Resolution Layer Case Study · Read & Diagnose
Mattress product pages.
A read of a sample of product pages from a large mattress retailer — where a shopper chooses by size first and by price second, and where a page's words and the data behind them are meant to say the same thing.
This is a real EKOM catalog analysis, with the client's identity removed. The client is a large mattress retailer that sells through stores and through its own product pages. EKOM opened 15 of its mattress product pages, spread across seven brands, one at a time in an ordinary browser, and read what each page shows a shopper, what its head carries, and the structured data embedded in it. This is a sample: 15 pages support patterns and examples, and every share below is a share of those pages.
What makes this category different is that a shopper chooses by size first. The same mattress comes in five to seven sizes, each with its own price, and the page is also expected to say what it is, what it was, and how it can be paid for over time. Those are the things a product page states in words, and they are also the things a search engine or an assistant reads from fields.
A shopper reads a page. A search engine and an assistant read what the page says about itself, and where the words and the fields disagree, or a field is missing, the machine has nothing to go on.
Brand names, product names and prices are not attached to individual findings here, because on a retailer's catalog a brand, a product and a price together can identify the retailer. Counts and shares are exact for the sample. Nothing was supplied by the retailer and nothing has been confirmed with it; each finding is a lead until someone with the product system agrees.
What's inside
- At a glance — the findings, the severity split and how the read was done.
- Pages that cannot name themselves — headings and titles that a search engine would need.
- What a size costs, and what a sale is measured against — size prices out of order and original prices with no stated basis.
- Financing a machine has to parse — a payment plan published as a sentence, and what is already strong.
- Where it traces back, and what's next — the three mechanisms the findings collapse into.
Nothing was supplied and no internal system was touched. Everything below was read from the retailer's own public product pages, and the headline checks were re-read two days after the first read.
The four findings
Theme
What it is
Severity
Pages that cannot name themselves
No readable main heading on any of the 15; an empty title on three
high
Size prices out of order
A Queen priced below a smaller size on two mattresses at the first read; one price across five sizes on a third
medium
Reference prices without a basis
10 of 15 pages show a sale against an original price, 22% to 79% off
medium
Financing held as copy, not data
A payment plan on all 15 pages, and in none of the structured data
high
How EKOM read this sample
1
Open
15 product pages spread across seven brands, each opened as a shopper would open it.
Prices are as shown for one store selection.
→
2
Read
The visible page, the page head, and the structured data inside the page.
Only what the page rendered was read.
→
3
Check
Sizes against sizes, the sale price against the original, and the words on the page against the data behind them.
4 findings, 3 mechanisms.
Signals from a sample, not a verified defect list. Some may prove intentional, which is why each is tied to the values behind it. Prices differ by store, and are stated for the store and the day they were seen.
Pages that cannot name themselves.
The lead finding is about identity. Before a shopper reads anything else, a page has to say what it is: a heading for the page, and a title for the search result.
High · Headings and titles
None of the 15 pages has a readable main heading, and three have an empty title.
On 14 pages the heading does not exist; on the fifteenth it exists with no text. Three pages, all on two brands, have an empty page title once loaded. The title is the headline a search result shows, so those three give a result nothing to name them by.
Field
What the 15 pages hold
Verdict
Main heading
None on 14 pages; empty on one
0 of 15 readable
Page title
Empty on three
disagrees
Share-preview title
Set on all three of those pages
present
The answer is already in the page
The product's name exists, and is not used for either field.
On the page re-read two days later, the share-preview title carried the full product name while the loaded page's own title was empty. A heading and a title taken from that name would correct every page on the same template, and the three empty titles sitting on two brands only is what a template-level cause looks like.
One question stays open and is reported as one: on the single page whose served source was also read, the source carried a title and the loaded page showed none, which points to something clearing it after load. That is the retailer's web team's to confirm.
What a size costs, and what a sale is measured against.
Two findings about price. The first is the price of each size of one mattress. The second is the original price a sale is measured against.
Medium · Size prices
On the first read, two mattresses priced a Queen below a smaller size, and a third cost the same in every size.
On one mattress a Queen is priced below the Twin. On another, the Queen is priced below the Twin XL and the Full. A third lists a single price across Twin, Full, Full XL, Queen and King. A shopper comparing sizes sees a smaller bed cost more, or no difference at all.
Stated for the store and the day
Prices differ by store, and one inversion did not reproduce.
The inversions were seen at one outlet on the first read. Two days later, with a different outlet selected, one of the two mattresses listed higher prices for the same sizes and no Queen below the Twin. The inversions may be promotions on the Queen and the flat price may be deliberate, so each is one question for the retailer. A check that price does not fall as size rises, run per store, would flag each before it publishes.
Medium · Reference prices
10 of the 15 pages show a sale against an original price, 22% to 79% off.
Two different mattresses carry the same original price. One mattress's original price is exactly two and a half times its sale price, and held that way on the re-read. These are stated as facts about the page: whether an original price was charged is the retailer's to answer. What the page cannot show is the basis of the comparison, and a reference price that carries its basis lets the page, and any feed reading it, say what it is measured against.
Financing a machine has to parse.
Every page makes a payment offer. None of the offers is in the data.
High · Financing
A payment plan appears on all 15 pages, and only as a sentence.
Each page states a monthly figure, a zero-percent rate and a number of equal payments in the page copy, and a lease line beneath it. The structured data on the same pages holds a product group, a product per size and a breadcrumb list, and nothing that describes a payment offer. An assistant comparing mattresses by monthly cost has to read and parse page copy to find any of it.
Where
What it holds
Verdict
The visible page
A monthly figure and a term, in a sentence
shown
The structured data
No payment or financing offer
absent on 15
The arithmetic behind the sentence is regular. On all 15 pages the monthly figure is the Queen price divided by the term, rounded up to the dollar, and exactly one term fits each page. The term varies by product from 6 to 72 months. That regularity is what makes the fix mechanical: the product-data specification used for shopping feeds carries an installment attribute, a number of months and an amount, and these pages already state both.
Everything a feed would need is already on the page, in a sentence.
What is strong. The structured data is otherwise rich: a product group, a product per size with its SKU, GTIN-12, price and availability, an aggregate rating and a breadcrumb list on every page, and a meta description on all 15. The displayed sale price matched the structured-data price on all 12 pages where the price block was captured. A lease costing more over twelve months than the sale price is how leasing works, and was checked and not reported.
Four findings sound like four problems. They collapse into three upstream mechanisms — and a mechanism can be fixed once and stopped from recurring. Each tell below is a pattern in the data; the mechanism is EKOM's reading of it.
A page template that carries a name for sharing but not for the page
The tell: the share-preview title is set on the three pages whose title is empty, none of the 15 has a readable main heading, and the three empty titles sit on two brands only.
Sizes priced one at a time, with no rule that price rises with size
The tell: one price across five sizes on one mattress, and a Queen below a smaller size on two others at the first read.
Prices and payment plans held as page copy rather than as fields
The tell: a monthly figure that follows a fixed rule on all 15 pages yet appears in no structured data; one original price shared by two different products, and another exactly two and a half times its sale price.
Writing a heading onto 15 pages fixes 15 pages. A template that takes its heading and title from the product name fixes the next product page added, and the one after it.
What this means, and what's next.
A shopper starts by choosing a size. A search engine starts with a title. An assistant starts with the fields. On these pages the first has what it needs, and the other two are reading a page that does not say what it is.
The cost follows the field. A page with no heading, or no title once it loads, may leave a search result with little to name it by. A size that costs less than a smaller one, or an original price with no basis, puts a question in front of a shopper at the moment of comparison. A payment plan that exists only as a sentence is a plan an assistant comparing monthly cost cannot see. A second clock runs underneath: more comparisons now start with an assistant, and an assistant reads the fields.
This pass read and diagnosed. The same structural understanding powers what follows — turning a diagnosed catalog into one a shopper, a search engine and an assistant can all read the same way.
1 · Apply the corrections the page already answers for itself
Take the heading and the title from the product name that already sits in the page's data, and express the payment plan the page already states as fields. Approved as patterns, not page by page.
2 · Settle the questions only the business can answer
Which size-price differences are promotions, what each original price refers to, and which financing term belongs to which product come from inside the business.
3 · Hold the line at intake
Keep ongoing intelligence where product data meets the page, so the next page published without a heading, or with a size out of order, is caught the day it goes live.
This is how EKOM moves a catalog from insight to impact —
and helps keep pages saying what the data says as the business grows.