Pricing the "Before" Seat From the Union of Every Option Invents Savings That Were Never There

How you price a user's existing seat decides whether a saving is real. Price it from the union of every license option each privilege could require and you inflate the before side; on one client that phantom inflation was 369,192 dollars of saving that never existed.

One client, two very different "before" totals

On one client, the before-and-after comparison first showed 694 users who looked cheaper after redesign, totaling 540,744 dollars in savings. After the pricing error was corrected, the same comparison showed 252 users cheaper, totaling 171,552 dollars. The 369,192 dollar difference was never a saving. It was the before side being priced the wrong way.

Nothing about the redesign changed between those two numbers. Only the method of pricing the existing seat changed. That is how much a modeling choice on the baseline, not the redesign, was worth.

Union pricing versus minimum-cost cover

Here is the mechanism in plain terms. A user holds many privileges. Each privilege can be satisfied by one or more license options. The question is how to price the seat that covers all of them.

The wrong way is the union. You take every license option that every privilege could possibly require, add them all up, and call that the user's before cost. This double counts, because many privileges can be covered by a license the user already needs for something else. The union assumes the user buys a separate license for every option, which no one does.

The right way is a minimum-cost cover. You find the cheapest set of licenses that together satisfy all of the user's privileges, and you price that set. One base license plus the attach licenses genuinely required, with no option paid for twice. This is a classic set-cover problem, and the answer is the smallest, cheapest set that covers everything the user actually needs.

Microsoft bills the cover, not the union. A user with a base license and the right attach licenses pays for that set, not for every theoretical option every privilege might name. So the only before cost that matches what the client actually pays is the minimum-cost cover.

Why the union inflates only the before side

The damage is one-sided, which is what makes it dangerous. The after side, the redesigned state, was already priced with a minimum-cost cover. The before side was priced with the union. So every bit of union overcount landed entirely on the before total, making the starting point look more expensive than the client's real bill and making the drop to the after state look larger than it is.

An inflated before and a correct after produce a saving that is partly fictional. That is exactly what the 369,192 dollars was. It was not money the redesign removed. It was money the before side never actually cost.

The fix and the check that enforces it

The fix is to price one population the same way on both sides, using a minimum-cost cover throughout. Once that was done on this client, the before total fell to match reality and the saving dropped to the 171,552 dollars that was actually there.

The check that keeps it honest is reconciliation. The per-user totals and the estate-level like-for-like figure are now required to match to within one dollar, and the report build refuses to run if they diverge. A gap between them is treated as a defect to find, not a footnote to write. That is the right posture, because a reconciliation that is allowed to drift is a reconciliation that hides the next union error.

Bottom line

A before cost priced from the union of options is not the client's bill. It is a larger, imaginary bill, and every dollar of the difference turns into a saving that cannot be delivered. On this client that was 369,192 dollars, more than two thirds of the original headline. Price both sides as a minimum-cost cover, reconcile them to the dollar, and the saving that remains is one you can stand behind.

Frequently asked questions