The average mid-size business runs between 40 and 80 SaaS subscriptions. Finance tracks the top 10. Nobody has the full list. We have done this audit enough times to know the number is always bigger than the CFO thinks.
The full cost of a SaaS stack is not just the subscriptions. Subscriptions are the visible tip of the iceberg. Hidden underneath are integration work, data reconciliation, context switching, vendor management, onboarding for every new hire, offboarding for every former one, security reviews, SOC2 compliance paperwork, and the opportunity cost of teams operating across 15 different UIs.
If you add all of that up honestly, the true cost of a SaaS stack is typically 3 to 5 times the subscription line on the P&L. The CFO sees the subscription cost and approves the renewal. The real cost compounds quietly in labor and friction, and nobody calculates it because the labor is distributed across every department.
The problem is especially bad in the tail of the stack. The bottom 20 tools that each cost a few hundred dollars a month and are used by 3 people. Each of them is too small to audit individually. Together they cost more than the top 3 tools combined, and half of them are replaceable by a 400-line internal app that consolidates three or four of them at once.
This is the case for custom software in the tail, not the head. You do not replace Salesforce. You replace the 6 little SaaS tools nobody notices that could be one internal app. Each consolidation saves subscription cost, reduces integration work, simplifies onboarding, and removes a vendor risk. The payoff is fast because the replacement scope is small.
The reason this does not happen is that nobody owns SaaS portfolio management. Finance approves renewals. IT manages access. Individual teams adopt new tools to solve specific problems. Nobody asks the meta question of whether the stack as a whole makes sense.
A quarterly SaaS audit is the antidote. Pull the full list. Identify the tail. Ask which ones could be consolidated. Build the consolidation. The return is typically measured in annual savings that equal multiple full engineer salaries, and the work is small enough to finish without a big transformation program.