How should we fix duplicate content and canonicals?

This is a technical SEO question — the kind that usually shows up from ecommerce SEO, developers. It rarely has a one-line answer, because the honest version of “How should we fix duplicate content and canonicals” is a shortlist of rival explanations, not a single cause. The job is to work through that shortlist with evidence and stop as soon as one of them is confirmed — not to write a report that mentions all of them.

The rival explanations

Direct answer: Duplicate-content problems come from five distinct sources — indexable facet combinations competing with category pages, conflicting canonical/sitemap/internal-link signals, variant consolidation decisions that need to differ by product type, HTTPS/www/parameter near-duplicates, and Google simply ignoring your canonical hints on the noisiest templates — and a single blanket rule rarely fixes all five.

Treat these as competitors, not a checklist. The point of naming five up front is to stop the first plausible-sounding one from becoming the story before the others have been checked.

• Facet combinations are indexable and competing with category pages.
• Canonical tags say A, sitemaps say B, internal links say C.
• Product variants (colour/size) should consolidate; some unique variants should not.
• HTTPS/www/trailing-slash and parameter order create near-duplicates.
• Google is ignoring our canonical hints on the noisiest templates.

What the evidence has to show

Direct answer: Crawl-based near-duplicate hashing by template, GSC's duplicate/alternate/Google-chose-different-canonical reports, log-based waste analysis on filtered URLs, an internal-link and sitemap membership audit per variant, and revenue-by-canonical-candidate data are what tells you which URL should actually win.

None of the five above survives on a hunch. Here is what actually needs pulling before any of them can be ruled in or out:

• Crawl hashes / near-duplicates; group by template.
• GSC Duplicate / Alternate / Google-chose-different-canonical samples.
• Log waste on filtered URLs.
• Internal-link and sitemap membership of each variant.
• Revenue by canonical candidate (which URL actually converts).

The decision rule

Direct answer: One indexable URL per unique user-intent. Facets that do not change intent are noindex or parameter-blocked. Where Google already chose a canonical and it converts, align the site to Google rather than fighting it.

What to tell the people around you

Direct answer: Engineering needs a written per-template policy and a freeze on new indexable filters — not a one-off fix that leaves the next facet combination to repeat the same problem.

The analysis is not finished until it produces something a non-specialist can act on. That means naming the situation, the cost of getting the first move wrong, and a specific ask — not a summary of the investigation.

• Situation — URL explosion is a policy problem. Every template needs a written rule before another facet ships.
• So what — Unruled facets tax crawl budget forever. The policy is the deliverable; the tickets follow it.
• The ask — A freeze on new indexable filters. Eng time to implement the policy on the two noisiest templates first.

Technical SEO writes policy. Platform team implements. Merch accepts which facets deserve pages.

How to act on this

1. Run a crawl-based near-duplicate hash comparison grouped by template to find where facet combinations are competing with category pages.
2. Pull GSC's duplicate, alternate, and Google-chose-different-canonical reports to see where your own signals and Google's actual choice disagree.
3. Check log-based crawl waste on filtered URLs and each variant's internal-link and sitemap membership.
4. Decide, per template, one indexable URL per unique user intent — facets that don't change intent get noindexed or parameter-blocked.
5. Where Google has already settled on a different canonical that converts, update the site's own signals to match it rather than continuing to fight the mismatch.

Frequently asked questions

What if Google keeps choosing a different canonical than the one we specify?

Check whether the URL Google picked actually converts. If it does, the pragmatic fix is aligning your own canonical signal to match Google's choice rather than repeatedly asserting a preference it keeps overriding.

Should all product variants (color, size) be consolidated onto one canonical?

Not automatically — some variants genuinely serve different search intent (a specific size or color someone searches for directly) and deserve to stay indexable, while others should consolidate. The revenue-by-candidate data is what tells the two apart.

Is noindex or a robots.txt block the right fix for unwanted facet URLs?

It depends on whether you want Google to crawl the URL at all. noindex still allows crawling (and can still leak crawl budget); a robots.txt block stops crawling but won't remove an already-indexed URL. Pick based on what problem you're actually solving.

Do HTTPS/www/trailing-slash variants really create duplicate-content problems?

Yes — inconsistent handling of these creates near-duplicate URLs that split signals, even though the content is identical. A single canonical, redirect, and internal-linking standard resolves this cleanly and permanently.

Who should own the ongoing policy once the initial cleanup is done?

Technical SEO writes the policy and the platform team implements it in code, so new facets follow the rule automatically instead of creating the same duplicate-content problem again on the next template.

Does duplicate content ever result in a manual penalty?

Ordinary near-duplicate or faceted content typically causes a ranking/consolidation issue rather than a manual action — reserve concern about an actual penalty for cases involving deliberate scraping or cross-domain duplication.

Can duplicate content exist across two entirely different domains we own?

Yes, and it's treated the same way — a canonical or 301 pointing to the domain that should own the content resolves cross-domain duplication just as it does within a single site.

Sources

1. Consolidating duplicate URLs with canonical URLs — Search Central
2. Canonicalization — Search Central
3. Duplicate content — Wikipedia
4. Canonical link element — Wikipedia

Where nqzai fits

nqzai runs this same rival-hypothesis framework against your own connected Search Console, Analytics, and audit history, and returns a keep / change / stop decision with the evidence named — including which of the explanations above it could not test, and what to connect to close that gap. No extra cost for the analysis itself; it reads measurements already on file.

Ask nqzai: “How should we fix duplicate content and canonicals?”