• Commercial SEO

Clothing store SEO: size and color break the index more often than anything else

The promotion of an online clothing store rests on the arithmetic of the catalog: one model is multiplied by sizes and colors, and fifteen thousand models give the search engine one hundred and one thousand addresses. Therefore, the work begins with counting, and not with texts - how many pages are actually in the store and how many of them the search engine received in general. In the lingerie store we measured, two independent methods yielded 115,456 and 112,646 items. The difference of 2,810 products is not a miscount, it is products that Google did not see.

See how it works
Price
after a free audit
Guarantee
30 days after project sign-off
Clothing catalog, our measurement 08/01/2026
115,456 search engine positions versus 112,646 addresses in the sitemap
after a free audit
Price
30
Guarantee

days after project sign-off

115,456
Clothing catalog, our measurement 08/01/2026

search engine positions versus 112,646 addresses in the sitemap

2.4%
The difference between the two calculation methods

— about 2,810 products outside the site map

50
Redirects inside the directory

out of 150 verified product addresses returned a 301, and all - to a different position

3,774
Duplicates in product subcards of the same store

records, individual addresses are repeated up to 55 times

1,544.6
The first answer: main vs. card

ms and 391.4ms on the same site

129
Directory structure

categories, 334 brands, three language versions

a monthly cycle with a crawl and indexing report
Work format
Who it's for

Situations where this service delivers results

Scenario 1 of 5

The catalog is growing, and two or three categories are entered from the search

The assortment is replenished every month, visibility is constant. It's almost always one of two things: the new addresses didn't make it into the sitemap, or they did, but they're competing with each other for the same request. We start with a count - without it, you can't see what's missing, the pages or the order in them.

We'll review your situation in a free audit
Cases

Tasks and results in numbers — all metrics measured by us

online retail underwear store, catalog of over 110,000 product items

Task
To understand why part of the assortment does not appear in the search, and to calculate the real size of the catalog: the numbers in the admin and in the site map did not match.
Solution
Measurement by two independent methods — traversal of the sitemap index of 42 submaps with full deduplication and page-by-page traversal of the standard search engine. Plus a random check of response codes on 150 random product addresses from the sitemap, hreflang analysis, and seven first response measurements for each page type.
Result
115,456 product items by search engine compared to 112,646 unique addresses in the site map — a discrepancy of 2.4%, that is, about 2,810 products are shown by the store in its own search, but not given to the search engine. In a sample of 150 product addresses, 50 returned 301, and all 50 led to a different position: 70B to 70A, L to XL, white to black. There are 3,774 duplicate entries in the commodity subcards, individual addresses are repeated up to 55 times. The first response: main 1,544.6 ms, card 391.4 ms, category 473.2 ms - the main slows down, the catalog is given normally. hreflang is set up by two mechanisms that contradict each other, the Ukrainian version has an invalid language code. Catalog: 129 categories, 334 brands, three language versions. Measurement 01.08.2026.
What's included

Complete list of work and what you get as a result

  • Measuring the catalog by two independent methods — a site map and a full-time search engine: two numbers from which all the work is calculated.
  • A decision by level of addresses (model, size or color) with the calculation of how many pages will be added or removed after its adoption.
  • Gluing variants of the same model with a canonical link: dimensions stop competing with each other for the same request.
  • Selectively check product address response codes and parse every 301 — no redirect should lead to another position.
  • Site map with submaps with deduplication of records so that the number of addresses in it matches the number of products in the catalog.
  • Landing pages for "type of clothing + label": size, color, fabric composition, season — for queries that are already visible in Search Console.
  • Dimensional grids as separate content with markup and links from cards, not a PDF somewhere in the footer of the site.
  • Product marking with size, color, price and availability; return policy in the markup, if you really have one.
  • Reduction of language versions to one mechanism: one set of hreflang, correct language codes, self-referencing canonical on each version.
  • Monthly report: the dynamics of crawling and indexing from Search Console, re-measurement of the directory, a list of what has been done and a plan for the next period.
When this service isn't right

What's not included — so there are no surprises at delivery

  • Photography and processing of product images
  • Filling cards with descriptions of the entire assortment
  • Assembling dimensional grids that the manufacturer does not have
  • Work with marketplaces and social networks
  • Item promises for common queries like clothing category name
Process steps

Transparent stages with approval at every step

Total duration:30–120 days

  1. Measurement of the catalog and the first snapshot of the state

    3-5 working days

    We calculate the catalog with the site map and the standard search engine separately, remove the response codes from a sample of product addresses, look at the language versions and the speed of the main card. At the output are numbers that we will check all the following months.

  2. Address-level solutions

    2-4 working days

    Together we decide what is considered the main page of the model. We show how many addresses will disappear or be added in each option, and fix the decision in writing: it is more expensive to replay it in six months than to make it now.

  3. Cleaning: canonical links, redirects, sitemap

    10-20 working days

    We glue the model variants, disassemble each 301 inside the catalog, remove duplicates from the subcards. The number of addresses in the index at this stage usually drops - as it should, and we warn about this before the start, and not at the moment when the graph has gone down.

  4. Language versions

    5-10 working days

    We reduce hreflang to one mechanism, fix language codes, make canonical self-referential on each version. Not everyone needs the stage - only those who have more than one version.

  5. Landing marks and dimensional grids

    10-20 working days

    We build pages for the "type of clothing + sign" connections based on queries that are already visible in your Search Console, and display dimensional grids with separate content with markup and links from cards.

  6. Monthly cycle

    every month, starting from the second

    Re-measurement, inspection and indexing report, list of completed works and the next batch of works. You make the decision to continue based on this data, not on promises.

Free measurement of the clothing catalog

We count the catalog in two independent ways and see how far the numbers differ. In clothing stores, the discrepancy is not an error, but a symptom: it shows how many products the searcher does not see.

What we measure

  • Two independent catalog measurementsSite map and standard search engine. In the measured underwear store, the difference was 2,810 items — 2.4% of the catalog, which are simply not on the site map.
  • The level at which the addresses are enteredModel, size or color. This determines whether you have 15k pages or 115k and whether they are competing with each other.
  • Where redirects lead inside the directoryWe are checking a selection of product addresses. Measurement finding: 50 out of 150 tested addresses were returning 301s, and all were in a different position, size, or color.
  • Doubles in undercardsHow many entries in the sitemap and how many of them are unique. In the measured store, 116,420 entries gave 112,646 unique addresses, some of which were repeated up to 55 times.
  • Language versions and hreflangHow many are there, whether they are coordinated with each other and whether they are started by two different mechanisms at the same time. We also met this.
  • The speed of the master against the cardOften they differ in times. In the measured store, the main one gave the first response in 1,544.6 ms, and the card - in 391.4 ms.

What you get

  • Two catalog numbers and the percentage difference between them.
  • Selective report on redirects within the directory with examples of addresses.
  • Status of language versions, sitemap and speed separately by type of pages in numbers.
  • A conversation for 30 minutes, where we go through the document point by point.

Timeline: 3-5 working days from the moment we receive the site address

Why is it free

Because the measurement is automated and takes us half a day, and without it it is impossible to distinguish a store that needs to clean the index from a store that needs to fill it. These are opposite jobs. The document is yours regardless of whether we continue to work together.

What's next

After the measurement, we make a plan in two streams: gluing the extra addresses and building the fitting under the connections "type of clothing + sign". Both are priced separately so you can only take one.

Short form: your contact and site URL

  • Contract, act and 30-day warranty

    Every project gets a written contract: scope, deadlines, amount, acceptance procedure. After delivery — act and invoice, then 30 calendar days of warranty.

  • Sole proprietor & bank transfer

    The contractor is a registered sole proprietor. Payment by invoice with closing documents.

  • Rights & access — yours

    Code, design and materials transfer to you after full payment. Domain, hosting, repository and analytics are registered to you.

  • Client portal instead of email chains

    During the project you get access to a portal: contracts, invoices, acts and project status in one place.

  • European clients

    Among our work — projects for Norway, Bulgaria, Moldova and Spain.

  • Verifiable numbers

    Every case in the portfolio comes with a link to a live site and a technical measurement.

  • Audit first, then pricing

    There is no price list on the site intentionally: the scope of the same work differs multiples between clients.

  • We say "no" when unsure

    If the task isn't ours or the deadline is unrealistic — we tell you upfront.

Did not find your case?

Describe how it works on your side — we will tell you whether “SEO for a clothing store” fits and what it means in your situation. No brief and no call: one question, one answer.

What affects the price

Why two seemingly identical tasks are priced differently

  • The level at which the addresses are enteredThe main multiplier of the volume. A directory of 15,000 models and a directory of 115,000 addresses from the same models are different numbers of pages to check, different sitemaps, and different amounts of gluing. We take this figure from measurement, not from words.
  • How many language versionsEach version multiplies the urls and adds another set of hreflang and canonical that should be consistent with the rest. There are three versions in the measured store, and contradictions were found precisely between them.
  • Site map statusA card with duplicates takes longer to unravel than a card with no duplicates at all. There are 3,774 duplicate entries in the product subcards of the measured store, individual addresses are repeated up to 55 times — this has to be reduced to any other work.
  • AccessWithout Search Console, we see the directory, but we don't see what's in the index, and we're working blind. Without access to the CMS, address and canonical link edits go through your developer — adding a reconciliation loop to each iteration.
  • State of dimensional grids and characteristicsTables from the manufacturer are started by the directory and tied to the brand. Grids that are pictures in the descriptions or do not exist at all are a separate volume, and the actual data for them is provided by the store, not us.
  • The pace of updating the assortmentA store where the collection changes twice a year, and a store where a batch with new addresses comes in every week, require different measurement frequencies. In the second case, part of the work becomes permanent.
Technologies & integrations

What we build on and what it connects to

Stack

  • Screaming Frog - directory traversal and response codes. Limitations: on hundreds of thousands of addresses, the bypass takes hours and rests on memory
  • traversal of the sitemap-index with deduplication — counts addresses, not lines: 116,420 entries yielded 112,646 unique addresses
  • standard search engine with an empty query and pagination - a second calculation method independent of the sitemap
  • selective verification of response codes: continuous traversal of hundreds of thousands of addresses does not pay off, sampling does
  • Google Search Console. Limitations: data is delayed, the list of addresses in the reports is incomplete
  • rel=canonical is a hint, not a directive: the search engine has the right to choose another page as canonical
  • hreflang - works only when started by one engine and all versions refer to each other
  • JSON-LD Product / Offer / MerchantReturnPolicy. Limitations: does not raise items, does not provide an extended view in the output
  • GA4

Integrations

  • Google Search Console
  • GA4
  • Google Merchant Center
  • Meta Pixel
  • Rozetka
The order of work: first cleaning the index or immediately the texts

How this option differs from the alternative

Where does it start?two independent counts of the catalog: in the measured store they differed by 2,810 items
Product addresses from 301 inside the catalogare sampled and managed before traffic hits them
The number of addresses in the index in the first monthscan crash: gluing variants removes duplicates, and this is the expected move
Dimensional gridsseparate content with markup and links from cards
What can be seen in the reportthe dynamics of crawling and indexing on a monthly basis plus re-measurement of the directory
What we need from you

We can't start without this — best to prepare in advance

  1. Access to Search Console - without it, we can see the directory, but not see what is in the index.
  2. Access to CMS or your developer: address rules, canonical links and sitemap are changed there.
  3. A solution that is considered the main page of the model, if the sizes and colors are entered separately. We show the consequences of both options, the choice is yours.
  4. Actual size grids by brands. We will not invent them - these are actual data for which the store is responsible.
  5. The terms of return and exchange are in the form in which they are legally designed: the markup includes what is on the site, not what is desired.
  6. One person from your side who can say yes to the directory structure.

If something is missing — let us know, we'll help you gather it or do it as a separate task.

FAQ

Most frequently asked questions — with concrete answers

Where do you start on my store?

From the calculation. We count the directory in two independent ways — a site map and a full-time search engine — and see how different the numbers are. In the underwear store we measured, they differed by 2,810 positions: the store shows 2.4% of products in its own search, but does not give them to the search engine. As long as these two numbers are not present, it is impossible to tell the store that needs to clear the index from the store that needs to fill it. They are opposite jobs, and starting them at the same time means paying for both.

We have a separate page for each size. should it be broken?

Not necessarily. A separate address under "bra 80B" catches the exact request, and this is the real advantage of such a scheme. Its price is a multiplied catalog in which the pages of the same model compete with each other: in the measured store, it is 115,456 items. Our recommendation is to leave separate addresses where there is really a demand for the size in Search Console, and glue the rest with a canonical link to the model. A fair boundary: canonical is a hint, not a command, and the search engine has the right to canonicalize another page. Therefore, we accompany the gluing with a link and a clean site map, and not just the tag itself.

Why are some products not included in the search at all?

Most often, because they are not in the site map, and there are not enough internal links for the bot to reach it by itself. This is verified by the same two calculations. Fair margin: A sitemap doesn't rank or force Google to index a page—it makes it visible to crawlers. A product without a description and without demand can be in the map, be indexed and not receive a single visit. How many addresses are actually in the index is visible only in your Search Console.

How harmful are redirects inside the directory?

The fact that they silently change the product. In our measurement of a sample of 150 product addresses, 50 gave 301, and each led to a different position: 70B to 70A, L to XL, white to black. A person from the search arrives at the wrong size and leaves - exactly at the step where the store has already spent money on it. On the scale of that directory, this is about 37,000 addresses. In the indexing report, they are in the redirect section and don't look like a crash: you can only see the problem when you look at exactly where each redirect leads.

Do you need texts in clothing categories?

Short and to the point - yes, canvases at the bottom of the page - no. What works is what answers the questions before buying: how to choose a size, what is the composition of the fabric, does it sit after washing, how to return it. This is also the best material for separate advisory pages that lead a person to the selection stage. In the agency's own project, a hosiery store, this layer consisted of a blog and seven advisory articles, including a brand size chart, and these were preserved in the domain's archival footprint as separate addresses alongside 7,565 product cards.

We have three language versions and something is wrong with them. what to check

First of all, whether they are wound twice. Finding from the measurement: hreflang is declared by two different mechanisms at the same time, the sets contradict each other, and the link to the Ukrainian version gives Russian content. Plus the language code: there was lang="ua" instead of uk, and on the third version no variables were inserted into the template at all. Fair margin: UI translation does not yet make the language version a separate market. If the names and descriptions of the products remain in the same language, you do not have two audiences, but two pages that share the same traffic.

Will the number of pages in the index drop after your work?

Most likely, yes, and this is the expected course of events. Gluing options and cleaning duplicates remove from the index pages that did not bring visits anyway, but only separated signals among themselves. We fix the figure before the start of work, so that later we do not argue about what exactly has changed. A fair boundary: Cleaning doesn't add new traffic—it stops spraying it. If there is nothing on your pages that your neighbor does not have, cleaning alone will not be enough for growth.

How do you report and how do you understand that there is movement?

We do not name the payback period: it depends on the margin, season and assortment, which we do not control. We give four things every month: the dynamics of crawling and indexing from Search Console, re-measurement of the directory using two methods, a list of what has been done and a plan for the next period. You decide to continue based on these data. We do not promise items for general queries such as the name of the clothing category - networks with different brand and link budgets compete for them.

Give the address of the store - we will calculate the catalog using two methods

In 3-5 working days, we return with two numbers of the directory, a selective report on redirects within it and the status of language versions. If it becomes clear from the numbers that you need more than SEO right now, you will hear it in the first email.

From measured cases115,456 product items by search engine compared to 112,646 unique addresses in the site map

View cases
  • Reply within 2 hours
  • No commitment
  • We work under a contract

There is no price list on the site on purpose: the same work differs several times over between two clients, and a “from” figure explains nothing in that case. First a free audit — we count your pages, duplicates and speed — then we name the sum and the deadline and fix both in the contract.