The Assortment

What this pack runs.

A range is reported as a count and a revenue line. Whatever the company calls them — SKUs, listings, titles, part numbers, models, brands — the count goes up, the revenue goes up, and nobody divides one by the other. The cost of carrying one item is real and never stated: listing it and maintaining its data, forecasting it, holding it, giving it shelf or slot, changing a line over to make it, and the planning attention it consumes all sit in overheads that are never divided by the item count. So every item looks free to keep, and only the revenue it adds is visible. This pack turns it into a sort and a multiplication. Sort by revenue, draw a line, and multiply the items below it by what carrying one costs; multiply their revenue by the tail's own margin rather than the range's, which always overstates it. Then the question that actually decides it: if those items went, some of what they sold would move to the rest of the range and would earn the rest's margin, which is higher — so the whole decision reduces to the break-even substitution share, how much of the tail's revenue has to move for pruning to be worth doing. It stops on five conditions: no item count, no revenue, no line drawn, no carrying cost, and no margin on the tail.

The casebooks these cases come from.

The pack reads nine sourced decisions through the framework. They are drawn from these casebooks, each one a set of verified records on a single decision type.

A verified record has passed a mechanical claim-versus-source check, a second model's read, and human approval. The count is what exists today, not a target.