Product taxonomy

A product category node is not a label. It is routing logic — for your PIM, for your channels, and for your buyers.

In a PIM, the category node triggers the attribute schema. Assign a circuit breaker to a generic "Electrical Components" node and it inherits a five-field template. Place the same breaker in "Electrical Components > Protection Devices > Circuit Breakers > Miniature Circuit Breakers" and it inherits thirty fields: rated current, breaking capacity, number of poles, trip characteristic, and mounting type. Those thirty fields are what a buyer filters on when speccing a panel. Without the deep node, the fields don't exist — and the product doesn't appear when the filter runs.

In channel navigation, the category node determines which facets are exposed. A product in the wrong node isn't ranked low in a filtered view. It is absent from that view. The buyer who drills into the right sub-category and applies their filters never sees it.

B2B catalogs amplify both problems. Electrical distributors, industrial MRO suppliers, and process manufacturers routinely carry 200,000 to 1,000,000 SKUs. A classification error multiplied across ten thousand similar items quietly kills an entire product line's discoverability. There is no error message. Products simply don't appear where buyers look.

Your internal taxonomy is only one of several hierarchies a product has to navigate. The major standards each serve a different part of the B2B supply chain.

GS1 GPC (Global Product Classification) uses a four-level Segment/Family/Class/Brick hierarchy, built for consumer and grocery supply chains. It is common in retail EDI workflows and some foodservice distribution, but thin on technical attributes for industrial products.

UNSPSC (United Nations Standard Products and Services Code) is the lingua franca of B2B procurement. If you sell through eProcurement punchout catalogs — Coupa, SAP Ariba, Jaggaer — your buyer's P2P system routes line items by UNSPSC code. A wrong code sends a hydraulic fitting to the wrong approval queue. A missing code can block the transaction entirely.

eCl@ss is dominant in European industrial and manufacturing sectors, with a thirteen-level hierarchy and rigorous attribute templates per class. If you sell technical products into Germany, Austria, or Switzerland, eCl@ss compliance is frequently a customer contract requirement, not an optional add-on.

ETIM (Electrotechnical Information Model) is the de facto standard for electrical, HVAC, and building-services products. Grainger, Rexel, and Sonepar have all adopted it. Each ETIM class defines a mandatory feature set — a contactor has a defined list of technical attributes that buyers expect to filter on, and any distributor listing that contactor without them is at a structural disadvantage.

Channel-specific trees — Amazon Business, Google Shopping, Grainger, and any marketplace you list on — each maintain proprietary taxonomies. They update without notice. A product correctly classified in one channel's tree can be miscategorized on another simply because the two trees diverged.

Most B2B catalog teams maintain one internal taxonomy and map outward to whichever standards their customers or channels require. That mapping layer is where errors accumulate and governance breaks down.

Stopping too shallow. Two or three levels feels manageable. "Fasteners > Bolts" looks clean. But at that depth there is no attribute schema granular enough to expose drive type, head style, thread pitch, or material — the filters a buyer actually uses to narrow 40,000 fastener SKUs to the four that fit their spec. Shallow categorization is not tidiness. It is a filter killswitch.

Organizing around your company, not your buyer. Supplier catalogs group products the way manufacturers think about them: by product line, brand family, or production process. Many catalog teams copy that structure into their PIM. Buyers search by application, media compatibility, regulatory standard, or installation context. A taxonomy built for the warehouse instead of the buyer quietly routes products into nodes nobody opens.

No governance. Without a classification guide and a responsible owner, the same product lands in different nodes depending on who imported it. One team puts a ground fault circuit interrupter under "Wiring Devices"; another puts it under "Circuit Protection." Both are defensible. Neither is consistent. Inconsistency at scale means facets that return incomplete results and buyers who learn not to trust your filters.

Confusing taxonomy with attributes. Taxonomy answers where a product lives. Attributes answer what it is. They are linked — the category node determines which attribute template applies — but they are not interchangeable. Adding more attributes does not fix a miscategorized product; the right facets still won't appear in the wrong node. And a perfectly correct node with an empty attribute schema still cannot be filtered. Both layers must be right.

Most taxonomy decisions are made from inside the catalog. Someone looks at the supplier sheet, picks the most plausible category node, and moves on. That is organizational logic. It is not necessarily buyer logic.

Buyer signals — the search terms buyers type, the filter combinations they run, the attributes they call out in RFQs — often point to different nodes than supplier structure does. An industrial valve classified under a manufacturer's product family might belong, from the buyer's perspective, in a category organized by media type ("Steam Valves" or "Chemical Service Valves") rather than by the brand line the manufacturer used to structure their catalog. Classifying against buyer intent, not supplier structure, is the difference between a product that enters the consideration set and one that sits in a node the buyer never opens.

This is where taxonomy intersects with enrichment. A classification that reflects how buyers search inherits the right attribute schema, surfaces in the right filter paths, and positions the product alongside its actual competitors in the buyer's browse session — not alongside unrelated products that happen to share a supplier's internal family code.

Anglera classifies each SKU to the deepest accurate node per channel — not mirroring the supplier's catalog structure, but mapped against how buyers in that category actually navigate, filter, and decide. The result writes back to the PIM, so your source of truth reflects where the product belongs in the buyer's world, not just where the supplier put it in theirs.

Taxonomy is infrastructure. Done well, it is invisible. Done badly, it is the silent reason a product with complete attributes and strong copy never makes it into a buyer's shortlist.