Audit a canonical URL inventory

Preserve distinctions until you have evidence to consolidate them.

Start from the URLs you actually serve

Combine a public sitemap with known navigation routes and relevant Search Console findings. Keep the source of each URL. A list assembled only from guessed keyword combinations is not an inventory of the live site.

Do not normalize away meaning

Hostnames are case-insensitive, but paths may be case-sensitive. A trailing slash or query parameter can alter application behavior. The checker intentionally keeps these distinctions. Inspect redirects and content before deciding that two addresses should become one.

Choose one preferred address

For genuinely equivalent public pages, choose a preferred URL based on the site’s existing structure and user needs. Keep internal links and sitemap entries consistent with it. Use a suitable redirect for retired equivalents and a canonical annotation where appropriate; verify the live result.

Treat canonical signals as a preference

Google describes redirects and rel=canonical annotations as strong signals and sitemap inclusion as a weaker signal. They express a preference; they do not force Google to select that URL. Keep the signals consistent, and check the selected canonical in Search Console once the page has been processed.

Do not substitute an indexing block

A robots.txt disallow is not a canonicalization method: Google may still index the URL without its content. A noindex directive removes a page from Search; it is not the recommended way to choose between equivalent public pages on one site. Keep genuine private or excluded pages out of Search for their own reasons, and use the canonical mechanisms for duplicates.

Record the change and watch for mistakes

Keep a mapping from old to preferred URLs, the reason for each change and the verification result. Avoid redirecting every missing URL to the homepage: visitors need a meaningful destination or a real not-found response. Recheck important inbound paths after deployment.

When the signals disagree

Disagreement usually looks like this: the sitemap lists one address, internal links point to a second, and a legacy redirect lands on a third. Work through it in a fixed order. First, establish what each address serves today — status code and content — because a signal pointing at a redirect or an error is noise, not a vote. Second, pick the preferred URL on structural grounds: the address your navigation and templates already treat as primary. Third, align the cheap signals — sitemap entries and internal links — in the same change as the canonical annotation, so no crawl ever sees a mixed state. Fourth, verify after processing: compare the user-declared and Google-selected canonicals in Search Console, and keep the old-to-preferred mapping with the reason for the choice. Most canonical problems are really record-keeping problems.