Skip to content
Get Ranked On TopIllustrative case studyCommerce SEO plan

A fictional business. The figures are a modelled scenario, not a client result.

SOMBRA PERSIANAS · MADE-TO-MEASURE BLINDS & SHUTTERS SHOP · MONTERREY

How a Monterrey blinds shop closed 3.6 million variant URLs and let Google’s light reach its categories

+131%

organic orders a month, once the blinds came down on 3.6 million variant URLs

Monterrey, Mexico · Home & real estate · 14 months · Commerce SEO plan

Further down, tilt the slats yourself: the whole engagement fits in one control.

THE PROBLEM

Two thousand blinds. Three and a half million doors.

Every colour and width combination was an indexable URL, so Google crawled millions of variants while the “persianas enrollables” category sat on page four.

Sombra Persianas makes roller blinds, romans and shutters to measure in Monterrey and ships them across Mexico. Its configurator was a pleasure for customers: pick a fabric and a colour, type your width and height to the centimetre. For Google it was a maze.

Every choice was a link, every link a new URL, and none of them said which page was the real one. The category pages that should have ranked for “persianas enrollables” sat on page four, behind thousands of near-identical variants competing with each other.

WHAT THE CRAWL AUDIT FOUND

  1. 3.6M

    Crawlable URLs

    About 2,000 sellable configurations (model, fabric, colour) multiplied by width and height parameters into 3.6 million URLs Google could follow.

  2. 91%

    Crawl spent on variants

    Twelve weeks of server logs: 91% of Googlebot requests hit width, height, colour, sort and session parameters. Categories and products shared the other 9%.

  3. 41 days

    Between category recrawls

    The 46 category pages were fetched on average once every 41 days, so new fabrics and prices took weeks to reach the results.

  4. 0

    Canonical tags

    No canonical, no noindex, nothing in robots.txt. For “persianas enrollables” Google chose its own canonical: a ?ancho=120&color=blanco variant.

WHERE GOOGLEBOT WENT

12 weeks before80%11%Month 1424%43%28%
  • Width, height & colour variants
  • Sort & session URLs
  • Categories & landing pages
  • Product pages

THE FIX

Close what wastes light. Open what sells.

Built on our Commerce SEO plan: ecommerce SEO for catalogues where faceted navigation, crawl budget and product data decide who ranks.

  1. 01 · FACET INDEXATION RULES

    7 open · 5 closed

    Twelve URL families, one decision each

    We scored every family on search demand, uniqueness and stock. Seven earned an indexable page; five became variants of a page that already existed. The rules live in the platform, not in a spreadsheet.

  2. 02 · CANONICALS & NOINDEX

    3.6M → 1,180 indexable

    Every variant points home

    Width and height variants canonicalise to their product, sort orders carry noindex, follow, and combination and session parameters were blocked in robots.txt only after the links to them were gone.

  3. 03 · CATEGORY PAGE REBUILD

    46 + 214 pages

    Categories worth the crawl

    The 46 categories and 214 landing pages got their own copy, measuring guides, FAQs and internal links, so the light that now reaches them finds something to rank.

  4. 04 · PRODUCT SCHEMA & MERCHANT FEED

    920 products

    Every blind in the shopping results

    Product and Offer schema with honest “from” prices, plus a Merchant Center feed for 920 products, so made-to-measure items show a price before the click.

TILT THE SLATS

Each slat is a family of URLs. You decide which ones Google sees.

An open slat is indexable. A closed one is canonicalised, noindexed or blocked. The light that gets through is Googlebot’s crawl: watch where it lands, and what it sells.

Tap a slat to tilt it. Shift the presets to compare.

WHERE THE LIGHT LANDS

  • Pages that sell · 9%
  • Variants & parameters · 91%
Indexable URLs
3.6M
Crawl on pages that sell
9%
Searches with a landing page
23,200
Organic orders / mo
2100%

MODELLED

5 open slats leak crawl into variants nobody searches for.

MONTH BY MONTH

Fourteen months, slat by slat.

  1. MONTH 1SEPT

    Crawl and log audit

    A full crawl stopped at 3.6 million URLs; twelve weeks of server logs showed where Googlebot really went.

    Crawl on pages that sell
    10%
    Orders
    208
  2. MONTH 2OCT

    Twelve URL families

    Every URL sorted into one of twelve patterns: categories, products, five facets worth a page, five mere variants.

    Crawl on pages that sell
    14%
    Orders
    214
  3. MONTH 3NOV

    Facet demand map

    Demand checked per facet: “persianas blackout” gets 5,400 searches a month, a 137 cm width gets none.

    Crawl on pages that sell
    26%
    Orders
    226
  4. MONTH 4DEC

    Canonicals live

    Width and height variants point to their product with rel=canonical; sort orders get noindex, follow.

    Crawl on pages that sell
    39%
    Orders
    249
  5. MONTH 5JAN

    Filters without links

    Size and colour filters became form state instead of links; combination and session parameters blocked in robots.txt.

    Crawl on pages that sell
    50%
    Orders
    281
  6. MONTH 6FEB

    214 landing pages

    Facets with real demand got clean static paths, unique copy and their filters pre-selected.

    Crawl on pages that sell
    57%
    Orders
    309
  7. MONTH 7MAR

    Category rebuild

    46 category pages gained buying guides, a measuring how-to and FAQs written with the Monterrey fitting team.

    Crawl on pages that sell
    61%
    Orders
    338
  8. MONTH 8APR

    Clean sitemaps

    XML sitemaps list only the 1,180 indexable URLs, with real lastmod dates.

    Crawl on pages that sell
    64%
    Orders
    362
  9. MONTH 9MAY

    Product schema

    Product and Offer markup with “from” prices on 920 product pages; reviews marked up where they exist.

    Crawl on pages that sell
    66%
    Orders
    391
  10. MONTH 10JUN

    Merchant feed

    920 products in a Merchant Center feed, so the blinds also show in free shopping listings.

    Crawl on pages that sell
    68%
    Orders
    418
  11. MONTH 11JUL

    Internal links

    Breadcrumbs and “popular in Monterrey” modules send crawl and shoppers to the landing pages.

    Crawl on pages that sell
    69%
    Orders
    440
  12. MONTH 12AUG

    Duplicates fade

    Search Console’s duplicate report falls below 2,000 URLs, from more than 400,000.

    Crawl on pages that sell
    70%
    Orders
    459
  13. MONTH 13SEPT

    Faster filters

    Filter JavaScript deferred; mobile LCP on category pages down to 2.1 s.

    Crawl on pages that sell
    71%
    Orders
    474
  14. MONTH 14OCT

    Measured

    Head term in position 3, 71% of crawl on pages that sell, 486 organic orders in the month.

    Crawl on pages that sell
    71%
    Orders
    486

RESULTS

More light on fewer pages, and the orders followed.

+131%

Organic orders / month
200300400500486SeptJanMayOct

34 → 3

"persianas enrollables" position
1102030343persianas enrollables

9% → 71%

Googlebot hits on pages that sell
9%Before71%After

3.6M → 1,180

Indexable URLs
Before3.6MAfter1,180

Modelled scenario for a fictional business — not a client result. Rankings and traffic are never guaranteed; see our Terms.

QUERIES THAT NOW LAND ON AN OPEN SLAT

Positions for Spanish search queries in Mexico, before the work and in month 14
Query (as searched)Searches / moMonth 0Month 14Change
persianas enrollables6,600343up 31 places
persianas blackout5,400414up 37 places
persianas a la medida4,400275up 22 places
persianas screen3,600382up 36 places
persianas monterrey2,400192up 17 places
persianas motorizadas1,600526up 46 places
persianas para recámara880637up 56 places
persianas grises720705up 65 places

YOUR CATALOGUE

Is Google crawling your filters instead of your categories?

We map every URL family in your shop, decide which ones deserve the light, and ship the rules with your developers.

Back to all case studies