Index bloat

Index bloat is a site having many more URLs indexed than pages worth indexing. Typical sources: tag and author archives, filtered and sorted lists, parameters, on-site search results, paginated duplicates, thin generated pages.

On a small site this dilutes crawl and link signals across junk URLs. It slows updates on key pages like /pricing or /guides/getting-started. It also creates duplicate or near-duplicate entries that confuse which page should rank.

Open Performance. Filter by Search type Web. Compare Total impressions and Total clicks to the number of pages that actually drive your business. Then open Pages. You often see thousands of indexed URLs but only a few hundred that earn clicks.

• Performance, Pages tab: sort by Clicks, then by Impressions to see deadweight URLs.
• Pages report, filter by URL contains “?sort=” or “?page=” or “/tag/” to spot templates.
• URL Inspection: test a sample bloated URL to confirm Indexing allowed and the selected canonical.

If your sitemap lists thousands of tag and filter URLs, that is a sign. Check the Sitemaps report to see what you are feeding Google.

• Tags and authors: /blog/tag/javascript, /blog/author/jane.
• Filters and sorts: /shoes?colour=red, /list?sort=price.
• Pagination variants: /category/shoes?page=2, /category/shoes?p=2.
• On-site search: /search?q=crm.
• Tracking parameters: ?utm_source=, ?ref=.
• Programmatic or AI stubs: empty location pages, faceted combos with no stock.

1. 1

   Crawl the site

   Run a crawler and export all URLs. Group by path and by parameters. A quick pivot by “?” in the URL is revealing.

2. 2

   Map URL to template

   For each group, write the template that outputs it: category, tag, search, parameter, pagination, product, article.

3. 3

   Decide intent

   Ask one question per template: should this rank? If not, it should not be indexed. If yes, it needs a clean canonical and a place in the sitemap.

Pick one control per template, then keep the sitemap to only pages meant to rank. Examples below show a fixed page that keeps a single clean URL indexed.

• Tags and authors: add meta robots noindex, follow. Remove them from the sitemap. Example: /blog/tag/javascript now noindex, posts still crawlable.
• Filters and sorts: set a canonical to the unfiltered list. Keep them indexable only if they have unique copy and demand. Example: /shoes?colour=red canonical to /shoes.
• Pagination: page 1 indexable, later pages either noindex or canonical to page 1 if they are near-duplicates. Link to page 2, 3 for users.
• On-site search: disallow crawl in robots.txt if infinite, and add meta noindex to results. Never include /search in the sitemap.
• Tracking parameters: always canonical to the clean URL. Strip UTM parameters server-side where possible.
• Thin stubs: remove, or 410. If needed for users, keep, but add noindex until it has content.

After changes, resubmit the sitemap that lists only target pages. In URL Inspection, Request indexing on a few representative URLs to nudge the recrawl. Watch the Pages and Performance reports over 28 days.

• Lock rules in code: template-level meta robots and canonicals.
• Validate on deploy: check for “?” in new internal links.
• Keep one sitemap or a small sitemap index for rankable pages only.
• Review Performance quarterly: pages with zero clicks and zero impressions are candidates for noindex or removal.
• Avoid linking to facets you do not intend to index. Use buttons or forms without crawlable hrefs.

A healthy small site might have 200 pages in the sitemap and, typically, a similar number indexed, not 8,000. /pricing, /features, /guides/… should dominate clicks and impressions.