Safety & personal protective equipment.

This is a real EKOM catalog analysis, with the client's identity removed. The client is a hardware and home-improvement distributor — Pyramex, Safety Works, Edge Eyewear, Forney, 3M and First Aid Only among the lines it carries. EKOM read one complete category of its public catalog, safety and personal protective equipment, and surfaced 21 flagged items, 20 of which are actionable catalog defects. The twenty-first was withdrawn before publication — it reported that certain product web addresses would break, and spot-checked against the live site they did not. The mismatch is still worth tidying; the breakage claimed for it was not. It is in no count here.

What makes this category different is what the buyer is actually choosing between. Nobody is comparing two objects they can hold. They are comparing a noise reduction rating, a lens tint, whether a lens is rated for UV, whether a mask is a respirator or a face shield. The specification is the protection. So the defects that matter here are not thin copy or a missing photograph — they are two fields on one product asserting incompatible things, on the attribute the product exists to provide.

One thing this study deliberately does not do. It reports every finding below as a factual inconsistency between fields on the same product — what the data shows, and all the analysis can establish from outside. Whether any of it carries a compliance consequence is a question for the business and its suppliers. No statute is named here, and none was named in the analysis.

Manufacturer and brand names are kept as they appeared, since on a distributor catalog they belong to the manufacturers rather than to the client. The client's own item numbers, product titles, specification key names and category taxonomy have been removed or generalized; field values shown in quotation marks are verbatim.

What's inside

• At a glance — the severity split, where the findings concentrate, and how the read was done.
• What a buyer runs into first — a hearing protector with three noise ratings, and a sibling whose dissenting number sits on the key a filter reads.
• The receipts — a record that disagrees with itself, four eyewear families each holding the other's lens color, and respiratory protection filed as fire equipment.
• Markers that outlived their meaning — a drop-ship sentinel on stocked, priced goods.
• Where it traces back, and what's next — the five upstream mechanisms most of these findings collapse into, each with the tell that proves it, and the order to fix in.

Nothing was supplied and no internal system was touched. Everything below was read from the public catalog.

Where the 20 concentrate

Every finding here is rated critical or high — a property of what this pass recorded rather than a clean bill of health. A single coherent category yields less low-grade noise than a mixed one.

How EKOM read this catalog

Three of the four contradicted-rating findings sit on hearing protection, where the contested field is the entire reason the product exists.

A third product in the theme fails the other way: its description stops partway through the rating, so a shopper reading that listing sees no completed number at all — not a wrong one, an absent one. Whether the text is cut at source, lost on the way to the page, or lost in EKOM's own reading of it is not settled from outside, and the study says so rather than guessing.

The receipt device: no outside data needed. Each row is one product, and the product's own fields are the proof.

Four separate eyewear families carry the same defect in the same shape: two products in a family hold each other's lens color. The name says one tint, the specification says the other — and on the sibling, the mirror image.

Frame color, anti-fog, UV protection and polarized are the same on both sides of each swap. Only one attribute moved, and it moved between the two records rather than away from them. A clear lens and a smoke lens are not interchangeable in the situation each is bought for: a buyer filtering for clear lenses to work indoors is shown a product the record describes as smoke, and the reverse. A fifth record carries one side of the same inversion with no partner among the findings, so unlike the four above it cannot be corrected by putting two values back where they belong.

Three findings, five products. Category assignment decides what a shopper can reach by browsing. Here it has placed respiratory protection under fire equipment, chemical splash goggles under hearing protection, and half-mask respirators under face shields.

Two further products break the catalog's own path grammar rather than its meaning: one is missing the root segment every other path begins with, and one carries an extra root from a different department. Both were detectable on shape alone, which is only possible because the grammar is otherwise consistent — four segments, one root, throughout the category.

That consistency is worth naming, because it is what makes this theme cheap to fix rather than expensive. The destination paths already exist on comparable products, so nothing has to be invented. And on every product re-filed here, the category leaf is wrong in exactly the same way the path is — the leaf is the last segment of the path, before and after — so re-filing is one decision rather than two that can drift apart later.

The catalog uses sentinel values to mark items it does not price or weigh directly — a convention, and a sound one. The finding is not the convention. It is that the sentinels have reached products the convention was not meant for.

A separate pattern in the same family: one brand answers the category's yes-or-no specifications with Y and N where the others write Yes and No. A facet reading that field builds two buckets that never merge, so a shopper filtering for anti-fog eyewear sees one brand's products or the others', never both. A separate oddity sits in the same specification block on one record — a key answered with its own name as the value — and is not counted as a finding, because nothing in the record says it is wrong.

None of the three is presented here as a defect to correct unattended. A repdigit price may be a deliberate "call for price" state. A placeholder weight may be a documented instruction to rate the shipment another way. A vendor's own vocabulary may be preserved on purpose, with the merging done at the facet layer instead. Each is a reasonable way to run a catalog — and each has to be a decision rather than an accident, which is the difference this pass can see and cannot settle.

20 findings sounds like 20 problems. The great majority of them collapse into five upstream mechanisms — and a mechanism can be fixed once and stopped from recurring. Each one below is admissible only because something in the data proves it.

Three moves, in order. The first needs no new data at all — and one correction that looks easiest deliberately waits.

This pass read and diagnosed. The same structural understanding powers what follows — turning a diagnosed catalog into one a buyer can specify from with confidence.