Why your month-end report disagrees with the floor

The controller says the month shipped at a 31 percent margin. The plant manager says the floor ran flat out and there is no way that number is right. They are both looking at the same ERP. In our experience the gap almost always comes from one of four places, and each one leaves a different fingerprint.

1. Labor posted late or not at all

Jobs close in the system on the day someone keys the last operation, not the day the part came off the machine. If labor from the last week of the month lands in the first week of the next one, this month looks cheap and next month looks expensive. The fingerprint: margins that swing in opposite directions from one month to the next, and a pile of open jobs with no activity in the final days of the period.

2. Material relieved at standard, bought at actual

Standard costs were set at the last update. Purchase prices moved since. The variance account is catching the difference, but it is posted in a bucket nobody looks at in the margin report. The fingerprint: product margins that look stable while the purchase price variance line quietly grows, and purchasing cannot explain why finance is surprised.

3. Work in process that never closed

A job gets 95 percent done, the last operation is skipped or the part ships from a different job, and the original job sits open with cost on it forever. Those dollars never hit cost of goods sold. The fingerprint: a WIP balance that only ever goes up, and a list of open jobs older than any lead time you quote.

4. Shipments and invoices in different months

The truck left on the 31st, the invoice was cut on the 2nd. Or the reverse. Revenue and cost land in different periods, and the margin for both months is wrong in opposite directions. The fingerprint: a shipment log and a sales journal that do not reconcile for the last and first few days of every month.

The four drift points, laid across the month boundary. Each leaves a different fingerprint in the data.

How to tell which one you have

Pull three lists before you argue about the number: open jobs with their last activity date, the purchase price variance trend by month, and shipments versus invoices for the last five days of the period. One of them will be obviously wrong. Fix that one first. In most shops two of the four are live at once, and the one that gets blamed is rarely the one that is largest.

A close everyone trusts is not a reporting project. It is the result of posting the work when it happens, closing jobs when they finish, and keeping standards near actual. The report is just where the discipline shows up.