Ask a finance team what software the business owns and you will usually get an answer — invoices, a reseller's order history, a renewals spreadsheet, a folder of PDFs. Ask what is actually installed across the machines, and the answer is almost always an estimate.
That gap is rarely dishonesty. It is drift.
A purchase record is created on the day of the purchase and then sits still. The environment it describes does not:
None of these involve a decision anyone would remember making. That is precisely why none of them reach the records.
There is a technical reason the picture stays incomplete. On Windows, a large share of everyday software installs per user rather than per machine — browsers, chat and meeting clients, several developer and creative titles. Those entries live under the signed-in user's profile, which means a discovery run over an office at night sees the machines but not the people, and under-reports without failing.
It is a small detail with a real consequence: a scan can complete successfully, produce a tidy report, and still be missing a category of software entirely. Knowing when a measurement is trustworthy is part of the work.
Reconciliation needs both sides. The purchased side already exists in writing. The installed side has to be measured — machine by machine, while people are actually signed in — before the two can be compared at all.
Until that happens, the difference between what an organisation owns and what it runs is not a number. It is a hope.
Follow Code & Clause Systems on LinkedIn for more notes like this.