Parameter handling

Long definition

URL parameters multiply URLs without multiplying content. A category page like /shoes/ exists at one URL conceptually, but in practice browsers may hit it as /shoes/?sort=price, /shoes/?utm_source=email, /shoes/?sessionid=abc123, /shoes/?color=red&size=10, and dozens more. Google's job is to figure out which are duplicates, which are real variants worth indexing, and which to ignore entirely.

Until April 2022, Google offered a manual "URL Parameters" tool in Search Console where you could declare parameter behavior — "sort changes order, doesn't change content" or "trackingid is just for analytics." Google removed it because internal data showed the algorithm was making better decisions than webmasters. Today, parameter handling is fully algorithmic.

The four main parameter categories:

• Tracking — UTM codes, fbclid, gclid, mc_cid. No content change. Google generally ignores and consolidates.
• Sort/order — ?sort=price-asc. Same content, different order. Google may pick one variant as canonical and consolidate.
• Filter/facet — ?color=red. Subset of content. May or may not warrant indexing depending on demand.
• Session/state — ?sessionid=..., ?ref=.... Per-user noise. Google usually consolidates but it's wasteful crawl-side.

What you can still do (and should):

1. rel=canonical to the parameter-stripped version for sort and tracking parameters. The strongest signal you have.
2. Strip tracking parameters server-side before they ever hit the response. Most CDNs and frameworks support this.
3. Use POST instead of GET for filters that genuinely don't need indexing.
4. Use clean URL paths for filter combinations that do warrant indexing — /shoes/red/size-10/ instead of ?color=red&size=10.
5. Configure pagination with self-referential canonicals on each page (not pointing to page 1 — Google deprecated rel=prev/next in 2019).
6. Reduce session-ID exposure with cookies instead of URL parameters where possible.

For ecommerce in particular, faceted navigation is the largest single source of parameter explosion. A 10-facet system can generate millions of URL combinations from a 1,000-product catalog. Without active management — robots.txt blocks on uninteresting facet combinations, canonicals on the rest — Google's parameter handling algorithms still help, but they can't dig you out of a deeply broken URL structure.

Common misconceptions

• "The GSC parameter tool still exists." It was deprecated April 2022. The setting page is gone from Search Console. Don't search for it.
• "Parameter handling is about ranking." It's about indexing efficiency and canonicalization. Done right, your important URLs concentrate signals; done wrong, signals scatter across thousands of near-duplicates.
• "Google ignores all tracking parameters automatically." Mostly yes for well-known ones (UTM, gclid, fbclid). Custom tracking schemes (?ref=internal-promo-123) may be parsed as content variants. Strip server-side or canonical defensively.
• "Robots.txt is the right tool to block parameter URLs." Sometimes, but it has trade-offs. Disallowed URLs may still appear in search as "indexed though blocked" because external links count. For low-value parameter combinations, canonicalization or noindex is usually better than Disallow.