Technical SEO · guide 4 of 7
Duplicate content and canonicals: what Google really does with copies
Duplicate content is usually not a penalty. Learn what Google actually does with copies, where they come from on your own site, and when to use a canonical tag or a redirect.
26 min readUpdated 9 Nov 2026
- 11 questions to test yourself
- 1 calculator
- 12-item checklist
On this page9 sections
Is Google punishing you for duplicate content? Almost certainly not
An audit tool came back red: “142 pages with duplicate content”. Or a supplier warned that the manufacturer’s product text will get you “penalised”. Or you found your own blog post, word for word, on a site you have never heard of. Is Google quietly holding this against you?
Almost certainly not. Google’s Search Central documentation says duplicate content on a site is not grounds for action unless it looks meant to deceive people and manipulate results. A shop that shows one product in two categories is not doing that.
But “no penalty” is not “no cost”. When Google finds copies, it picks one to show and files the rest away. If it picks the wrong one, or your page’s strength is split between copies, you lose out with no warning message. By the end you will know where copies come from, which tool fixes which kind, and how to check your own site in about 20 minutes.
One more thing to watch for on the way: one of the most common ways owners “fix” duplicates makes the problem worse. We will come back to it near the end, once you can see why. First, test what you already believe.
Myth or fact?
Five things owners hear about duplicate content
1.If two pages on my site share the same text, Google will penalise the whole site.
2.Google may show a different version of my page than the one I want, even if I have marked my favourite.
3.Google can tell you, for any page, which version it chose as the main one.
4.Copying large amounts of other people’s content to fill a site can get it into trouble.
5.Using the manufacturer’s description word for word puts your shop at risk of a Google penalty.
What Google actually does with copies: it picks one
When Google finds several addresses showing the same or almost the same content, it groups them together and chooses one as the canonical: the main version, the one that can appear in search results. The others become duplicates. They are still known to Google, but people searching usually see only the chosen one.
This is called canonicalisation, and it is mostly helpful. Nobody wants ten identical results for the same page. The trouble starts in two places.
First, Google may choose an address you would not have chosen, such as one with a tracking code on the end, or an old http version. Second, the signals that help a page, especially links from other sites, can be spread across several copies instead of all pointing at the one you care about. Google tries to gather those signals onto the canonical, but it can only do that cleanly when the copies are clearly copies.
So the goal is not to “avoid a penalty”. It is to make Google’s choice easy and to make sure it chooses your version.
Make a guess
Take one inner page, say your prices page. Counting only http or https, with or without www, and with or without a slash at the end, how many different addresses could it be reached at?
Where duplicates come from (most of them you never made on purpose)
Owners picture duplicate content as someone copying and pasting text. In practice, most copies are created by the website software itself, without anyone noticing. These are the usual sources on a small-business site:
- Host and protocol versions: http and https, www and no www, all loading the same pages instead of redirecting to one.
- Trailing slashes and capital letters: /prices and /prices/, or /Prices, served as separate pages.
- Tracking and session codes: ?utm_source=facebook or ?sessionid=123 added by ads, emails or the shop system.
- Shop filters and sorting: ?colour=red&size=42&sort=price creates a new address for every combination (our guide to filters and faceted navigation covers this in depth).
- The same product in several categories: /shoes/running/trail-x and /sale/trail-x showing one product.
- Printer-friendly versions, AMP pages and separate mobile addresses such as m.example.com.
- Near-identical location pages: “Plumber in Varna”, “Plumber in Burgas”, with only the city name changed.
- Staging or test copies of the site that were left open to Google, such as dev.example.com.
- Your content published elsewhere: articles you syndicated to a news site, or pages copied by scrapers.
Quick quiz
Your shoe shop’s main category page is https://example.com/shoes/. Which of these is most likely NOT a duplicate of it?
Pick one. Your first answer is the one that counts.
Try the numbers
How many addresses can your shop filters create?
Possible filtered addresses
2,520 addresses
Addresses for each page you actually want
252 to 1
Simple multiplication from your own numbers, not a measurement of your site. The +1 counts “any” for each filter, and real shops often have more filters, so the true number is usually higher.
What your result means depends on its size. A few hundred filtered addresses with canonical tags pointing at the clean category page is a tidy, normal shop: Google reads each copy, sees the tag and merges it (step 3 of the check below shows you how to confirm it).
Thousands or tens of thousands is a different problem. Google still has to fetch every address to read its canonical tag, so the tags tidy up what is shown but do little about the crawling. That is where Google’s faceted navigation documentation suggests deliberate robots.txt rules for the filters you never want in search. For a worked example, read the Ambra Forno study on our Case studies page: an illustrative, made-up Bologna bakery shop whose filters produced 40,000 addresses, of which 24 earned their own indexable page. Our guide to filters and faceted navigation walks through that decision step by step.
Canonical tags: how to name the main version
A canonical tag is one line in the hidden head section of a page that says “the main version of this page lives at this address”. On the copy at /shoes/?sort=price-low it points to https://example.com/shoes/. On the main page itself it points to its own address, which is called a self-referencing canonical and is perfectly normal.
Most website platforms (WordPress with an SEO plugin, Shopify, Wix, Squarespace) add canonical tags automatically. Your job is usually not to write them, but to check they say what you mean.
Three things matter. Use the full address, including https:// and your domain, as Google recommends. Use one canonical tag per page; two conflicting ones are a common template bug, and Google may then ignore both. And remember it is a strong hint, not an order: Google weighs it against your redirects, your internal links and your sitemap. If your menu links to /shoes, your sitemap lists /shoes/ and your canonical says /Shoes/, you are asking Google to guess.
Before and after
One page, two sets of signals
Mixed signals
Canonical tag: https://example.com/prices
Menu link: https://www.example.com/prices/
Sitemap: http://example.com/prices
Search Console: “Duplicate, Google chose different canonical than user”
Every signal agrees
Canonical tag: https://example.com/prices
Menu link: https://example.com/prices
Sitemap: https://example.com/prices
Other versions redirect permanently to https://example.com/prices
- Google picks a canonical by looking at all the evidence, not just the tag. When the tag, the links and the sitemap name the same address, there is nothing to resolve.
- This is an illustration with a made-up site, but the Search Console status is real and appears in the Page indexing report.
Quick quiz
A developer set every page on your site to have a canonical tag pointing to your homepage, “to concentrate the strength”. What is the likely effect?
Pick one. Your first answer is the one that counts.
This happens by accident more often than on purpose, usually through a template that fills in the homepage as a default. The Wattlebark Pet Co. study on our Case studies page shows it in an online pet-food shop: the server pointed every canonical at the homepage and the correct address was filled in by JavaScript only later, so Google merged hundreds of products into one address. The business and its figures are illustrative, but the mistake is a real and common one. The quickest way to catch it on your site is in the 20-minute check below.
Redirects: when a copy should simply stop existing
A canonical tag keeps the copy reachable and points Google to the original. A redirect goes further: anyone, person or Google, who asks for the old address is sent straight to the new one. Google treats a permanent redirect (a 301 or 308) as a strong signal that the destination should be the canonical. A temporary one (302 or 307) is a weaker signal, because it says the move might be undone.
Use a redirect when nobody needs the copy: the http version of your site, the www or non-www version you are not using, an old page you rewrote under a new address, a product you merged into another. Use a canonical tag when the copy has to keep working for people: a sorted or filtered view, a tracking link, the same product in two categories.
Another tool often gets mixed into this, and it does not do the job. A noindex tag keeps a page out of the results, but it does not tell Google which page is the original, and Google’s documentation advises against using noindex to choose a canonical within one site. A third tool, robots.txt, is the most tempting of all; it gets its own warning in the last part of this guide.
Side by side
| Tool | Copy still opens? | Names the main one? | Use it for |
|---|---|---|---|
| Permanent redirect (301) | No, visitors move on | Yes, strongly | http to https, www to non-www, old or merged pages |
| Canonical tag | Yes | Yes, as a strong hint | Sorting, filters, tracking codes, products in two categories |
| Noindex tag | Yes | No | Pages that should not be in Google, such as thank-you pages |
| Robots.txt block | Yes | No, Google can’t read it | Planned crawl rules for huge filter sets, never as a duplicate fix (see below) |
What would you do?
Two articles that say the same thing
This is an illustrative example. A garden centre wrote “How to plant tulip bulbs” in 2021 and, forgetting about it, “Tulip planting guide” in 2024. Both are about 600 words and give the same advice. Neither shows up well in search, and a gardening blog links to the older one.
What is the best fix?
Copies on other sites: scrapers, partners and manufacturers
Finding your article copied word for word on a stranger’s site is upsetting, but it is rarely a ranking emergency. Google usually shows the original rather than the copy, though not always. If a copy does outrank you and it uses your work without permission, you can file a copyright removal request through Google’s legal help pages.
Quick quiz
A regional news site asks to republish one of your articles. What should you ask them to do with their copy?
Pick one. Your first answer is the one that counts.
If you syndicate, put that request in writing before you hand the article over.
Manufacturer descriptions are the third case. They are not a penalty risk, but hundreds of shops may carry exactly the same paragraph, so Google has no reason to prefer your page over theirs. The fix is to add what only you can add: your own photos, the question customers always ask you in the shop, sizing advice, what it pairs with. Our guide to product pages shows how to write a description only your shop could write, starting with the products that matter most.
How to check your own site in 20 minutes
You do not need an expensive crawler to find the duplicates that matter. Google Search Console, which is free, shows which pages Google treats as copies and which version it chose. Work through these four steps with a cup of coffee, then use the cards: for each status, guess first whether to leave it alone or fix it, and flip to check.
Step by step
A 20-minute duplicate check
Step 1 of 4
Try every version of your address
Type your domain into the browser in its variants: http and https, with and without www. Every version should land on the same single https address. Then try one inner page with and without a slash at the end. If two versions load the page without the address bar changing, you have a redirect to fix.
Our free domain health check runs this test for you: it shows where the http and www versions of your domain end up, so you can see a missing redirect in seconds.
Test yourself
Think of your answer, then flip the card.0 of 4 flipped
The “fix” that makes duplicates worse (and four more to avoid)
Here is the mistake promised at the start. Because “duplicate content” sounds alarming, owners and developers often reach for the quickest-looking fix. See if you can spot it before reading on.
What would you do?
The audit says 142 duplicates
This is an illustrative example. A shoe shop’s audit report lists 142 duplicate pages. Almost all of them are sorted and filtered category addresses such as /shoes/?sort=price-low and /shoes/?colour=red. The developer offers three ways to deal with them.
Which would you approve?
Four more well-meant fixes that do real damage. It is worth a quick look to make sure none of them has happened on your site:
- Page 2, 3 and 4 of a category point their canonical to page 1, so products listed only further down become harder for Google to discover.
- Whole sections are set to noindex out of fear, including pages that should rank, such as a second location page with genuinely different content.
- A canonical points to an address that redirects or returns an error, so the hint leads nowhere.
- The language versions of a site point their canonical to the English page. Translations are not duplicates; each language version should point to itself and use hreflang (a tag that tells Google which page is for which language) to link the versions. Our guide to multilingual websites explains how.
Your score, and your checklist for this week
Rate your site on the six things that matter most. “Not sure” is an honest answer: it simply means that check is your first job. Then use the checklist to work through the rest at your own pace; no developer needed for the first pass.
Score yourself
How tidy are your copies?
Typing the http and www versions of my domain always ends on one https address
My important pages have a canonical tag pointing to their own address
My menu, sitemap and canonical tags use the same version of each address
I have not used robots.txt as a quick fix to hide duplicate pages
URL Inspection shows Google chose the same canonical I declared for my key page
I do not have two pages answering the same question
0 of 6 answered. Answer them all to see your result.
Checklist
Duplicate content, point by point
Saved on this device only
Questions people ask
How much text can repeat before a page counts as a duplicate?
There is no published percentage, and you should be wary of anyone who quotes one. Google groups pages that are the same or very similar in their main content. A shared header, footer and opening hours on every page is normal and not a concern.
I have similar pages for each town I serve. Are they duplicates?
If only the town name changes, Google may well treat them as near-duplicates and show just one, and pages made in bulk to catch searches can fall under Google’s doorway page policy. Give each page something real about that place: jobs you did there, travel times, local photos. Our guide to location pages goes through it.
Will fixing my canonicals improve my rankings?
Nobody can promise that. What it does is remove confusion: Google shows the address you chose and stops splitting signals between copies. Whether the page then ranks depends on how good and how relevant it is.
Can I use a canonical tag to point to a page on another domain?
Yes, a canonical can point to another site you own, for example when the same article lives on two of your domains. For other people republishing your work, Google now prefers that they use noindex instead.
You reached the end
Your summary
This guide has 11 quick questions
Try them to see how much you already know. Your first answer is the one that counts.
Technical SEO: 0 of 7 read
- 1. Why isn't my page on Google? 12 causes and fixes: not read yet
- 2. Schema markup: what it does and which types a small business needs: not read yet
- 3. Multilingual websites: reach customers in every language without confusing Google: not read yet
- 4. Duplicate content and canonicals: what Google really does with copies: this guide, not read yet
- 5. Image SEO: faster photos that Google can find and understand: not read yet
- 6. Moving or redesigning a site without losing your Google traffic: not read yet
- 7. Core Web Vitals explained without the jargon: not read yet
0 of 57 guides read · Saved on this device only
New guides by email when they are published.
Unsubscribe in one click.
What to do next
- Type the http and www versions of your domain into a browser and check every one ends on the same https address.
- Open Search Console, go to Indexing, then Pages, and look for “Duplicate, Google chose different canonical than user”.
- Run URL Inspection on your most important page and compare the user-declared and Google-selected canonical.
- Find two pages that answer the same question, merge them, and redirect the weaker address to the stronger one.
Check your own site
Run our free site check: most of the technical points above, in under a minute.
Run the free checkNext in Technical SEO
Technical SEO
Image SEO: faster photos that Google can find and understand
The photo you are proudest of may be the reason your homepage feels slow. Here is how to check in two minutes.
26 min readStart reading
See it applied in an illustrative case study
Invented businesses and modelled figures, not client results: how the work in this guide fits into a whole engagement.
- Signal · 7 monthsIllustrative+189% organic consultation requestsKardia Varna
Rebuilding a slow clinic site until it had the heartbeat of a market leader
Private cardiology clinic · Varna · Bulgaria
- Signal · 5 monthsIllustrativeIndexed pages 38 → 410Kaze Aircon Services
A JavaScript booking app was hiding Singapore's busiest aircon servicer from Google
Air-con servicing, 60 technicians · Singapore · Asia-Pacific