1

I am working on a website where some pages link to a pdf. Here is the high-level description of the current set up:

Page A - This is a normal web page, lets say, a page that talks about instruction manuals for a particular piece of furniture, how rare they are, etc, and has a link to a PDF of the actual instructions for that piece (or multiple links if there are several versions). The URI of this page might be something like:

http://www.example.com/antique-xyz-game-table-manuals

...and it would have links to one or more "Page B"

Page B - This is a "pdf viewer" page... basically has the site's header and footer, and in between is an iframe of the PDF using the google document viewer (the actual pdf is hosted on the main site, it just uses the google viewer to embed it). This page also contains a direct link to the pdf which would either open it in the browser or download it depending on the user and whatever client they use. Example URIs of these pages would be:

http://www.example.com/file/view/f?=xyz-assembly.pdf
http://www.example.com/file/view/f?=xyz-maintenance.pdf
http://www.example.com/file/view/f?=xyz-parts.pdf

Page C - We'll call the direct URI of the PDF "Page C". Each "Page B" would have a link to the direct pdf. Example URIs would be:

http://www.example.com/file/xyz-assembly.pdf
http://www.example.com/file/xyz-maintenance.pdf
http://www.example.com/file/xyz-parts.pdf

A few more assumptions:

  • Page A is original content talking about the manuals, and linking to them. It is NOT a duplicate of the manual.
  • The PDFs themselves have their own entries in the sitemaps.xml and have no issues getting indexed by the search engines.
  • The UI is how it is, and it is how the customer wants it. This is not a question about the UI.

The question(s):

  1. On the direct link in Page B, to the actual PDF, Should I be using a rel="canonical" meta tag, and if so, should it point to the actual PDF or should it point to the page A?
  2. On Page A - should there be any rel attribute on the link itself?
  3. Any other SEO factors I should consider with this type of set-up?

Thanks in advance and let me know if any further clarification is needed.

  • Which document do you want to rank in the SERP: the original PDF, the viewer page or the overview page only? – Seb Jan 30 '17 at 8:25
  • The overview page (Page A) and the actual PDF – GWR Jan 30 '17 at 12:22
2

So you have three types of pages of which you want two to appear in the search engine results. Therefore you can use rel="canonical" link elements in order to tell the search engine which page of two pages it should index and serve as a result and which one it should skip.

The rel="canoncal" attribute has to be placed in a <link > element in a HTML document's <head> section:

<html> …
  <head> …
   <link rel="canonical" http://www.example.com/file/xyz-assembly.pdf" >
   …
   </head>
   …
</html>

For your setup this means:

http://www.example.com/antique-xyz-game-table-manuals

This page should be indexed and ranked. No need to specify duplicates or alternate versions, as long there are none of them. Basically it is a good practice to mark up these page with a rel="canonical" to themselves to avoid duplicate content issues with URL variations.

http://www.example.com/file/view/f?=xyz-assembly.pdf

This page is a duplicate of

http://www.example.com/file/xyz-assembly.pdf

As you only want the PDF file to rank in the SERPs you make use of rel="canonical" to the PDF document (despite the reader's menu the documents are identical).

To speak more generally:

  • Each overview page has a self referential canonical link element in the <head> section of it's source code.
  • Each viewer page has as canonical link element pointing to the real PDF document in the <head> section of it's source code.
  • The alternate Links are not needed in terms of SEO.

For your specific Questions

  1. On the direct link in Page B, to the actual PDF, Should I be using a rel="canonical" meta tag, and if so, should it point to the actual PDF or should it point to the page A?

Specify the rel="canonical" link in the <head> section of the viewer page pointing to to the actual PDF. If you cannot access the HTML source of the viewer page you may set up a "canonical" header for viewer pages pointing to the actual PDF file. (For more detailed information on how to implement "canonical" headers see: How To: Advanced rel="canonical" HTTP Headers (Moz.com)

  1. On Page A - should there be any rel attribute on the link itself?

Not for SEO reasons.

  1. Any other SEO factors I should consider with this type of set-up?

Maybe, but it depends on further information to judge this :)

As you have direct links to the PDF files you may think of using optimized anchor texts. F.e.: "Assembly Guide for Table Type XYZ (PDF)" instead of "xyz-assembly.pdf". Make sure the PDF link is the first to be crawled by the search engine.

Make sure you do not mix "noindex" and "canonical"! maybe you think of marking the viewer pages as "noindex" in order to keep them out of the search engine's index. This would hurt the canonical set-up.

In order to save crawling resources you may set up your overview page in a way that it only serves links to the actual PDF files ind leverages Java Script or something alike to enable the viewer mode. This would avoid search engines crawling the viewer pages. But you would stay with the canonicalization of viewer page and PDF file, as users may link to the viewer URL from outside of your page.

  • "Make sure the PDF link is the first to be crawled by the search engine." Could you clarify why you would recommend this? – Luke Rehmann Apr 5 '18 at 19:24
  • This should be clarified: if you have multiple links pointing to the PDF document you have to make sure that the first one of those links is the one with the optimised anchor text. The first of multiple links is the one that is taken into consideration for connecting keywords to the link target. – Seb Apr 5 '18 at 19:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.