Markdown Timing & Depth Against a Season

How to read this sheet

Two margins are reported because they move in opposite directions and only one of them is banked. Margin per unit sold is highest when nothing is marked down, and margin per unit bought is lowest, because everything that failed to sell is salvaged. A markdown meeting that reviews the first number and not the second reliably decides to wait. Elasticity is the input that decides the answer and the one worth measuring rather than assuming: a cut of d multiplies units by (1 - d) to the power of minus the figure entered, so 2.5 turns a fifth off into roughly a half again in units. Read it off what actually happened the last time these goods were reduced, not off a category average, and if it is not known, run the sheet at a low and a high value and see whether the decision even changes - often it does not. This model has no cost for teaching customers to wait, and that omission is the real reason retailers hold price longer than any arithmetic supports. It will therefore always lean early, and the earliest-week input exists to contain it: set the date policy will not cross and the sheet answers inside it rather than arguing with it. Also absent: a markdown ladder, since real seasons take price down in two or three steps and this sizes the first and most consequential one; competitor moves; stock held across sizes, where the curve breaks long before the total sells through and the last third is effectively unsellable at any price; and the effect of a markdown on the goods next to it on the floor, which can be larger than the effect on the goods marked.