Duplicate Content and Canonical Tags, Explained
There is no duplicate content penalty. There is something worse and less discussed: your own pages quietly competing with each other and splitting the signals that should have concentrated on one.
Google does not penalise duplicate content, but it does have to choose one version to rank, and it splits ranking signals between the duplicates while deciding. A canonical tag tells Google which version you consider authoritative, consolidating those signals onto one URL.
The Penalty That Doesn't Exist
"Duplicate content penalty" is one of the most persistent myths in SEO. Google has stated repeatedly that ordinary duplication is not penalised — it is an extremely common, usually accidental technical condition, and penalising it would break most of the web.
What actually happens is quieter. When Google finds several URLs with substantially the same content, it picks one to index and ranks that. The others are filtered out. If the version it picks is not the one you wanted, you have lost nothing to a penalty and everything to a coin flip.
The real cost is dilution. Links, internal links and engagement that should have accumulated on one URL are spread across several. Each duplicate is individually weaker than the single consolidated page would have been, which is why fixing duplication often produces an immediate, visible improvement.
The exception is content duplicated deliberately at scale to manipulate rankings — scraped pages, or hundreds of near-identical location pages with only a place name swapped. That is a spam problem, and it is treated as one.
Where Duplication Comes From
Almost none of this is deliberate. These are the ordinary technical conditions that produce multiple URLs for one page, usually without anyone noticing.
Protocol and host variants
http and https, www and non-www. Four possible addresses for every page. Each should permanently redirect to one chosen form — this is a server configuration rather than a content problem, and it is the first thing to settle.
Trailing slashes
/services and /services/ are different URLs to a crawler even when they render identically. Pick one convention, redirect the other, and make sure internal links use the chosen form consistently.
URL parameters
Tracking codes, filters and sort orders generate endless variants of the same page. Common on e-commerce, where filtering can produce thousands of crawlable URLs for one category. Canonical tags are the usual answer.
Printer-friendly and AMP versions
An alternate rendering of the same content at its own URL. It should canonicalise to the main version rather than compete with it.
Boilerplate across thin pages
Location or service pages that share ninety percent of their text with only a name changed. Not duplication in the technical sense, but it produces the same problem — several pages none of which is distinctly about anything.
Using Canonical Tags Correctly
A canonical tag is a hint, not a command — Google can disregard it if the evidence contradicts it. Getting these four things right means it usually will not.
- 01
Every page declares itself canonical by default
The normal case is a self-referencing canonical: the page points at its own preferred URL. This settles parameter and slash variants automatically, since each variant declares the same canonical form.
- 02
Use absolute URLs
Full URLs including protocol and domain. Relative canonicals are resolved against the current page, which produces the wrong target in exactly the situations where the tag was supposed to help.
- 03
Point at a page that actually exists and is indexable
A canonical pointing at a redirect, a 404 or a noindexed page is a contradiction and gets ignored. Verify the target returns 200 and is itself canonical.
- 04
Never let a template hardcode one canonical site-wide
The classic catastrophic bug: every page declaring the homepage as canonical, telling Google the entire site is one page. The site looks perfect to visitors and collapses to a single indexed URL. Check by viewing source on three unrelated pages.
Canonical, Redirect or Noindex?
Three tools that overlap enough to be confused. Choosing wrongly either fails to fix the problem or removes a page you wanted.
| Use when | Effect | |
|---|---|---|
| Canonical tag | Both URLs should stay reachable, but one is authoritative | Consolidates signals onto the canonical; both remain accessible to visitors |
| 301 redirect | The old URL should no longer exist | Passes nearly all ranking value and sends visitors to the new page |
| noindex | The page is useful to visitors but should not be in search | Removes it from the index; it can still be crawled and pass internal link value |
| Nothing | The pages are genuinely different | Two pages on related but distinct topics are not duplication and need no intervention |
Common Questions
What comes up when diagnosing duplication.
Is Google ignoring my canonical tags?
It can, if the evidence disagrees — internal links pointing at the other version, or the canonical target being noticeably thinner. Search Console's URL Inspection shows the canonical you declared and the one Google actually selected, which is the fastest way to spot a disagreement.
Does republishing my content elsewhere hurt me?
Syndication is fine when the copy points its canonical back at your original. Without that, the larger site's version may be the one indexed. Agree the canonical before syndicating, not after.
How similar is too similar for location pages?
If swapping the place name would make one page read as the other, they are too similar. Genuinely local content — the areas covered, local specifics, real examples — is what makes a location page a page rather than a template fill.
Do I need canonicals on a small site?
Self-referencing canonicals are cheap insurance on any size of site, and they cost nothing once the template does it automatically. They also handle tracking parameters, which appear the moment you run any campaign.