01What a canonical tag tells Google
A canonical tag is a link element in a page's head, written rel=canonical, that names the address you consider the main one. It matters only when the same content, or very similar content, can be opened at more than one URL. Google's own documentation gives four reasons to name a main address: to choose which URL people see in results, to consolidate signals for duplicate pages, to simplify tracking metrics, and to avoid spending crawl time on copies. If you name nothing, Google picks the version it judges best. The same page also says none of the methods is required and that your site will likely do just fine without one.
- It names the main address among duplicate or very similar pages
- Google calls it a strong signal, and Google can still choose a different address
- Without one, Google chooses the version it thinks is best
- Google's documentation says none of the methods is required
02The one case where a small site needs it
A five-page site of unique pages rarely needs anything. The case that does is a page that must stay reachable at more than one address: the same product list opened with sort or tracking parameters on the end, a printable copy of an article, or one text you publish on two sites. Those are our examples from how sites are built, not a list from Google. If an old address can simply disappear, use a redirect instead. Google calls a redirect a strong signal that the target should become canonical, calls rel=canonical a strong signal as well, and calls a sitemap a weak one. The tag is for when both addresses have to keep working.
- Same page, several live addresses (parameters, print version, shared text): add the tag
- An old address that can go away: redirect it, do not tag it
- Unique pages with one address each: nothing to add
- A sitemap is a weak signal, a redirect and a tag are strong, and the methods can be combined
03How a canonical tag goes wrong
Because Google reads the tag as a strong signal that the named address is the main version, a template that names the home page on every page tells Google that your service pages are copies of it. That is our reasoning from Google's description, not a case we have measured. Other slips are a tag that points at an address that does not exist, and a tag missing from the pages that need it. Google also advises a self-referencing tag on the canonical page itself. None of this repairs thin pages, and none of it moves a site: for a move, redirects do that work. If you suspect your site has this problem, send us the URL. Our SEO page says the technical audit is free and that we tell you what is broken in the first reply.
- One tag in the template naming the home page on every page: the pages read as copies
- A tag pointing at an address that returns an error
- Mixed signals: a redirect says one thing, a tag another
- A tag does not fix thin content or a site move, redirects handle the second
04How to check your own pages
Open one page, view its source and search for canonical. On a page with its own address, the tag should name that same address. Then paste the URL into the inspection bar in Google Search Console. The tool shows a User-declared canonical, which is the tag you wrote, and a Google-selected canonical, which is the address Google chose when it found similar pages. If the two differ on a page you care about, you have a real finding. The Page indexing report also names the problem statuses: Duplicate without user-selected canonical, and Duplicate, Google chose different canonical than user. Start with your service pages.
- View the source of a page and search for canonical
- Search Console URL inspection: compare User-declared canonical with Google-selected canonical
- Page indexing report: look for the two duplicate statuses by name
- A mismatch on a page that brings you enquiries is the one to fix first
05Where PROOF48 is the wrong choice
If your site is a handful of unique pages, each reachable at one address, you may not need us for this at all. Read the source, check Search Console and stop. Our SEO page lists sites that are technically broken, with bad canonicals or duplicate titles, as a good fit, and it lists anyone who needs customers this week, or expects page-one results in under 60 days, as not a good fit. We promise no position, because Google decides that. The page also says we work on sites we did not build, fixing what we can and documenting what we cannot.
- Unique pages with one address each: you probably do not need us for this
- Bad canonicals or duplicate titles: a good fit
- Customers needed this week, or page one in under 60 days: not a good fit
- No position promised, because Google decides that
Redirects first: a salon rebuild where old addresses had to keep leading somewhere
SCP Shine & Style, a beauty salon in Sector 3, Bucharest, was rebuilt and its old WordPress addresses were 301 redirected to the new site, so nothing already indexed broke. That is the redirect half of this article's advice. The case study does not mention canonical tags, and it records no ranking or traffic figure, so we claim neither.
See the build: SCP Shine & StyleQuestions, answered
Does every page need a canonical tag?
No. Google's documentation says none of the methods is required and that a site will likely do fine without one. If you add them, Google recommends a self-referencing tag on the canonical page itself.
Canonical tag or 301 redirect?
If the old address can disappear, redirect it. Use the tag when both addresses must stay live. Google calls both strong signals and says they can be combined.
Will a canonical tag improve my rankings?
It consolidates duplicates and tells Google which address to show. It promises no position, and neither do we.
How do I see which canonical Google chose?
Paste the URL into the inspection bar in Search Console and compare User-declared canonical with Google-selected canonical.
Do you fix this on a site you did not build?
Yes. Our SEO page says we run the audit on whatever stack you are on, fix what we can and document what we cannot.
Where PROOF48 fits
A canonical tag names the main address when the same content lives at several. Most small sites of unique pages do not need one, and Google says a site will likely do fine without. When two addresses must stay live, add the tag; when an old one can go, redirect it. PROOF48's SEO page starts with a free technical audit and promises no position.
Cosmin Tiu · Founder, PROOF48
Cosmin has worked in online marketing and software development since 2018, and has built with AI every day since January 2023. He still writes code from scratch daily, in Python, React, Rust and C#. PROOF48 ships business websites in 48 hours and invoices afterwards, so every opinion on this blog was paid for in real builds, not in theory.