A true-up is not a correction. It is the one moment the organization finds out which of its assumptions was wrong, and most organizations spend it arguing about the number instead.
This guide shows how to decompose the variance by driver before booking it, route each component to its functional owner with the assumption quoted, report gross movements before netting, and keep a record of which assumptions failed. Skip it and you repeat the same variance every quarter, paying the full cost of being wrong and collecting none of the information.
What this guide is for
Actuals have arrived and they do not match. This guide is about what to do with that information, and specifically about how to stop the true-up from being the end of the conversation when it is the most useful beginning available.
The failure it addresses
A true-up gets treated as an accounting event. The difference is quantified, the entry is booked, the rate is adjusted, and the period closes. Everyone involved has done their job correctly and the organization has learned nothing.
The information was in the variance, not in the entry. A variance is the only reliable signal that an assumption held somewhere upstream is no longer true, and it arrives already attributed to Finance because Finance is where it becomes visible. That attribution is the problem. The assumption that failed was almost never Finance’s.
There is also a quieter failure. A true-up that nets to something small looks like a good period. It can just as easily be two large movements in opposite directions, one in coverage and one in affordability, cancelling each other out. Netting is how an organization congratulates itself on an accrual that was wrong twice.
The use cases you can run
Decompose before you book
Split the variance by driver before anything is recorded: volume, mix, rate, lag, and eligibility. Booking first fixes the number and ends the inquiry, because once the entry exists the question feels answered.
Route each component to its owner
Every decomposed piece goes back to the function that owns the assumption underneath it, with the assumption quoted. Not as blame. As information that function has no other way of receiving, because they do not see the financial consequence of their own operating change.
Ban the net
Report gross movements, both directions, before any netting. Then net.
Write the assumption obituary
For every assumption that failed, record what it was, when it stopped being true, how long it took to notice, and what would have made you notice sooner. Keep the file. It becomes the most honest forecasting asset the organization owns.
What good looks like
Variances are decomposed before they are booked. Each component has a named owner outside Finance. Gross movements are visible before netting. And the organization can say how long it takes to notice that an assumption has stopped being true, because it measures that.
What it costs to skip it
The same variance, next quarter, with the same explanation. An organization that trues up without governing the true-up is paying the full cost of being wrong and collecting none of the information.
Where this connects
The accrual this trues up is Accrual Design. The reason a variance that surfaces in reporting was created three layers upstream is the 5 Layers of GTN™. The room where the routing in use case 02 actually happens is the Alignment Labs™. The patient-side view of the same ten exits is at Where did your treatment stop? What is actually testable at the first three checkpoints after launch is set out in Your First Ninety Days.
Former employers and products referenced are matters of professional history. The views expressed are my own and are not made on behalf of any current or former employer, and no commercial outcome for any named product is attributed to me personally. All trademarks belong to their respective owners.