Canonical tags and noindex directives are frequently treated as interchangeable ways to deal with unwanted URLs. They are not, and using the wrong one is among the more common causes of ecommerce indexation problems.
The short version: a canonical says “this content belongs to another URL, consolidate them”. Noindex says “do not include this page in search results at all”. One merges, the other removes.
What each actually does
Canonical
- A hint, not a directive. Google may select a different URL if other signals contradict it.
- Consolidates ranking signals onto the target URL.
- The page stays crawlable and its links continue to be followed.
- Appropriate where two URLs genuinely represent the same thing.
Noindex
- A directive, and it is obeyed.
- Consolidates nothing. Any signals the page had are simply removed from consideration.
- The page is still crawled initially, though it tends to be crawled less over time.
- Appropriate where the page should not appear in results at all.
The critical difference for ecommerce: canonical preserves value, noindex discards it. If a filtered URL has accumulated links, canonicalising it passes that to the category. Noindexing it does not.
Which to use, by case
Filtered category URLs
Canonical to the base category. These are genuinely the same products in a narrower view, and users reach them deliberately. Canonicalising consolidates rather than discards.
Sort orders and view toggles
Canonical to the default view. Same content, different arrangement.
Products in multiple category paths
Canonical to one product URL. This is the textbook canonical case.
Internal search results
Noindex. These are not duplicates of anything, they are generated on demand and infinite in number. There is nothing to consolidate them into.
Thank-you, cart and account pages
Noindex. They should not appear in results and have no canonical equivalent.
Tag archives duplicating categories
Usually noindex, unless the tag has genuine standalone demand, in which case treat it as a real category instead.
Getting it wrong, in both directions
Noindexing what should be canonicalised
The costlier mistake. Noindexing filtered URLs that have accumulated links throws that value away rather than passing it to the category page. The URLs disappear and so does whatever they were contributing.
Canonicalising what should be noindexed
Canonicalising internal search results to a category is meaningless, because they are not the same content. Google is likely to ignore the canonical and index them anyway.
Using both together
Contradictory. A page cannot simultaneously be a duplicate of another and be excluded. Where both are present, behaviour is unpredictable. Pick one.
Canonicalising to a redirecting URL
Breaks the chain. Canonical targets should return 200, not redirect.
The robots.txt interaction
Neither works if the page cannot be crawled. Blocking a URL in robots.txt prevents the crawler seeing the canonical or the noindex, so an already-indexed page can remain indexed indefinitely.
- To remove an indexed page: allow crawling, serve noindex, wait for it to drop.
- To prevent crawling of never-indexed URLs: robots.txt is appropriate.
- Never both at once on a page you want removed.
More detail under noindex tag issues and duplicate without user-selected canonical.
Frequently asked questions
Which is better for filter URLs?
Canonical, in most cases, because filtered views are genuinely the same products and canonicalising consolidates their value into the category. Reserve noindex for URLs with no canonical equivalent.
Does noindex pass link equity?
Not usefully. Links on a noindexed page can still be followed initially, but such pages are crawled progressively less. Do not rely on them as a route for internal linking.
Can I use both to be safe?
No. They contradict each other and the outcome is unpredictable. Decide whether the page is a duplicate or should be excluded, and apply one.
Why is Google ignoring my canonical?
Usually because other signals disagree: internal links pointing elsewhere, sitemap entries, or genuinely different content. Canonicals are a hint, so make every signal agree.
Consulting CTA
If your store has index bloat and you are unsure which URLs to canonicalise and which to exclude, book an SEO consultation for an indexation review.