URL parameters

Parameters create alternate views of one base page. /hoodies?sort=price, /hoodies?color=red and /hoodies?sort=price&color=red are different URLs.

Filters, sorts and pagination help users. Crawlers see them as separate addresses. That multiplies pages fast and can drown the canonical in duplicates.

Tracking tags like UTM add more combinations. Each click from an email or ad can spawn a fresh copy of /pricing with ?utm_source=newsletter and friends.

You have little crawl budget. Let crawlers spend it on /hoodies and /pricing, not on 200 versions with sort, page and colour toggles.

Duplicate URLs split signals. Links and clicks land on variants, not the clean page. Rankings wobble when crawlers cannot pick the right canonical.

UTM parameters pollute internal links if you let them. That leaks tracking junk into your own crawl space and index as of 2026.

• Run a site: search for question marks: site:yourproduct.com ? in the query. Scan which paths repeat with different keys.
• Export top pages in Search Console Performance. Sort by Page. Look for clusters of the same path with ? and without.
• Open a few parameter URLs in the URL Inspection tool. Check the Canonical chosen by Google and if it matches your preferred clean URL.
• Crawl your site with your tool of choice. Count URLs containing ?. Group by parameter key: sort, page, color, utm_source.
• Check your server logs if you have them. Are bots spending time on ?sort, ?ref or ?utm links instead of the base pages?

A healthy small site has parameter URLs present but not indexed. The clean page holds impressions and clicks for its topic, like /hoodies, not /hoodies?sort=price.

Google’s URL Parameters tool was retired in 2022. You cannot declare parameter behaviour in Search Console now. Use on site controls instead.

• Canonical tag: on parameter URLs, point rel=canonical to the clean URL. Example: /hoodies?sort=price canonical to /hoodies.
• Robots.txt: disallow crawl of stable parameter patterns you never want crawled, like /*?sort= and /*?ref=. Do not block pages you need indexed.
• Noindex: if a parameter view must exist for users but should not be in the index, add meta robots noindex,follow to that template.
• Consistent parameter order: always output keys in the same order, like ?color=red&size=m, to reduce duplicate variants.
• Internal links: never link to parameter URLs in menus, breadcrumbs or body copy. Always link to the clean version.

Fix one template at a time. For a collection page, set canonical to the base, block crawl of ?sort in robots.txt, and remove any internal links that carry ?sort.

• Blocking with robots.txt but keeping internal links. Bots will still try, and the variants will linger in reports.
• Noindexing a URL that is disallowed in robots.txt. If it is disallowed, Google may not see the noindex and the URL can remain known.
• Canonical from a weak variant to another variant. Always set to the clean, crawlable, indexable page.
• Letting UTM leak inside your app. Stripping and redirecting to the clean path avoids split signals.
• Faceted filters creating infinite combos like ?color=red&color=blue&in=stock&price=low. Constrain filters, and expose a few crawlable facets only if they have search demand.