Catalog & Assortment Data Services

Price monitoring tracks a list you supply. Assortment requires discovery — you cannot count what a retailer ranges by checking products you already know about, so entire categories have to be enumerated including products you have never seen.

That makes completeness matter more than freshness. A price feed missing 2% of SKUs is slightly incomplete; an assortment count missing 2% is a wrong number, and the error compounds across period comparisons. It is a heavier collection problem with a different cost profile.

By mapping published category paths to a normalised taxonomy using path structure, product attribute profiles and the actual products in the category — a category is defined by its contents more reliably than by its name. Mapping confidence is delivered per record.

We retain the published path alongside, so counts are auditable, you can remap to your own taxonomy without recollecting, and retailer conversations can use their category language rather than ours.

Usually, and we flag the cases where we cannot. Classification requires continued absence across multiple runs plus disappearance from category listings, not just an unavailable product page.

The distinction matters in both directions: a stock-out recorded as a delist corrupts range analysis, and a delist recorded as a stock-out sends supply chain chasing a decision that has already been made. Ambiguous cases carry a confidence score rather than a determination.

Weekly for full category counts, since range changes on a weekly to monthly rhythm and daily full crawls cost far more for little additional signal. Daily is worth it for change detection — monitoring category listings for additions and removals rather than re-enumerating everything.

That hybrid is the usual configuration: weekly full counts for breadth analysis, daily change detection for newness and delisting alerting.

We detect it and flag it, because it is the largest source of false alarms in assortment data. A retailer splitting one category into three produces an apparent range expansion that never happened.

Because we retain published paths alongside normalised categories, restructures are identifiable as restructures rather than being silently absorbed into period-over-period comparisons. Flagged periods can then be excluded or adjusted rather than misread.

Yes — shared SKU counts, exclusive counts per retailer and overlap percentage by category, tracked over time. It requires product identity matching across retailers, which is delivered with confidence scores.

Overlap is one of the more strategically useful outputs, because differentiation is usually asserted internally rather than measured. Retailers are frequently surprised by how high overlap actually is in categories they consider distinctive.

SKUs, deduplicated for multi-placement. A retailer listing the same product in four categories has one SKU, not four, and counting placements inflates range breadth by a variable and misleading amount.

Variants are handled per your definition — whether a colour or size counts as a separate SKU differs by category and by how you plan internally, so we agree it during scoping rather than imposing one rule across everything.

Yes, using retailer-specific own-label brand mappings rather than name matching, since many retailer brands do not carry the retailer's name and some are exclusive third-party brands.

Own-label share by category, and new own-label introduction detection, are among the most requested fields here — particularly in grocery and general merchandise where own-label expansion is the defining competitive dynamic.

We quote individually, and here the driver is category crawl scope rather than SKU count — because the SKU count is an output, not an input. Retailer count and category breadth determine the work.

A defined category set across a few competitors at weekly cadence sits at the lighter end. Full-catalogue enumeration across many retailers with daily change detection sits considerably higher. One scoping call, a free pilot on your own categories within 48 hours, then a fixed monthly quote. Request a quote.