Services
AI Solutions
TL;DR: Most companies pay for entire SaaS platforms but use a fraction of what they license. According to Pendo's 2019 Feature Adoption Report, 80% of features in the average software product are rarely or never used. For companies with mature processes, the real SaaS cost goes far beyond the subscription fee — it includes workarounds, trapped data, vendor lock-in, and lost competitive advantage. This article breaks down the four hidden layers of the SaaS tax.
You pay per seat, per month, for a project management tool with 40 features. Your team uses eight of them. The other 32 are clutter — tabs nobody clicks, dashboards nobody reads, integrations nobody asked for. You are paying for software you don't use, and you have been for years.
This is not a hunch. According to Pendo's 2019 Feature Adoption Report — still the largest study of its kind, and no one has published data to contradict it — 80% of features in the average software product are rarely or never used. Dig deeper and the picture gets worse: just 12% of features drive 80% of daily usage (Pendo, 2019). Not 50% or 60%. Twelve.
Now multiply that across your entire stack. The average company runs 275 SaaS applications(Zylo 2025 SaaS Management Index). Each one built for the broadest possible market, each one carrying features designed for someone who is not you. SaaS feature usage across your portfolio is almost certainly lower than you think. And the subscription fee is only the beginning of what it costs.
The line item on your P&L says "software subscriptions." What it does not say is that you are paying at least four separate taxes, only one of which shows up on an invoice.
This is the obvious one — the money you spend on licenses nobody uses.
Enterprises waste an average of $21 million per year on unused SaaS licenses, up from $18 million the year before — a 14% increase (Zylo 2025 SaaS Management Index). And 53% of licensed SaaS seats go unused*(Zylo, 2025). More than half. Not underused. Unused.
SaaS is now the third-largest operating expense for many organizations, behind only salaries and rent. Unlike salaries and rent, most companies have no structured process for auditing what they actually get for it. Per-employee SaaS costs reached $9,100 in 2025, with SaaS inflation running at nearly five times the standard G7 inflation rate (Vertice 2026 SaaS Inflation Index).
SaaS bloat is not a minor line-item problem. It is a budget category growing faster than almost everything else in the organization.
This one does not show up in any financial report. It shows up in your employees' calendars.
When your CRM does not talk to your project management tool, and neither talks to your analytics platform, someone has to bridge the gap. Custom scripts. Manual exports. Copy-paste routines that become someone's actual job. 72% of organizations struggle with disconnected data across their SaaS tools (Future Processing, 2026).
The average company runs 15 duplicative training apps, 11 project management tools, and 10 collaboration apps (Zylo 2024 SaaS Management Index). Each one was purchased to solve a problem. Together, they create new problems — and someone on your team is spending their week making those tools cooperate instead of building what your business actually needs.
Every integration you build, every workflow you wire into a vendor's platform, every year of data you accumulate in someone else's database — all of it raises the cost of leaving.
47% of enterprises cite data migration as a significant barrier to switching providers (Flexera 2023 State of the Cloud Report). Buyers are catching on: 74% of SaaS buyers now evaluate switching costs before purchase, up from 47% in 2018 (Deloitte Tech Trends 2023). But by the time you are evaluating switching costs, the lock-in has already done its job.
This is not a failure of planning. It is vendor lock-in working exactly as designed. The more deeply embedded the tool, the more expensive it becomes to replace —and the less leverage you have at every renewal.
This is the cost that never appears in a spreadsheet, but it may be the largest of the four.
When your process runs inside a SaaS tool, it runs the way the vendor designed it. Your workflow conforms to their data model. Your metrics are the metrics their dashboard offers. Your competitive advantage — the way your specific business operates differently from every other company in your market — gets flattened into someone else's default settings.
For early-stage companies, this is fine. Off-the-shelf tools accelerate speed to market. But there is a crossover point. The more specific and mature your process becomes, the more the generic tool holds you back. You stop adapting the tool to your business and start adapting your business to the tool. That is the opportunity tax: the delta between how your business could operate and how your software allows it to operate.
None of this is accidental. SaaS products are built for the broadest possible market. Every feature serves someone. No single customer uses them all. The vendor's incentive is to add features that reduce churn across millions of users — not to remove features that one customer does not need.
The result is a structural misalignment. The more your process matures, the more you diverge from the "average user" the product was designed for. You pay the same price. You use less of the product. Your SaaS feature usage declines while your bill stays the same — or climbs.
Vendor lock-in reinforces this. According to IDC (via Third Stage Consulting), enterprises with ten or more integrations into a Salesforce have 40% lower churn rates. That is not a happy customer metric but an engineering moat. And 60% of vendors deliberately mask their rising prices (Vertice 2026 SaaS Inflation Index).
Meanwhile, 58% of customers who feel trapped by a vendor will eventually leave anyway — and become detractors (Gartner, 2022). The lock-in does not create loyalty. It delays defection and increases the cost when it finally happens.
The pattern is clear: vendors optimize for retention through dependency. Customers optimize for outcomes. Over time, those incentives diverge. And the customer absorbs the difference as a tax.
For many companies, SaaS remains the right choice. This is not an argument against subscriptions. Rather, it's an observation that the math has changed, and most companies have not recalculated.
The cost of building software has dropped dramatically. LLM inference costs fell 280-fold between 2020 and 2024, from $20 to $0.07 per million tokens (Stanford AI Index 2025). Alphabet now generates 30% of its code using AI (Stanford AI Index 2025). Open-source frameworks cover more ground than ever. Infrastructure costs continue to fall.
It's not that SaaS is inherently bad. It accelerates time-to-market, that's certain. But when your company finds its niche, the question is if it's still paying for software you don't use because the alternative used to be too expensive — and whether that assumption still holds.
"The cost of code has dropped dramatically. Has anyone recalculated whether building the 10% you actually need is cheaper than buying the 100% you don't?"
We're entering a highly debatable "buy vs build" territory. And for build, the question on everyone's mind is 'yes, but what about security, maintenance, and ongoing development costs?' We'll cover that in the next articles.
Buy vs build
