I think I shouldn't use canonical tags for this situation.

[skuba]

Hi,
I am putting together SEO requirements for a project and it includes adding canonical tags.
I definitely want to add canonical tags to all pages so that pages that are sorted differently or pages with URLs with tracking parameters or sessions IDs are still referring back to the original URL.

But there is this situation that I am not sure.
Let's say a user browse the site in different flows. For example, one flow is GENDER, then CATEGORY, then BRAND. And another could be CATEGORY > GENDER > BRAND

Both pages will have very similar content. But should they be linked canonically? I don't think so.

Let's say the title of one page is GENDER CATEGORY BRAND at store.com , and the other page is CATEGORY GENDER BRAND at store.com

There are 2 different pages. More ways to get found by seachers. Why would I want them to be treated as one?

I appreciate your feedback.

Thanks

[g1smd]

I posted some stuff vaguely related to this topic just a few weeks ago in...
[webmasterworld.com...] if that's of any use. :)

[aakk9999]

Hi g1smd, I read your post, but I have a question which I think is similar to what skuba is asking.

Lets say that there are a number of products that can be grouped by categories. Lets say that there is sufficient number of products to span a number of pages when listed at (for example) 10 per page.

So (to take skuba's example), the list shown for GENDER, CATEGORY, BRAND order shows different products to list shown in CATEGORY, GENDER, BRAND.

I would like to rank the first page of the first list for GENDER and the first page of the second list for CATEGORY.

Would this cause a problem ?

[skuba]

G1smd, thanks for the reply. i did read your message, but not sure it's what I am asking. I am also talking about how the page Title will change depending on the user flow. Wouldn't it be better to have those 2 or 3 different pages indexed?

Thanks

[TheMadScientist]

Wouldn't it be better to have those 2 or 3 different pages indexed?

Sorry for hopping in, but here is the general rule of thumb I think g1smd was talking about:

Are they actually different pages, or pages with essentially the same content and a different title? If the content on the pages is significantly different from one page to another, then they are different pages, but if the content is 'essentially the same' somewhere else on any page of the site, then the pages are 'duplicates' regardless of the title.

You're really the only ones who can answer the question, because it's a site and situation dependent answer... If the content of the pages is unique based on category order then they are different pages, but if the content is essentially the same anywhere else on the site using different navigation changing the order of the words in the title or the order the products are displayed in and allowing the page to be indexed is not something I know of usually being recommended, because the content is still the duplication of another URL.

If it's unique it's unique.
If it's not unique it's duplication.

You're the only ones who can answer the question...

[skuba]

That makes sense.
But if we just care if the content is the same or different enough, I don't understand why it's recommended to use canonical with SORTED pages. But not recommended to use with PAGENATION.
Thoughts?

[TheMadScientist]

Usually, in my experience pagination shows 'items' '1 through 20' on page 1 and '21 through 40' on page two, so the content of the pages is different as you go through the pagination.

There's more than one page of products...

With sorts, you are 'generally' changing the order, and while on 100 products at 10 per page it could be significantly different at 20 per page (or something) it's entirely possible 15 of the products are the same, but the order they are displayed in has changed.

So, generally, with pagination there's usually a definite 'uniqueness' between pages 1 and 2, while with a 'sort' it could be pages 1 and 2 are identical to pages 1 and 2 using the 'regular view' but the order they are displayed in is different.

The answer again is situational...

I have a couple of sites where the 'sort' is different in 'sections', but I allow the 'sorted' pages to be spidered and indexed in a couple of the sections, because it's significantly different than the 'show everything in a to z order' section and the 'show me these specific types only' section is 'unique and useful' in and of itself, because about 2/3 of the 'items' are eliminated and people often look for the 'specific item' in the section, so it's useful and cuts down on 'noise' helping people find the 'item' they were looking for.

There are other sections where the sort does not significantly change the information displayed, except for the 'order' or only eliminates a few 'items' from the section, so it's essentially the same as the 'show me everything a to z section' except for the order the 'items' are displayed in. I do not allow those sections to be indexed, because they are for my visitors who would rather view the information in a different order...

So, my personal view is: If it's significantly different and can 'stand alone' as a unique section rather than 10 pages of 20 products displayed in different orders I allow the sections to be indexed and searchable. If the 'sections' all display the same 200 'items' so the sum of the ten pages is not significantly different from one section to another, then I only allow one to be indexed.

IOW: If 200 products produce 10 pages per section and all 200 products are displayed within each of 3 sections. I pick a section to allow to be indexed. (I would have 3 10 page sections with only one indexed.)

If 200 products produce 10 pages in one section, and only 3 pages in another section and 4 pages (completely unique from the 3 page section) in a 3rd section I would probably allow all 3 sections to be indexed. (I would have one 10 page section, one 4 page section and one 3 pages section all indexed.)

If the 3 page section and the 4 page section of the preceding example were 'essentially the same' meaning they basically contained the same information in a different order and one had a bit 'extra' mixed in I would probably only allow one of the two to be indexed. (I would have one 10 page section and one either 3 or 4 page section indexed.)

Hope this makes a bit of sense and gives you some ideas.

[freejung]

The trouble is, it's not up to you whether the different versions of the page are treated as duplicates or not. The Canonical tag is a suggestion, not a hard directive. If the two URLs contain essentially the same content, they will probably be treated as duplicates regardless of how you use the Canonical tag. In that case, it might be better to use the same canonical URL for both, so as to have at least some control over which URL gets canonicalized. Otherwise, Google will just decide for you.