By: Audrey Denise B. Cachuela
Ask your finance team what the company pays for software each year and you’ll get a number from memory. Ask what it costs to keep those applications working together, and you’ll get a long pause, because that number lives in IT maintenance hours and workarounds that never reach an invoice. That blind spot is the real price of SaaS sprawl, and Rajeev Jaswal, co-founder and CEO of Argo Intelligence, has run into it from the CIO’s chair.
Jaswal spent more than 25 years in enterprise IT, including roles as CIO at Rapid7 and divisional CIO at Red Hat, and watched the cost of running software outgrow the cost of buying it. The visible number is already large. Organizations in Zylo’s 2026 benchmark managed an average of 305 SaaS applications and spent $55.7 million a year on them, up 8% even though their portfolios stayed largely unchanged (Source: Zylo, 2026).
SaaS sprawl, also called software sprawl or application sprawl, is what happens when software piles up across an organization with little centralized oversight. Most of that growth has a good reason behind it. Finance needs forecasting tools, sales needs a CRM, HR needs somewhere to keep employee records, and operations needs to track inventory. Each application can improve its own process.
Trouble starts when applications have to trade information or support work that crosses departments. A customer service rep might need account details from three systems to resolve a single ticket, and the task picks up extra steps even when every tool works exactly as designed.
Software stack costs run well past the subscription line. The hidden costs of SaaS sprawl collect in three places: the integrations IT keeps alive, the records employees reconcile by hand, and the minutes people burn hopping between applications. Each gets absorbed into work people were already hired to do.
Software integration costs keep growing after launch
The first place that work shows up is integration. Picture a business adding a customer service platform that needs records from the CRM, order details from the ERP, and delivery information from an inventory system. IT has to configure each connection, set permissions, decide how data will flow, and keep it all accurate.
Getting connected is the easy part. Vendors keep updating applications, changing APIs, revising authentication requirements, and renaming fields, and any one of those changes can break an integration somebody then has to track down.
That upkeep rarely appears as a line item of its own, which makes it hard to isolate and easy to ignore. The struggle shows up clearly in survey data: 95% of the 1,050 IT leaders surveyed say integration is a hurdle to implementing AI effectively, and the average organization has connected only 29% of its applications (Source: Salesforce, 2025).
If every application needed a direct link to every other one, 10 systems would produce 45 possible connections and 50 systems would produce 1,225. Most enterprises avoid building things that way, using integration platforms and shared services to cut the number of direct connections. The calculation still shows how potential relationships multiply faster than the applications themselves.
Whoever owns that upkeep keeps owning it for as long as the application stays in the stack, and every hour spent there is an hour unavailable for other work. The business pays for complexity twice.
One revenue question, four different answers
Connected systems can still produce conflicting numbers. Imagine a CEO asking on Monday morning, “How much revenue did we book last week?” The sales team checks the CRM, finance pulls from the ERP, an analyst consults the data warehouse, and someone remembers the spreadsheet one department has maintained by hand for years because certain adjustments never made it into the official reporting system. That’s four sources and four slightly different answers, and now the team has to work out why they disagree.
Perhaps the CRM counts deals that finance hasn’t yet recognized as revenue, and the warehouse refreshes overnight while the ERP already holds fresher transactions. By Tuesday employees are comparing records, and by Wednesday they’re settling definitions. The answer lands on Thursday.
Jaswal watched this play out for years before he built anything to fix it. Asked how long it could take to get a straight answer to a simple business question when he was a CIO, he said: “Days. Sometimes longer, for a question you could state in one sentence. The data lived in four systems, the person who understood how they joined together was on leave, and the existing report answered a slightly different question, built a year earlier.”
Data silos get especially expensive when departments keep overlapping records without agreed definitions or clear ownership. A customer can carry one address in the CRM and another in billing, and finance and sales can define revenue differently. The expense appears when employees must investigate those differences before they can trust the numbers, and over time that investigation becomes part of the reporting routine.
The cost reaches decisions too. A leader who needs an answer on Monday can’t always wait until Thursday, and by the time the report arrives, the chance to act may be gone, which turns disconnected data into a problem well outside IT.
The hours employees lose between applications
The last cost sits with the people who produce those numbers. Whoever builds the revenue report starts in the CRM, hops to the ERP to check invoices, opens the warehouse for history, and finishes in a spreadsheet (there’s always a spreadsheet). Each application holds one piece of the answer, so the employee has to remember where everything lives and carry context from one system into the next.
Workers in an Asana Work Innovation Lab survey of 3,004 U.S. and U.K. knowledge workers reported losing 57 minutes a day switching between collaboration tools and another 30 minutes deciding which one to use for a given task (Source: Asana Work Innovation Lab, 2023). That research measured collaboration tools only, which limits how far the numbers stretch to other enterprise applications, though it still shows how fragmented software use eats time.
Take a conservative hypothetical: 100 employees each spend 15 minutes a day navigating applications, locating records, or moving information between systems by hand. Across 220 working days, that comes to 5,500 hours a year, which at an illustrative fully loaded labor cost of $60 an hour equals $330,000. That figure shows how small repeated chores accumulate, and it makes no claim as an industry average or a guaranteed saving from reducing SaaS sprawl.
Tab-hopping carries a second cost in the thinking it forces, because employees have to understand how information relates across systems and check whether records are current before any analysis can start.
Some context switching is unavoidable, because different business functions need specialized software. That leaves unnecessary navigation and manual information gathering as the part worth attacking, and employees who can reach what they need without hopping between applications spend more time on work that calls for their expertise.
Why adding applications can make application sprawl worse
Stacks reach this size one reasonable purchase at a time. A company rarely decides to run dozens of applications all at once, and each addition gets approved on its own business case. The total only becomes visible after dozens of those decisions have piled up.
The difficulty arises when those decisions get judged separately, without asking how each application will affect workflows across departments. A tool that saves one team several hours a week can add work for another, and each department judges the benefit it sees and rarely the cost it creates elsewhere.
Jaswal said, “The request always arrived with a good business case, because the team asking had done the work. What they could not tell me was what it would cost the three teams downstream, and neither could I, not without deeper investigation. So the honest answer is that I approved some of them and found out later because of time constraints and business imperative.”
Those purchases may still be worth making, provided the company first understands the operational impact. Application count says little about efficiency. A company with 50 well-connected systems can run more smoothly than one with 20 that demand constant manual coordination.
Consolidation carries its own trap. Dropping an application trims licensing costs but can also force employees to recreate its functionality by hand. Effective SaaS management means understanding how each application supports a business process and whether its combined operating cost still holds up.
How to audit your software stack for hidden costs
You can start measuring the cost of managing multiple software applications without evaluating all of them at once. Pick the five recurring workflows that matter most, such as monthly financial reporting, customer onboarding, invoice processing, sales forecasting, and inventory reconciliation, and begin with one.
Document that workflow from start to finish and list every application involved, including the spreadsheets and departmental tools that never made it into the official software inventory. Then count the handoffs, meaning every time information moves between systems, departments, or employees. Some happen automatically, and others require someone to export data, request access, or update a record by hand.
Plenty of handoffs serve a purpose, because approvals, financial controls, and security checks protect the business. The audit should flag the steps that exist mainly because information cannot travel efficiently.
Next, count how many employees need access to multiple systems to finish the workflow, and watch for the people who spend their days gathering and preparing information for others. Some of that repeated collection and reconciliation may be simplified.
Finish by estimating the time those activities consume. Suppose a weekly revenue reporting process involves four applications, six handoffs, and three employees who collectively spend four hours a week gathering and reconciling information. That works out to 208 hours a year, and at an illustrative labor cost of $60 an hour, it comes to $12,480 in employee time.
Repeat the exercise across all five workflows, and the places where unnecessary work concentrates start to show. The audit exists to locate work the company pays people to do because its software does not operate together, and cutting applications is only one possible outcome.
How to cut SaaS sprawl costs without replacing your software stack
Once the hidden costs are visible, the instinct is to rip out applications, and established enterprise systems make that expensive. An ERP may support critical financial processes, a CRM may hold years of customer history, and a mainframe application may run transactions the business depends on daily. Replacing any of them means migration, testing, integration changes, and retraining, which adds a fresh set of costs.
Several lower-disruption options exist: better integrations, consistent data definitions, tighter access management, and tools that change how employees retrieve information from applications they already use. Which one fits depends on where the audit found the unnecessary work.
One example of that last option is Argo Intelligence, the company Jaswal co-founded, which connects with SaaS applications and mainframe systems so employees can work across authorized systems using plain-language requests, with no need to replace existing software. A finance leader could request revenue information across connected systems without opening every application involved. The platform is designed to operate within existing permission structures, retain records of where information originated, and require human approval for actions that change business systems.
A common interface across existing systems addresses the employee-facing costs of a fragmented stack. It is also one more application, so the test is the same one the audit applies: whether it removes more steps than it adds. Inaccurate data, duplicate records, and unresolved integration problems stay with the company, which makes clear data ownership and sound governance necessary alongside it.
The next software purchase or renewal on your desk is the right moment to put numbers behind the decision. Audit the SaaS sprawl in your five most important workflows this week, and you will know which handoffs cost the most and which tools are worth keeping.
