A dealer portal behind a four-thousand-product catalogue

Four thousand products is not a large catalogue by retail standards. It is large enough that the interesting problems stop being about the catalogue.

Modern Houseware imports and distributes everyday household products across North America and abroad. Dealer pricing, promotions and order processing were being managed outside any single system, which in practice meant they were being managed by people who knew things.

The price is the product

A consumer store has one price per item. A distributor has a price per item per dealer, and that price moves with contracted discount tiers, volume breaks, regional schemes, seasonal promotions and whatever was agreed on a phone call last March. Building the catalogue page is the easy half. Deciding what number to print on it is the rest of the engagement.

We modelled price as something resolved rather than stored: a function of product, dealer, quantity and date, evaluated against the schemes active at that moment. That part is straightforward. The part that matters operationally is what happens next. Once a price is resolved for an order line, it is written onto the line and never recalculated. The order carries the number it was placed at, together with the rule that produced it.

That single decision is what makes a dealer dispute answerable eighteen months later. Without it, reprinting an old invoice runs today's rules over last year's order and produces a different total, which is the sort of thing that ends a distribution relationship.

Catalogue scale, and where it actually hurts

Four thousand active lines creates problems that four hundred does not. Browsing stops working as a way to find anything, so department structure has to carry real weight and search has to be good rather than present. Product data quality becomes visible in aggregate: a few items with missing images or inherited descriptions disappear in a small catalogue and stand out in a large one, because they cluster in whichever department was loaded last.

Then there is the tail. Most of the revenue comes from a small part of the range, and the rest still has to be orderable, priced and visible. A dealer who cannot find one obscure item on the portal will phone the order in, and once they have phoned one order in they phone the next one too. The portal has to be complete before it can be convenient.

The back office is the deliverable

The visible part of this engagement was a customer-facing catalogue with department navigation. The part that carried the value was the engine behind it: user management, order processing, billing and reports, with dealer-specific discount and scheme visibility inside the portal itself, and payment and inventory integrated against the distribution workflow.

Putting the schemes on screen was the change people noticed. Before, a dealer knew their discount and had to trust that it had been applied. After, the entitlement sits next to the price. Support volume for pricing questions is a good proxy for whether a distribution portal is working, and it is the number worth watching in the first six months.

None of this is technically demanding work. It is detailed work, which is a different kind of difficult, and it is the reason distributor portals get quoted too low. The catalogue takes a fortnight. Agreeing what a dealer's price actually is takes considerably longer, and most of that time is spent in rooms with the people who currently hold the answer in their heads.