Most distributor catalogues are organised around how the distributor buys. Products sit under the supplier who sells them, or under the warehouse aisle they live in, and the customer is expected to work backwards from that. It is efficient for the distributor and slow for everyone else.
ERAM's catalogue is organised the other way round: around what the product is, independent of who made it. That is what a taxonomy is for, and it is why ours goes deeper than most.
The shape of it
The ERAM Master Taxonomy classifies to five levels. Sector, branch, group, family, leaf. As at version 1.5 it holds:
| Level | What it is | Count |
|---|---|---|
| L1 | Sector | 20 |
| L2 | Branch within a sector | 195 |
| L3 | Group within a branch | 715 |
| L4 | Family within a group | 1,530 |
| L5 | Leaf — where a product is classified | 5,702 |
Scroll the table sideways to see all columns.
Five thousand seven hundred leaf categories is a lot. It is also the point. A leaf such as Wall Split Inverter — 12000 BTU/h (3.5 kW) is a thing a person actually asks for. Air conditioning is not.
Why depth helps rather than hinders
The obvious objection is that five levels means five clicks. It would, if the taxonomy were the navigation. It is not, and conflating the two is the most common way a good taxonomy produces a bad catalogue.
The taxonomy is a classification scheme, not a menu. L1 and L2 drive navigation. L3 and L4 drive filtering. L5 drives search relevance and reporting.
A buyer never walks the tree. They type what they want, and the depth of the classification is what makes the search return the right forty products instead of the right four hundred. Depth is a precision instrument that happens to live behind the interface.
Attribute-led classification
Two of the newest sectors — Electronics & Components, and Metalworking, Machine Tooling & Precision Manufacturing — are classified attribute-led, and the distinction matters.
In a conventional branch, classification continues until the category describes one kind of product. In an attribute-led branch it stops at the component or tool type, and the remaining distinctions are held as attributes instead. Value, package, tolerance and rating for a capacitor; ISO designation, dimensions, grade and coating for a cutting insert.
The reason is arithmetic. Classifying every capacitor value, package and tolerance combination as its own leaf would produce tens of thousands of categories describing one product family. Nobody navigates that, and nobody maintains it. But a buyer filtering to 100 nF, 0603, X7R, 50 V finds their part in one action — because those four facts are attributes on a record, not four more levels of menu.
Two coding layers, and why
Every classified item carries two identifiers doing different jobs.
- A hierarchical code in the form NN.NN.NN.NN.NN, which describes where the product sits in the tree. It governs navigation and reporting, and it is never reused.
- An operational SKU in the form EG-NNNNNN, which is flat and stable. It goes on the invoice, the label, the delivery note and the order.
The separation exists so that reclassifying a product does not change its SKU. A category can be corrected, split or moved without invalidating a single historical document. Distributors who use the classification code as the order code discover this problem later, and expensively.
Classification is not a stocking commitment
Worth stating plainly, because it is easy to misread: adding a sector to the taxonomy defines how its products would be structured, coded and searched. It does not by itself mean the range is stocked, or that a distribution agreement is in place.
The two are deliberately separate. Classification is a data decision and can be made in advance; supply is a commercial decision with agreements, stock and working capital behind it. Keeping them apart is what allows the catalogue architecture to be built properly before the commercial conversation concludes — rather than being retrofitted afterwards, which is where most catalogue debt comes from.
What it changes for a buyer
- Search returns the right thing. Precision at the leaf is what separates a useful result set from a list.
- Filters mean something. Attributes are structured data, so temper, diameter, refrigerant, pressure rating and standard family are filterable rather than buried in a description.
- Comparison is like for like. Two products in the same leaf are genuinely comparable, which is the precondition for offering alternatives honestly.
- The documentation follows the product. Datasheets, declarations and certificates attach to a classified record, not to a folder someone has to remember to look in.
Governed, not improvised
A taxonomy this size only stays coherent if changes to it are controlled. Ours are: every addition or amendment goes through a numbered change request, codes are never reused or reassigned, and each sector is reviewed against a set of written classification boundaries that decide the hard cases — whether a pre-insulated pipe classifies with the pipe or the insulation, whether a fenestration component classifies to its material or to the opening it belongs to.
Sectors 19 and 20 were added under change requests CR-019 and CR-040 respectively. All existing codes were unchanged.
Figures cited are from the ERAM Master Taxonomy Structure v1.5. For catalogue, classification or data-integration questions, contact technical@eram-group.net.