The Resolution Layer · Case Study in BriefEKOM
Industrial motors distribution.
A cold read of one servo-motor category from an industrial distributor's web catalog — where a buyer picks a motor by the specification on the item page, and about three in four pages have none. One of a set of anonymized EKOM analyses — what surfaced, and where it leads.
Lead finding · most of the category has no specification table
Three of every four servo-motor pages have no specification table — and the gap is one product line.
A bare page carries a title, an abbreviated ERP line and the manufacturer's order-code text. Power, speed and mounting appear as prose, and no field holds any of them, so a filter, a comparison table or an assistant has nothing to read. 94% of the largest line is bare, against 2% of its sibling line — same manufacturer, same category, same site — and the Series field is blank on 83% of the category. The split follows the line, not the item, which is what makes it fixable at the line.
Also surfaced
Photos named for a different motor
Dozens of items display a photo file named for a smaller motor family. The same few files recur across larger-frame items, which points to an attachment made by family. EKOM read the file names, not the pixels; whoever opens the image confirms it.
Values that contradict their own page
A shaft listed in inches on a page whose description says millimetres. An ERP line stating a power rating ten times what the item's own table says. A horsepower that sits about 12% above what its own kilowatts convert to. Each was confirmed on the live page.
One value, several names
One standard flange mount in four spellings, so a Mounting Type filter splits one population four ways. One manufacturer family named with and without a prefix. A few DC drives filed under AC servo motors, and a part number and a barcode keyed into one field.
Context: a category with one clean spine
Item numbers are unique across the whole category, every page title leads with its own item number, and every page loaded. Not counted as a finding — but it is why every finding can be cited to an item a reader can open, and why the fix is a pattern rather than a rewrite.
How EKOM read it. The engine listed one whole category from its own listing pages, read every item page without any file handoff or credentials, and profiled what each field is for from how the pages themselves use it. It then held every item against its siblings: the filled page beside the bare one, the description that gives millimetres where the table gives inches, the siblings that agree where one does not. Every candidate was recounted against the archived pages, and the headline values against the live ones, before anything was written down.
From diagnosis to a record everyone reads the same way. The ten findings collapse into four upstream mechanisms — a product line loaded without its specification, values that look keyed or converted by hand, free-text values with no controlled list, and values attached at the wrong level or merged. Fixing the pages fixes today's catalog; fixing the mechanisms fixes the next load before it goes live. Then the same read runs on every new batch before it is published, so the next bare page is caught in the file rather than found by a buyer.