The inventory is the hard part, not the analysis.
Most companies cannot produce a complete list of what they pay for monthly. Some of it is on a departmental card, some on a founder’s personal card, some renewed annually two years ago and forgotten.
Start with the bank and card statements rather than with what people tell you. The gap between the two lists is itself a finding, and it is usually the most interesting part of the exercise.
Then ask the people using it, not the person who bought it.
The person who selected a system knows what it was supposed to do. The person using it daily knows which three features they actually touch and which spreadsheet they keep alongside it because the system cannot do the thing.
That spreadsheet is the most valuable artifact in the audit. It is the documented gap between the software you own and the process you actually run.
The three leaks.
Departed employees still holding seats. Offboarding almost never includes a licence review, so seat counts ratchet upward and never come back down.
Duplicate capability. Two departments each hit the same problem eighteen months apart and each solved it. Neither knew about the other, and both are still paying.
Modules you already own. Platforms bundle capability that requires enabling and configuring. It is extremely common to be paying separately for something included in a system already in use.
What to do with the result — and what not to.
Do not cancel anything on the strength of low usage alone. The dangerous system is the quiet one that turns out to be load-bearing for a single critical process nobody mentioned. Establish what depends on it first.
Consolidation is a conclusion, not a starting assumption. Sometimes the right answer is to keep everything, fix two integrations, and reclaim nine seats — which is a cheaper and far less disruptive outcome than a migration.
And be wary of anyone taking a percentage of the savings. It points the incentive directly at cancelling things you need.