Skip to content

URL parameters

URL parameters are name and value pairs after the question mark; some change what a page shows, some only track, and each combination is another address Google may crawl.

Last updated

On this page

Definition

  • The query string is the part of a URL that starts at the first ? and ends at a # or at the end of the URL (RFC 3986, section 3.4). Google recommends an = between name and value and an & to add further parameters, as in https://example.com/category?category=dresses&sort=low-to-high&sid=789.
  • Some change what the page shows, such as filtering by color and size. Others do not: referral and tracking parameters such as utm_source and gclid, and session identifiers such as sid and PHPSESSID. Google's URL structure guide notes that referral parameters, shopping sort parameters and session IDs can create a large number of URLs.
  • The part of a URL after # (the fragment) is not a parameter: Google says it generally does not support fragments in crawling and indexing, so they have no effect on crawling.
  • Every different combination of names and values is a different URL, even when the page shows the same content.

Why it matters

  • Faceted navigation, meaning filtering and sorting through URL parameters, can in Google's words generate infinite URL spaces. A crawler cannot tell that a new URL is useless until it has crawled it, so it fetches a very large number of them, and has less time left for your new, useful pages.
  • URLs that do not change the content make duplicates: Google's own example is a dress page URL carrying ?gclid=ABCD. Google picks one URL to represent the copies, and is not obliged to pick the one you prefer.
  • A session ID in a URL gives each visitor an address of their own. Google advises avoiding it where possible and using cookies instead.
  • The URL parameters checker and the url-parameters-links rule read a page's links to its own site and report tracking and session parameters from their list. They leave alone parameters that change the content, such as sort, page and color.

Example

Three links to the same page, each a different case:

  • https://shop.example/offers?utm_source=banner: a tracking parameter; the content is the same.
  • https://shop.example/offers?PHPSESSID=7f3a9c1d: a session ID; the content is the same.
  • https://shop.example/offers?sort=price: a different ordering of the listing.

For the first two, remove the parameter from internal links, and give the page a canonical URL without parameters:

html
<link rel="canonical" href="https://shop.example/offers" />

Google treats a redirect and the rel="canonical" annotation as strong signals of the preferred URL, and listing it in a sitemap as a weak one. As for sort=price, if you do not need it to appear in Google Search, the first option Google's faceted navigation guide mentions is blocking its crawling in robots.txt.

Common mistakes

  • Tagging internal links with utm_: each is another address for the same page, and the url-parameters-links rule explains the effect on analytics. Keep utm_ for email, ads and other sites' links.
  • Putting the session ID in the URL instead of a cookie.
  • Using robots.txt to consolidate duplicates: Google says not to, and that a blocked URL may still be indexed without its content.
  • Separating parameters with commas, semicolons or brackets: Google says crawlers find those hard to recognise as separators, and recommends &.
  • Treating rel="canonical" as a command: Google counts it as a signal, and may choose another URL.
  • Treating all parameters alike: those that change the content, such as sorting and the page number, are a different decision from tracking.