Skip to content

Case study · Ecommerce

An online store that stopped wasting its crawl

Cut indexed filter pages by 80 percent and lifted indexed product pages from 48 to 91 percent.

Engagement: 5 monthsServices:Technical SEOContent and on-page SEO

Who: An independent online store with about 4,000 products, a small in-house team and one part-time developer.

Products indexed
91%from 48%
Organic orders a month
340from 210
Filter pages in the index
11,800from 60,000
Crawl requests spent on filter pages
18%from 71%
Variant pages with the correct canonical
100%from 35%
Top 100 products with original descriptions
100from 12

What changed

The store's filters could be combined in thousands of ways, and each combination created a page that search engines found. Google spent its attention on sixty thousand near-identical pages and left half the real products unindexed.

We gave the developer a ranked list of changes in plain language: how filters should be handled, which variant pages should point where, and which product pages needed their own words rather than the manufacturer's.

Over five months the filter pages in the index fell from sixty thousand to under twelve thousand, and indexed products rose from 48 to 91 percent. Organic orders climbed by more than half.

Products indexed, percent
  • Start48
  • Month 157
  • Month 269
  • Month 482
  • Month 591

The situation

A store with 4,000 products had 60,000 filter pages indexed and half its products missing from Google.

What they wanted

  • Get every product that is for sale into Google
  • Stop search engines spending their time on filter pages
  • Give the developer a short, clear list of what to change

What we found

  • Filters created endless near-duplicate pages
  • Canonicals pointed the wrong way on variants
  • Product pages repeated manufacturer text

What we did

  • Fixed filter and canonical handling
  • Rewrote the top 100 product pages
  • Added product markup that passes validation

How the engagement ran

  1. Weeks 1 to 3

    1. Find where the crawl goes

    We read the server logs and the index report together to see how much of Google's attention went to filters, variants and real products.

  2. Weeks 4 to 10

    2. Fix the rules

    Filter handling, canonicals and the sitemap were changed in small releases, each checked before the next. Nothing went live without the developer's yes.

  3. Weeks 11 to 18

    3. Rewrite what earns

    The hundred products with the most search demand got their own descriptions, sizing notes and answers to the questions buyers ask.

  4. Weeks 19 to 22

    4. Mark up and measure

    Product markup was added and validated, and we compared indexing, crawl share and orders with the start.

What surprised us

  • A single site setting was creating most of the duplicate filter pages
  • Rewriting the top hundred products mattered more than touching all four thousand

What they keep

  • A ranked list of fixes written for the developer
  • Rewritten copy for the hundred best products
  • A crawl and indexing report kept up to date weekly

How we measured it

Indexing comes from the search engine's own index report, checked weekly against the product list. Crawl share comes from the server logs. Orders are those the store's analytics attributes to organic search, compared with the same months a year earlier to allow for season.

The report told us what was wrong in language our developer could act on the same day.
Head of ecommerce, online store

What is next: Category page rewrites and a review programme.

Ready to be found by Google and named by AI assistants?

Book a 30-minute call. We prepare a free audit of your site first, so you leave with a written read whether or not you hire us.