Rel="canonical" meta tag for category pagination pages - how to use it correctly?

The question about the use of rel=”canonical” on the pagination pages often confuses even the experienced search engine optimizers. And the beginner would get lost immediately, without any chance of recovery. Indeed, the Internet is full of conflicting articles and confronted advice. If summing it up, they are as follows:

• use rel=”canonical” meta tag that points to the first page of the category.
  https://example.com/category has rel=”canonical” pointing to https://example.com/category
  https://example.com/category?p=2 has also rel=”canonical” pointing to https://example.com/category
  and so on
  
• use rel=”canonical” meta tag that points to each pagination page.
  https://example.com/category has rel=”canonical” pointing to https://example.com/category
  https://example.com/category?p=2 has rel=”canonical” pointing to https://example.com/category?p=2
  and so on

*In this article, we only consider techniques related to the use of rel=”canonical”. The options with robots meta tag are the topic for another conversation.

Do not forget about Google recommendations, namely rel=”next” and rel=”prev” directives. In this support article Google insists on using these links to help its crawlers to understand the structure better. Also, it is stated there that Google prefers the second variant – using canonical meta tag for each category page.

What then can you argue at all?

The problem is that Google advices are often found being contradictory.  For example, if using the very literal approach for this issue type, we would face hitting another Google claim – not to have duplicated titles and content. Despite Google says that it can recognize pagination pages and to interpret and index them correctly, SEO-teams with experience are used to trust such statements very carefully. And they have all good the reasons for this. Even the correct implementation of rel next & prev does not guarantee you the absence of such a picture:

This fact looks like is speaking in favor of the first approach type. However, no one is assured to be avoiding these warnings, whatever variant of rel=”canonical” meta tag use on the category page you choose. It only proves again, that you still can’t rely on Google on every clause. Why do these “HTML improvements” appear in the Search Console if Google is supposed to understand that pagination pages are… obviously the pagination pages?

What happens when we use own rel=”canonical” on each page of the category (combining it with rel next & prev)?

We use rel=”next” and rel=”prev” to indicate the position of the current page in relation to the neighboring within the navigation unit. We use rel=”canonical” meta tag pointing to itself to tell the search bot that all these pages are preferred for indexing. It is important if we have different structures in the URL parameters (if we use some filters, for example). That is extremely important in different CMS systems, like Magento f.e., that can create thousands of pages for such technical needs.

This approach allows to be sure that every product on even the deepest page will be seen by the search bot and will get indexed. However, we leave it up to Google to decide which of all created pagination pages will be shown in search results. Well, actually it’s always up to Google, despite all rel=”canonical” meta tags in the world, but let’s talk more on that later.

And what if we use rel=”canonical” pointing to the first page of the category (with rel=”next” and rel=”prev”, of course)?

Rel next & prev play the same role. Still, we give a clear signal to Google that the first page is the preferred one. There is one more essential aspect – the page can receive all link juice and the Page rank from the pages that have it as a canonical one. Why just can? Because it is still up to Google to decide.

Both rel next-prev and canonical tags are just hints that are not mandatory for Google

In contrast to meta robots tag, rel=”next/prev” links and rel=”canonical” meta tag are recommendatory in their nature. That means that Google can just ignore them with a poker face if it decides that some other page is… more canonical, for instance. Examples can be found in the new Search Console:

This does not entail any sanctions though, you are just going to find that Google shows in the search results not the page that you told it to make canonical, but the one that it thinks is more appropriate.

What method do big brands use?

There is one golden rule – if you doubt how to act, copy a leader’s behavior. Well, or just pick the arithmetic average.

Armed with this approach, I decided to conduct a little research. Sportswear online stores were chosen as a general direction. Region – the USA. This combination promised a prohibitive level of competition. And as a consequence – the highest level of quality.

First, I used Serpstat to find the most appropriate resources. Then, with the help of Similarweb tool, I gathered some general information about them. And with the majestic method of manual digging in the code I checked each resource for the presence of the appropriate tags. A little transcript before we move on to the most interesting part:

Website – the address of the website being audited.
Keywords (Google.com) – number of keywords in Google.com (USA) database associated with the resource.
Monthly visits (Apr 18) – number of overall (not unique!) visits during the April, 2018, according to the Similarweb database.
Organic search % – the percent of organic traffic (to make sure the website is well perceived by Google).
Rel=next&prev used – that is clear.
Category page canonical – is the rel=”canonical” meta tag used on the category and pagination pages. If yes – in what exact way:
1st page – canonical tag is used and points to the first page from all pagination pages;
Each page – each pagination page points to itself;
Inf. scroll – infinite scroll is used;
Not used – rel=”canonical” is not used at all.

So, here is what I got.