Composite, drawn from patterns across engagements, not one specific client.
A routine tool audit for a business I worked with, not triggered by any specific complaint, just a periodic check, turned up over two thousand dollars a month in subscriptions nobody was actively using. Nobody had noticed, because each individual charge was small enough to not raise a flag on its own.
How it stayed invisible
Fourteen tools, spread across four different departments, each approved separately by a different person, at a different time, for a reason that made sense in the moment. No one person had visibility into the full list. Finance saw the total spend category, not which specific tools it was made of. Each department saw their own tools, not anyone else's.
That's the exact structural gap that lets subscription bloat hide: the approval is decentralized, but the review never is.
What the audit actually did
Not a fancy process. A single spreadsheet: every active subscription, who owns it, when it was last actually used, cross-referenced against login and usage data where available. Four tools hadn't been logged into in over six months. Three more overlapped almost entirely with a tool a different department already paid for, nobody had realized.
What changed
Seven tools cut immediately. Two consolidated into one, since they served the same function for different teams. Total: just over two thousand a month, recovered without touching a single working process.
The lesson, generalized
Subscription bloat isn't usually a spending problem, it's a visibility problem. Nobody's being careless on purpose, the approvals were all individually reasonable. The fix isn't stricter approval, it's a periodic, centralized review that nobody's job naturally includes doing, so it has to be deliberately scheduled or it never happens.