Operations
Part of Ecommerce SEO: Fix a Stale Shipping Promise
Ecommerce SEO best practices for filter and variant URLs
Ecommerce SEO best practices for filter and variant URLs: robots.txt, noindex, canonicals, pagination and redirects that keep a store's URL space small.
What to take away
- Block low-value filter and sort URLs in robots.txt, and keep the useful combinations as landing pages.
- Canonicalize variants that add nothing to the preferred product page, and use noindex only on pages that stay crawlable.
- Give each page in a series its own crawlable URL and a self-referencing canonical.
- Redirect a moved URL directly with a 301, and return 404 or 410 when nothing replaces it.
Filters, variants and paginated lists can each multiply URLs faster than the catalog grows. Ecommerce SEO practices for them are rules about which addresses exist, which one is preferred, and what happens when one disappears.
Decide which filter and sort URLs may exist
A filter URL earns a place only when it matches a real selection task. Expose valuable facet combinations as landing pages, and block the rest so they cannot create endless crawl paths. Keep tracking parameters out of preferred URLs and use one parameter order.
Google's guidance on faceted navigation says that if you do not need those URLs indexed, use robots.txt to disallow crawling of them. Its reasoning is that crawling filtered pages costs server resources for negligible benefit, so allow the individual product pages and one unfiltered listing page.
Sort order, sort direction and view switches, such as sort= or view= parameters, rearrange or restyle the same products. Treat them as their own URL type and keep them out of crawl paths with the same robots.txt rule.
| URL type | Rule | Check |
|---|---|---|
| Valuable facet combination | Its own landing page with a clear purpose | Linked internally and listed in the sitemap |
| Low-value filter | Disallowed in robots.txt | A crawl shows no runaway URL growth |
| Sort or view parameter | Disallowed in robots.txt | No sort= URL in the sitemap or internal links |
| Tracking parameter | Absent from preferred URLs | Canonical points to the clean address |
| Parameter order | One fixed order | The same filters always give the same URL |
| Internal ID or session value | Not part of the path | Paths read as plain words |
A filter built on a URL fragment, the part after the hash sign, has no effect on crawling, because Google Search generally does not support fragments in crawling and indexing.
Noindex and robots.txt do different jobs. A blocked page is never fetched, so it can never show its noindex tag. Put a robots meta noindex on pages that stay crawlable but should not appear in results, and do not stack the two on one URL.
Standardize paths in lowercase with readable words separated by hyphens, and avoid internal IDs, repeated keywords and session values. Google's documentation on URL structure warns that poor structures can create inefficient or infinite URL spaces. Check each new pattern against the ecommerce SEO checklist before a template ships.
Handle variants and pagination with opposite rules
Variants and pagination need opposite rules. A variant gets its own URL only when it serves a distinct customer or search need, while every page of a series keeps its own address. Work through the decision in order.
- Ask whether shoppers or searchers need this variant as its own page.
- If it does, give it a variant-specific URL.
- If it does not, canonicalize it to the preferred product page and keep selection usable.
- Give each paginated page a distinct crawlable URL and a self-referencing canonical.
- Never canonicalize every page in a series to the first page.
Use rel="canonical" for genuine duplicates, such as a variant that adds nothing of its own. Align canonicals with internal links and the XML sitemap, and list only preferred, indexable URLs in the sitemap so the signals agree.
Google no longer uses rel="next" and rel="prev" as an indexing signal. Ordinary crawlable links between the pages of a series, each with its own URL, do the work instead.
For localized equivalents, use rel="alternate" hreflang with reciprocal annotations and a canonical suited to each locale. Keep country-specific URLs limited to the markets the store serves.
Retire and move URLs without breaking them
When a URL must change, send a direct HTTP 301 redirect to a relevant replacement. Do not send unrelated retired products to the home page. A shopper looking for a discontinued item should land on the page that helps them most.
Federal web guidance on Digital.gov says to use 301 redirects to send searchers from the old page to the new page. Its guidance on decommissioning sites adds that you should implement redirects so search engines and bookmarked visitors do not get 404 errors.
When nothing replaces the page, return 404 or 410. Use 410 when the removal is known to be permanent, and 404 when it is not. Watch for discontinued products left as broken links.
Test a failure case before release, such as an unintended filter URL or a discontinued product, and document how the team will correct it. Review the redirect and status rules after platform releases, catalog changes, international expansion and acquisitions, using the signals in ecommerce SEO mistakes.
Test each URL pattern before release
Choose representative product, category, editorial, cart and policy templates, and check the URL each one produces. In Search Console, the URL Inspection tool shows whether Google can crawl a given URL and which canonical it selected.
Performance Review Checklist
- Representative product template
- Representative category template
- Representative editorial template
- Representative cart template
- Representative policy template
- Large images and galleries
- Third-party tags and consent states
- Weak networks and personalization
Search Console Help explains the Core Web Vitals report as field data grouped by similar URLs and status categories. Use it to find groups needing work, then test single URLs with diagnostic tools. Field coverage can be unavailable for low-volume groups.
Give every URL rule an owner
Each URL rule needs one accountable owner, even when other roles contribute. The SEO lead owns URL standards and indexation rules, approves exceptions and resolves conflicts between organic search requirements and commercial requests.
Developers implement the rules, and merchandisers decide names, variants, availability and retirement. An analytics or data owner keeps the benchmark definitions. When a page group cannot meet the standard, assign an exception owner and an expiration date, since permanent undocumented exceptions become new platform defaults.
Run checks at catalog imports, template or routing releases, locale launches and product retirements. Record the expected URL, canonical, status and structured data beside the result. A plan built around focused business priorities decides which page groups are checked first.
Common questions
Should every filter combination be indexable?
No. Only combinations that serve a real selection task should exist as landing pages; the rest stay out of the sitemap and internal links.
Do variants always need their own URL?
No. Only variants with a distinct purpose get their own URL, and the rest are canonicalized to the preferred product page.
Where should a retired product redirect?
To a relevant replacement with a direct 301. If there is no replacement, return 404 or 410, not the home page.
