A return is three events, not one. It is a refund, an inventory movement and a cost event, and most setups record only the first. That is why December margin often looks different from the November margin that was reported at the time.
The three paths a returned unit takes
Back to sellable stock, at the cost layer it left with. Into rework, which adds cost before it can be sold again. Or written off, which removes the unit and leaves its cost as an expense. Each path belongs somewhere different in the numbers, and the difference between them is not small at peak volumes.
What happens when only the refund is recorded
Revenue is reduced, which is correct. Stock is not increased, so inventory is understated. Cost of goods is not adjusted, so margin is overstated on the original sale and never corrected. Multiply that across a return rate of ten or fifteen percent on a heavy promotional month and the reported margin drifts away from the real one.
The timing problem
Returns arrive after the period that produced them. A refund processed in December often belongs to an order from November, which means a month that was already reported keeps moving unless the link back to the original order is intact. Brands that reconcile on payout totals rather than order detail usually cannot make that connection, so the adjustment lands in the current month and quietly distorts both.
Bundles make it harder
A returned bundle has to put its components back, not just reverse a revenue line, or the stock figures for every product sharing those components drift.
What good looks like
The return is recorded where it happened, carries its cost path, and stays linked to the order it came from. The margin on the original sale gets corrected rather than re-estimated.
Focal treats a return as an inventory and cost event as well as a refund, at the cost layer the unit carried, with the link back to the original order intact.
