The typical U.S. company now spends a median $145,000 a year on software and runs about 35 paid tools. The problem is not simply subscription costs. Duplicate applications, unused licenses, decentralized buying, shadow IT, AI add-ons, renewal surprises, and security gaps can turn a manageable SaaS stack into an expensive operational liability.

For many American businesses, the software budget no longer looks like a handful of predictable subscriptions. It looks more like a second utility bill: recurring, distributed across departments, increasingly difficult to audit, and capable of growing quietly for years.

That is the central problem behind today's SaaS sprawl.

Cledara's September 2026 U.S. data puts the median annual software spend for a company at $145,000, with the middle 50% of companies spending approximately $61,000 to $350,000 per year. The same dataset shows a typical U.S. company running about 35 paid software tools, with the middle half using between 23 and 53.

That figure should not be interpreted as saying every American business wastes $145,000. It does not. Software can be one of the highest-leverage expenses a company makes. A $20,000 CRM can be worth far more than its price if it improves sales execution. A $10,000 security platform can prevent losses that are difficult to quantify.

The issue is whether companies know which software is producing that value.

Increasingly, they do not.

BetterCloud's 2026 State of SaaS research found that the average number of SaaS applications has started rising again after two years of consolidation, driven substantially by artificial intelligence. Organizations now deploy an average of 27 AI-powered SaaS applications, representing roughly 22% of the overall portfolio.

For U.S. businesses, that creates a new management problem. The old SaaS challenge was controlling subscriptions. The new challenge is controlling an environment where virtually every department can discover, trial, purchase, connect, and abandon software without necessarily creating a clean record for finance, IT, security, or procurement.

The result is not merely "too many apps."

It is a lack of institutional visibility into what the business owns, what employees use, what data those applications contain, when contracts renew, and which tools could be eliminated without disrupting operations.

SaaS sprawl describes the uncontrolled or poorly coordinated growth of cloud-based software applications throughout an organization.

It can happen in obvious ways. A marketing team purchases a new project-management platform. Sales adopts another reporting tool. Human resources adds an employee-engagement platform. Engineering introduces a specialized developer service. Then individual employees start subscribing to AI assistants, transcription tools, design applications, research platforms, and automation services.

None of those decisions necessarily looks unreasonable in isolation.

The problem appears when the company looks at them collectively.

SaaS sprawl therefore has several dimensions:

- Financial sprawl: too many subscriptions, duplicate licenses and unnecessary upgrades.

- Operational sprawl: employees use different tools for similar processes.

- Security sprawl: applications receive corporate information without consistent controls.

- Identity sprawl: former employees or contractors retain accounts and access.

- Contract sprawl: renewals occur across different departments and purchasing channels.

- Data sprawl: business information becomes distributed across dozens of vendors.

- AI sprawl: employees adopt AI applications before governance policies catch up.

The most expensive version of SaaS sprawl is not necessarily the company with the largest number of applications. It is the company that cannot explain why each application exists.

A six-figure annual software budget sounds large for a small or midsize company, but it can disappear surprisingly quickly.

Imagine a 200-person U.S. company spending $145,000 annually on software. That is roughly $12,083 per month.

At first glance, that may seem reasonable.

Now divide it across departments.

The company may spend $35,000 on collaboration and productivity, $23,000 on engineering and DevOps, $17,000 on marketing, $10,000 on people operations, $9,000 on sales and revenue tools, $7,000 on data and analytics, and additional amounts across finance, security, customer support, legal, and AI.

Cledara's current U.S. dataset illustrates this fragmentation. Its median annual spending figures include approximately $44,000 for collaboration and productivity, $32,000 for AI platforms and model operations, $23,000 for engineering and DevOps, $17,000 for marketing and growth, and $9,000 for sales and revenue.

The danger is that every individual department can make a perfectly defensible purchasing decision while the organization as a whole becomes inefficient.

This is how software waste becomes invisible.

No single $29 monthly subscription looks like a problem. Fifty such subscriptions become $17,400 annually. Add unused enterprise seats, overlapping platforms, premium tiers and forgotten renewals, and the economics change quickly.

The bigger cost, however, can be the labor required to manage the resulting complexity.

💡 Discover More from Cloud & SaaS

AI & Automation

AI Agents vs. AI Assistants: What’s the Difference—and Which One Should Your Business Use in 2026?

AI & Automation

The Future of Generative AI in Creative Industries

AI & Automation

Machine Learning in Healthcare: Transforming Diagnosis and Treatment

AI & Automation

How Artificial Intelligence Is Redefining Workplace Productivity

Unused licenses are one of the easiest forms of software waste to identify and one of the easiest to underestimate.

A company hires aggressively and buys 100 seats. Six months later, 14 employees leave, 11 change roles, and several teams stop using the application.

The invoice does not automatically become smaller.

Some vendors charge by assigned seats rather than actual usage. Others have minimum commitments or annual contracts that make immediate reductions impossible.

Zylo's 2025 SaaS Management Index reported that organizations were not using 52.7% of licenses, with estimated average waste on unused licenses reaching approximately $21 million across its large enterprise dataset. That number is an enterprise benchmark rather than a representation of the typical U.S. company, but the underlying problem—paying for unused capacity—is highly relevant.

For a smaller business, the dollar figure may be much lower. The principle is the same.

The practical question is not simply "Do we have unused seats?"

It is "Why are we still paying for them?"

Duplication is more complicated because different tools can appear to serve different purposes while solving essentially the same problem.

A company might have one platform for internal project management, another for marketing projects, a third for product launches and a fourth for cross-functional task tracking.

Each department may insist that its preferred platform has unique features.

Sometimes that is true.

Often, the overlap is substantial.

A sensible SaaS review therefore should not begin by asking, "Which software should we cancel?"

It should begin with:

"What business capabilities are we paying for?"

That distinction matters.

If three applications provide project management, the company should compare them based on actual workflows, adoption, integration requirements, security, switching costs and measurable outcomes—not simply monthly subscription price.

Software contracts have a peculiar characteristic: companies often spend more time evaluating a $200 office purchase than a $40,000 annual renewal.

That happens because renewals are easy.

The vendor sends an invoice. The budget already exists. Employees are accustomed to the application. Nobody wants to disrupt operations.

So the renewal happens.

Then it happens again.

BetterCloud's 2025 research found that renewal management remains difficult, with some organizations relying heavily on automated cancellation alerts that may arrive too late to support meaningful negotiation or sourcing decisions.

A mature organization should treat major renewals as business decisions, not administrative events.

At least 60 to 90 days before renewal, the owner should be able to answer:

- How many people actively use the product?

- Which departments depend on it?

- What alternatives already exist internally?

- What has the company paid over the previous 12 months?

- Has the vendor changed pricing?

- Are AI or usage-based charges being added?

- What would migration cost?

- What happens if the company does nothing?

If nobody can answer those questions, the company is not managing the renewal. The renewal is managing the company.

Artificial intelligence has changed the SaaS problem because AI is no longer confined to specialized technology departments.

It is appearing inside sales platforms, customer-service systems, productivity suites, design tools, meeting software, developer environments and standalone applications.

That creates a strange situation: a company may think it is buying one software product while its employees are effectively consuming several layers of AI-enabled services.

Zylo's 2026 SaaS Management Index reports that spending on AI-native applications increased 108%, while usage of applications in the broader artificial intelligence category increased 181% in its dataset. The report also says 78% of IT leaders experienced unexpected charges associated with AI features or consumption-based pricing during the previous year.

This is important because traditional SaaS budgeting assumes relatively predictable seat-based pricing.

AI can introduce a different model.

Costs may depend on usage, tokens, automation volume, API calls, premium models or feature tiers.

That means finance departments can encounter a new problem: the employee count has not changed, but the software bill has.

For businesses that have adopted AI quickly, software governance therefore needs to include consumption monitoring, not merely license counting.

Shadow IT sounds like a security violation.

Often, it begins as an employee trying to solve a legitimate problem.

A salesperson needs to convert a file. A marketer wants to analyze a dataset. A manager needs to record a meeting. A designer wants to remove a background. A developer needs a specialized testing environment.

The employee searches online, finds a tool, creates an account and gets the work done.

From the employee's perspective, that is productivity.

From IT's perspective, it may be an unknown application containing company information.

From finance's perspective, it may be another subscription.

From security's perspective, it may be another external system holding corporate data.

This is why simply banning unauthorized applications rarely solves the underlying problem.

Organizations need a way for employees to request tools quickly, understand approved alternatives and explain the business requirement.

BetterCloud's research found that nearly 60% of IT professionals remained concerned about shadow IT in 2025. Its 2026 research goes further, reporting that 21% of organizations discovered new unsanctioned SaaS or AI tools during the prior year.

The lesson is straightforward:

If employees consistently bypass IT, the organization should investigate what is slowing them down.

Sometimes the answer is poor governance.

Sometimes it is poor internal service.

Usually it is some combination of both.

Every SaaS application creates another place where business data can live.

That does not automatically make the application unsafe. Established cloud services can have excellent security programs.

The problem is visibility.

Security teams need to know which applications exist, what information they contain, who has access, whether multifactor authentication is available, whether the application is still being used and what happens to the account when an employee leaves.

This becomes especially difficult when applications are acquired outside centralized procurement.

A 2025 1Password report notes that employees increasingly use applications outside IT oversight, while its research highlights the relationship between SaaS sprawl, shadow IT and access-management challenges.

Consider a departing employee.

HR marks the employee as terminated.

IT disables the employee's corporate identity.

But what happens to accounts the employee created independently?

The employee may have used a personal email address. The company may not know the account exists. Business documents may remain inside it.

This is not just an access problem.

It is an asset-management problem.

SaaS sprawl tends to become particularly painful during periods of growth.

A 20-person startup can often remember who owns what.

A 200-person company cannot rely on memory.

A 2,000-person organization certainly cannot.

Growth introduces new departments, managers, locations, vendors and purchasing habits. Acquisitions can multiply the problem by bringing in entirely separate technology stacks.

BetterCloud's 2026 report found that mid-market organizations experienced one of the largest jumps in application counts, with the average portfolio increasing from 116 to 164 applications in its measured segment.

This is why SaaS management should not be treated as an administrative cleanup project.

It is infrastructure.

As the company grows, the organization needs a repeatable answer to four questions:

What do we have?

Who uses it?

Why do we need it?

What happens when we no longer need it?

If those questions cannot be answered consistently, software costs will tend to grow faster than organizational understanding.

The good news is that companies do not need to eliminate dozens of applications overnight.

A practical SaaS audit can start with visibility.

Combine information from:

- Corporate credit cards

- Accounts payable

- Procurement systems

- Expense reports

- SSO directories

- Identity providers

- Browser or endpoint data where appropriate

- Vendor contracts

- Department budgets

- Employee reimbursement records

The goal is to find both sanctioned and unsanctioned applications.

A spreadsheet is enough for a first pass.

For each application, record the vendor, product, owner, department, annual cost, renewal date, number of licenses, number of active users, data sensitivity and contract term.

Do not begin by canceling anything.

First build the map.

Every significant application should have an accountable owner.

That does not necessarily mean IT.

A CRM might belong to revenue operations. An HR platform may belong to people operations. A developer platform may belong to engineering.

The owner is responsible for proving that the application still serves a business purpose.

Without ownership, software becomes organizational orphanware.

A license assigned to an employee does not necessarily mean it is being used.

Where possible, review actual activity.

Look at login frequency, feature usage, workflow volume and active users.

But be careful with simplistic conclusions.

A rarely used application might still be essential. Legal software, emergency-response platforms and specialized compliance tools may have low usage but high business value.

Usage is evidence, not the entire decision.

Create categories such as:

- Collaboration

- Project management

- CRM and sales

- Marketing

- Finance

- HR

- Security

- Data and analytics

- Engineering

- AI

- Customer support

- Legal and compliance

Then identify overlapping capabilities.

The objective is not to force every department onto one platform.

It is to identify unnecessary fragmentation.

A simple renewal calendar can produce immediate benefits.

For major contracts, set internal review dates 90 to 120 days before renewal.

That gives the business enough time to:

1. Measure adoption.

2. Review alternatives.

3. Negotiate pricing.

4. Reduce seats.

5. Consolidate applications.

6. Plan migrations if necessary.

The worst time to discover that a tool no longer fits the business is after the renewal deadline.

The answer is not "the most expensive applications."

The best candidates are usually applications with low utilization, overlapping functionality and weak business ownership.

A useful prioritization model is:

Keep: High usage + high business value + clear owner.

Review: Moderate usage + important workflow + potentially better alternatives.

Consolidate: Similar functionality exists elsewhere.

Reduce: Valuable application but excessive seats or unnecessary premium features.

Replace: Poor value, weak adoption or unfavorable economics.

Retire: Little usage, no clear owner and no critical dependency.

This approach is considerably safer than imposing a blanket software-cutting target.

A company that cuts a $40,000 system employees genuinely depend on may save $40,000 in software while creating $100,000 in lost productivity.

That is not optimization.

It is cost displacement.

Traditional budgeting asks, "How much did we spend last year?"

A stronger approach asks, "What are we buying, what are we consuming and what business result does it support?"

For each major application, finance and IT should understand four numbers:

Contracted cost: What the company agreed to pay.

Actual consumption: What employees are really using.

Cost per active user: What the application effectively costs based on real usage.

Business value: What measurable outcome the software supports.

For AI-enabled services, add a fifth:

Variable consumption: What portion of the cost changes with usage.

This becomes particularly important as vendors move toward hybrid pricing models.

A $100,000 annual contract may look predictable until usage-based AI charges add another $20,000.

The most overlooked SaaS cost is often internal labor.

Someone has to:

- Approve purchases

- Configure accounts

- Negotiate contracts

- Provision users

- Remove users

- Answer employee questions

- Review security

- Integrate systems

- Reconcile invoices

- Monitor renewals

- Manage data

- Support migrations

If 100 applications each generate a modest amount of administrative work, the total can become significant.

That is why SaaS management is increasingly moving toward automation.

BetterCloud's 2025 research found that IT organizations were increasingly centralizing SaaS management and using automation to reduce operational burden. The report also noted that the IT-to-employee ratio had reached approximately one IT employee for every 108 employees in its survey population.

The implication is important for executives.

The question is not simply:

"How much software do we buy?"

It is:

"How much organizational complexity does every new software purchase create?"

Companies that want to get control of SaaS spending do not need a year-long transformation program.

A focused 30-day project can establish the foundation.

Week 1: Discover

Collect financial, identity, procurement and employee data.

Create the first master application inventory.

Week 2: Classify

Assign owners, categorize applications and identify duplicate capabilities.

Flag applications with no owner or unclear purpose.

Week 3: Prioritize

Focus on upcoming renewals, high-cost unused licenses, duplicate systems and unauthorized applications.

Do not attempt to solve everything simultaneously.

Week 4: Act

Reduce unnecessary seats, consolidate obvious duplicates, establish renewal deadlines and create an approval process for future software purchases.

Most importantly, publish the new rules internally.

Employees need to know how to request software and what happens after they request it.

Good SaaS governance should feel like a faster path to getting work done—not another bureaucratic obstacle.

A useful software policy can be surprisingly short.

It should explain:

- Who can purchase software.

- What types of purchases require security review.

- How employees request new applications.

- Who owns each application.

- How vendors are reviewed.

- How licenses are reviewed.

- How renewals are approved.

- What happens when an employee leaves.

- How AI applications are evaluated.

- What data employees may enter into external tools.

The policy should also distinguish between risk levels.

A free browser extension and a platform processing customer financial records should not go through identical approval procedures.

Overly rigid policies create workarounds.

Risk-based policies create accountability without unnecessarily slowing employees.

No.

Consolidation can create its own problems.

A business may eliminate specialized applications and force teams onto a general-purpose platform that technically covers the same functionality but performs worse for important workflows.

That can reduce software costs while increasing employee frustration.

It can also create vendor concentration risk.

If one provider becomes responsible for CRM, collaboration, identity, analytics and workflow automation, an outage or pricing change can affect a much larger part of the business.

The objective should therefore be rationalization, not maximum consolidation.

Keep specialized tools when they create meaningful competitive or operational value.

Eliminate them when their differentiation is mostly organizational habit.

The SaaS problem crosses organizational boundaries.

For CFOs, the issue is financial visibility and recurring expenditure.

For CIOs, it is architecture, operational complexity and employee productivity.

For CISOs, it is access, data exposure and third-party risk.

For procurement teams, it is vendor leverage and contract management.

For department leaders, it is whether employees have the tools required to perform their jobs.

The most effective organizations therefore treat SaaS management as a cross-functional discipline.

No single department sees the entire picture.

Finance sees invoices.

IT sees accounts.

Security sees risk.

Procurement sees contracts.

Employees see workflows.

Only by combining those perspectives can the organization determine whether an application is genuinely valuable.

The $145,000 figure is useful because it makes SaaS spending tangible, but the deeper issue is not the dollar amount.

It is the management model behind it.

Software has become so easy to acquire that many organizations still manage it as though it were difficult to obtain. Employees can sign up for services in minutes. AI tools can be tested instantly. Teams can purchase specialized platforms without waiting for a traditional IT project.

That flexibility is valuable.

Trying to reverse it completely would be counterproductive.

The better response is to build a company where software remains easy to acquire responsibly.

That means maintaining a current inventory, assigning ownership, measuring usage, reviewing renewals, controlling sensitive data and giving employees a fast route to approved tools.

The strongest SaaS strategy is therefore not "buy fewer applications."

It is make every application accountable.

When a company knows what it owns, why it owns it, who uses it, what it costs and when it should be reconsidered, the $145,000 software budget stops being an invisible collection of subscriptions.

It becomes a managed business investment.