire premise behind an opinionated, scored directory.
| Source | Segment | Avg apps |
|---|---|---|
| Zylo (latest index) | All companies | 305 (240 median) |
| Zylo 2025 | Small business | 152 |
| Zylo 2025 | Large enterprise | 660 |
| Productiv 2021 | All companies | 254 |
| Productiv 2021 | Enterprise | 364 |
How much SaaS spend is wasted?
Companies waste an average of $21 million a year on SaaS they don’t use, a 14.2% jump year over year, per Zylo’s 2025 index. The prior year’s figure was $18 million (Zylo 2024), so the waste is growing faster than the portfolios that produce it.
The mechanism is simple underuse. Zylo found companies use only 49% of the licenses they provision — you buy two seats to get one active user. On a per-head basis, SaaS spend reached $4,830 per employee in 2025, up 21.9% in a single year, which means the waste compounds against a rising base.
Waste at this scale changes how buyers should shop. The question stops being “which tool has the most features” and becomes “which single tool will actually get used,” because the second seat you don’t need is pure loss.
Why do companies end up with duplicate tools?
Duplication comes from decentralized buying. Lines of business now control 70% of SaaS spend while IT manages just 26.1%, per Zylo’s 2025 index, so five teams each buy their own tool for the same job with no one reconciling the overlap.
The result is measurable redundancy. The average company runs 15 duplicative online-training apps, 11 project-management tools, and 10 team-collaboration apps, per Zylo’s 2024 index. Each duplicate carries its own contract, its own admin, and its own security review.
This is where a shared, opinionated shortlist pays off. When every team pulls from the same project-management or CRM pick, the eleven tools collapse toward one.
What is shadow IT costing teams?
Shadow IT — software bought and run outside IT — accounts for more than a third of the average company’s apps, per Zylo’s 2024 index. Productiv put the figure even higher, at 56% of all apps owned outside IT.
The risk is concentrated in expensed tools. Zylo found 65% of employee-expensed apps score “Poor” or “Low” on risk, and only 2% rate “Excellent.” Two-thirds of IT leaders (67%) name rogue software purchases among their top SaaS challenges.
Shadow IT is not a discipline problem so much as a discovery problem: teams reach for a tool because finding the sanctioned one is harder than expensing a new subscription.
Is AI making SaaS sprawl worse?
AI is now the fastest-growing driver of SaaS cost. Spending on AI-native applications grew 75.2% year over year, per Zylo’s 2025 index — faster than any other category, and largely outside traditional procurement.
The pricing model is the new sprawl. 66.5% of IT leaders reported surprise SaaS charges from consumption-based and AI pricing, where a per-seat contract turns into a per-token bill nobody forecast. Roughly 90% flagged security concerns about AI tools entering the stack.
AI raises the stakes on tool selection because the cost is now variable. Picking the AI writing or AI coding tool your team will standardize on — rather than letting ten seep in — is the difference between one predictable bill and ten surprising ones.
How do you cut SaaS sprawl?
Consolidation is already the mandate: 53% of IT teams consolidated redundant apps last year and 44% were asked to cut SaaS spend outright, per BetterCloud’s 2024 State of SaaSOps. The durable fix is upstream of any dashboard — buy fewer tools in the first place.
The 80/20 moves that work:
- Standardize one tool per job. Collapse the 11 project-management tools to one shared pick before you negotiate a single renewal.
- Buy for usage, not features. With only 49% of licenses used, the tool your team will actually adopt beats the one with the longest feature list.
- Make the sanctioned choice the easy choice. Shadow IT shrinks when the approved tool is obvious and documented, not buried in a wiki.
- Treat AI tools as consumption risks. Pick a standard, cap the seats, and forecast the variable bill before it surprises you.
- Re-check quarterly. Portfolios are flat but prices are not — the waste hides in renewals, not new logos.
That is the job tools8020 does by default: one scored, hands-on-tested pick per category, so the shortlist is made before the sprawl starts. See the companion data on how AI search changed software buying and AI coding tools by the numbers.
Frequently asked questions
How many SaaS apps does the average company use in 2026?
About 305 on average with a 240 median, per Zylo’s latest SaaS Management Index. Large enterprises run far more — around 660 — while small businesses average 152. Portfolio size has flattened, so the current problem is cost and utilization rather than raw app count.
How much money do companies waste on unused SaaS?
The average company wastes $21 million a year, up 14.2% year over year, per Zylo’s 2025 index. The root cause is underuse: only 49% of provisioned licenses are actually used, so organizations routinely pay for seats that no employee opens.
What percentage of SaaS is shadow IT?
More than a third of the average company’s apps are shadow IT, per Zylo, and Productiv put it at 56% of all apps. The security exposure is real: 65% of employee-expensed apps score “Poor” or “Low” on risk.
Is AI increasing SaaS costs?
Yes. AI-native application spend grew 75.2% in a single year, the fastest-growing category, per Zylo. Consumption-based AI pricing also drove surprise charges for 66.5% of IT leaders, turning fixed per-seat budgets into variable per-usage bills.
What is the fastest way to reduce SaaS sprawl?
Standardize on one tool per job before renewal season. BetterCloud found 53% of IT teams already consolidate redundant apps. Choosing fewer, better tools upfront beats managing sprawl after the fact, because unused second and third tools are pure cost.
How much does SaaS cost per employee?
SaaS spend reached $4,830 per employee in 2025, a 21.9% year-over-year increase, per Zylo’s 2025 index. Because roughly half of licenses go unused, a large share of that per-head spend produces no return.