The best catalog decision is the one a seller never has to make

The orderable catalog had grown past ten thousand raw records. Every one of them was there for a real reason, and together they were unusable. A seller facing that list ends up searching it, then asking a specialist.

The underlying problem is a demand signal spread across thousands of near-identical items. It slows quoting, pushes deals toward whatever the seller already knows, and makes investment decisions guesswork.

I did the demand analysis and then owned the reduction: which products exist, which patterns they collapse into, and what a seller is actually asked to decide.

Analyze the demand first: which configurations customers actually order, and what shape they have. Then reduce those raw records into 671 orderable products sitting behind 3 governed purchase paths.

The reduction follows the demand rather than the organization chart. Design patterns covering 85% of orders became the spine, and the rest became exceptions with a route.

Anything not on a governed path is now visibly harder to buy. That is the mechanism working, and it is also a real cost to the small number of customers who wanted the unusual thing.

Apparent catalog complexity fell by roughly fifteen times, and global sellers were left with one configuration decision rule instead of a search. Roadmap investment could be pointed at the standardized configurations the demand analysis showed were carrying the portfolio.

People remember the fifteen-times figure. The 85% was more useful: knowing which design patterns carried most of the orders made the reduction safe rather than a guess.