Resolution Layer Case Study · Read & Diagnose
Tire specifications.
A catalog read across the specification tables of an online tire retailer — where a figure a shopper relies on to choose a tire disagrees with the standard it belongs to.
This is a real EKOM catalog analysis, with the client's identity removed. The client is an online tire retailer that publishes a specification table for every size of every model it lists. EKOM read the specification tables for a random sample of models from six major brands, one page at a time in an ordinary browser, and ran arithmetic checks on every size row. Eight findings are reported.
This is a sample, not a census. The counts below describe what was read and should not be scaled to the catalog, and nothing here has been confirmed with the retailer. Some of the values may originate upstream of it; a shopper reads them on the retailer's page either way.
What makes a tire's specification table different is what rides on it. Shoppers choose a tire by its load rating, its tread depth, its width and its fit to the wheel, and those numbers are read by the shopper and by any search engine, comparison tool or assistant that takes the table as the tire's record.
For a tire, a wrong load rating is a figure a shopper could rely on when choosing a tire for a heavier vehicle.
Product names, model names, sizes and field values that could identify the retailer have been removed or generalized.
What's inside
- At a glance — the severity split, where the findings concentrate, and how the read was done.
- The numbers that carry weight — a load rating that contradicts its own index, and a tread depth with a stray zero.
- The same size, more than once — repeated rows, and product pages that do not match the table.
- Smaller items, and what is strong — outliers worth a second look, and the checks the table passes.
- Where the fixes group, and what's next — the three kinds of fix the findings fall into, and the order to fix in.
Nothing was supplied and no internal system was touched. Everything below was read from the retailer's public specification pages and the sitemap it publishes.
Where the findings concentrate
Theme
What it is
Findings
Load rating
A maximum load that contradicts the load index
1
Tread depth
A stray zero beside the correct row
1
Repeated rows
Exact duplicates; product pages that do not match the table
2
Outliers
A width, an inflation pressure and a rim range outside the pattern
3
Gaps
Core figures missing on some sizes
1
How EKOM read this catalog
1
Read
Each model's specification table, opened one page at a time in an ordinary browser and copied as shown, not retyped.
Every size each model lists, from a random sample of models.
→
2
Measure
Test each row against the standards it should follow: the load index against the standard load table, overall diameter against the size, revs per mile against diameter.
Where the table agrees with the standard, that is reported too.
→
3
Compare
Look for rows that repeat, contradict a sibling row, or sit outside the pattern the rest of the table sets.
Eight findings, five themes, three kinds of fix.
Machine-surfaced signals from a cold read of a sample, checked against the table's own standards — not a verified defect list. Some will prove intentional, which is why each is tied to the values behind it.
The numbers that carry weight.
A tire's load index is a standard code for the weight it can carry, and the maximum-load figure on the same row should be that code's value. On three sizes in the sample it is not.
Critical · Maximum load contradicts the load index
The load index says one capacity. The maximum-load figure on the same row says another.
On three sizes across two tires, the maximum load does not match the load index beside it. On one of them it states the capacity of a tire six index steps higher, overstating what the tire is rated for by a large margin per tire — a figure a shopper could use to pick a tire for a heavier vehicle. The other two read low.
Size
What the maximum load holds
Direction
A winter tire, size one
The value for an index six steps lower
low
A winter tire, size two
A value that appears in no standard load table
low
A tire of another type
The value for an index six steps higher
high
Every other size in the sample with a readable load index lists the matching maximum load, within a few pounds of rounding. These three are exceptions on a table that otherwise follows the standard, which is what makes them findable and fixable.
High · Tread depth with a stray zero
Eight sizes of one tire list a tread depth an order of magnitude above the real one, beside an identical row that reads correctly.
On one performance tire, eight sizes carry a tread depth far above any real tire's. Each sits next to a row for the same size that is identical in every value except one, which reads as a normal tread depth. The pattern reads as a copy of the size with a stray zero, left in the table beside the correct row. A shopper comparing tread life sees a figure no passenger tire has.
The same size, more than once.
Each model page lists every size in a single table, and a size should appear once. In the sample, some appear several times, and for some models the number of product pages does not match the number of rows in the table.
Medium · Duplicate rows
More than two dozen size rows are exact duplicates, spread across nine models.
The same size and every value repeated. One model lists the same size four times, another lists one three times, and one performance tire carries eleven duplicated pairs. Several other sizes appear more than once with small differences in weight, tread width or country of origin. Those may be genuine manufacturing variants, but nothing on the page labels them as such.
To a shopper a repeated row is clutter. To a system that counts sizes, joins on size or builds a filter, it is a different answer depending on which row it reads first.
Medium · Product pages against the table
On several models the number of product pages does not match the rows the specification table lists.
At the widest, a model has noticeably more product pages than its table has rows. The table and the site count the product differently, and from outside there is no way to tell which is complete. Product pages were counted from the sitemap the retailer publishes for search engines.
This is the finding most dependent on the retailer's own records, and one that may turn out to be intended — a page per part number, for example, where a size has more than one.
Smaller items, and what is strong.
Four further items are leads rather than defects. Each is a value that sits outside the pattern its neighbors set, and each may turn out to be legitimate, which is why they are rated Low.
Item
What the table shows
Severity
Section width
One size lists a width about eight tenths of an inch wider than the size implies; two rows of another size disagree with a third by about four tenths of an inch
low
Inflation pressure
One extra-load size lists a pressure about ten psi above its siblings, and a few metric light-truck sizes list lower than the usual figure. Some light-truck sizes are legitimately rated lower
low
Rim width range
One range is not in the half-inch steps every other range uses
low
Missing core figures
Twenty sizes lack at least one of diameter, revs per mile, tread depth or section width. Most are specialist tires, including a trailer line that lists no revs per mile on any of its sizes
low
And what is strong here
This is a catalog whose specification data mostly holds together, which is why its errors are easy to see. On every size where both figures are given, the overall diameter agrees with the tire size, and every revs-per-mile figure agrees with its diameter, within 3%. Wherever a load index can be read, it matches its maximum load except on the three rows on page 4. A rule written against a table this consistent is practical rather than a line-by-line rewrite.
Eight findings sound like eight problems. They group into three kinds of fix — and a fix can be built once and run on every row, so the next one is caught as it lands. Each group below is read from the values behind it.
Repeated rows
What the data shows: the rows with the stray zero each sit beside a row identical in every value but one, and the exact duplicates repeat every value. This covers the tread-depth finding and the duplicate rows. The fix: collapse exact repeats, and check each new row against the rows beside it as it lands.
Values a standard can test
What the data shows: the wrong maximum load on two of the three rows is a real value belonging to another load index, and on the third a value that appears in no standard table; the outlier width, pressure and rim range each break a pattern the rest of the table keeps. This covers the load-rating finding and the three outliers. The fix: test every row against the load table and its sibling sizes, and raise exceptions for review rather than change them.
Page and table out of step, and gaps in the table
What the data shows: the number of product pages differs from the number of table rows on several models, and the missing core figures sit mostly on specialist lines. The fix: reconcile pages against table rows on a schedule, and flag sizes that lack a core figure.
Correcting the rows found corrects those rows. Checking every row against the standard it should follow corrects those and every row that arrives after them.
A specification table is a record, not a page. When a shopper compares tires, when a search engine builds a result and when an assistant is asked which tire carries a given load, the table is where a disagreement gets settled. Wherever a downstream system takes its data from that table, it takes the error with it, and a copy carries no mark of where it came from.
What this class of defect costs is the wrong tire, not a sale. A capacity above what the tire is rated for can reach a shopper choosing for a heavier vehicle. A tread depth an order of magnitude above the real one reaches a shopper comparing tread life. A repeated row gives a system that counts sizes a different answer depending on which row it reads first.
The load rating raises the stakes above the rest. Tread depth and width are matters of life and fit; a maximum load a shopper may rely on touches safety, and an error there reaches a driver as a specification, not as a typo. It is why a load figure that disagrees with its own index should be among the first rows resolved. Fixed at the pattern level, each of these is fixed once, and a standing check catches it if it comes back as the catalog grows.
Three moves, in order. The first draws on the table and the manufacturers' own data; one correction that looks mechanical deliberately waits.
This pass read and diagnosed. The same structural understanding powers what follows — turning a diagnosed catalog into one every channel can read with confidence.
1 · Apply the corrections the table already answers for itself
Collapse exact duplicate rows. Remove the stray-zero rows once the retailer confirms the pattern. Add plausibility checks for width, inflation pressure and rim range that raise rows for review rather than change them. Each is a pattern-level check, not a row-by-row edit.
2 · Settle the questions that gate the rest — starting with load rating
Setting a maximum load from its load index looks like a mechanical fix, and it waits for one answer: which of the two figures the manufacturer stands behind. So do which repeated rows are real variants and why product-page counts differ from table rows. Those answers come from the retailer's data owners first.
3 · Hold the line at the source
Keep ongoing intelligence where manufacturer data meets the published table, so the next stray zero or mismatched load figure is caught as it lands rather than found later by a shopper.
This is how EKOM moves a catalog from insight to impact —
and keeps every specification reading right as the line grows.