How to correctly mark up different versions of the same document which are non-canonical

Question

The use case is a document which has multiple versions which are all simultaneously available. For example, documentation on a product for each version of that software:

• /v1/install-guide
• /v2/install-guide
• /v3/install-guide

These are not the same content in the sense that it would be incorrect to specify in the v1 and v2 pages that the canonical URL is v3. If I searched in Google for 'product install guide v2' I'd expect to be able to find v2. Each version of the document would be slightly different but they'd have a lot in common with each other.

So the question is: can I mark up these pages in some way to signal to Google and friends that the latest version (v3) is preferred, so a simple search for 'product install guide' is more likely to show the v3 page instead of the v1 or v2 page?

Answer 1

given your question I have to make the following assumptions:

1. your website is offering a section for install-guides for a product
2. guides differ from product version but stay relevant for customers using older product versions
3. your customers are aware of that and know exactly which version they may need

First: this indeed is not the use-case for a rel="canonical".

One approach of dealing with situations like yours is structure: You can implement a static guide page for your product that always contains the most recent guide. This page is linked from the product or category page, so it gains a lot of importance.

Scheme:

exapmle.com/product/guide/
exapmle.com/product/guide/guide-v1
exapmle.com/product/guide/guide-v2
exapmle.com/product/guide/guide-v3

Each time a new guide is released it gets published to the static page and the old one moves down to a new URL deeper in the URL structure. Those "old guide" URLs are linked to from the static page, but not from the category pages. Therefor they will be seen "less important"

Make sure, the v1, 2, 3, … n versions are only linked from the main guides page and each to each. The main page must be well linked across your whole site.

Compared to your existing URL structure this give more relevance to the guide published on exapmle.com/product/guide/.

• /v1/install-guide
• /v2/install-guide
• /v3/install-guide

In this structure each URL is "worth" the same and Google needs to rely on other signals to decide which URL is the most relevant.

The next step is optimizing the guide pages in a way search engines quickly understand the difference between the individual versions:

• individualize the page's title, e.g.: <title>Version 1.0 install Guide for Product XY</title>
• individualize the page's main heading, e.g.: <h1>How to Install Version 1.0 of Product XY</h1>
• individualize the page's description, e.g.: <meta name="description" content"Here you can find your guide to install version 1.0 of Product XY […]" />
• make use of the anchor texts to point out different versions when linking to the guides <a href="http://example.com/product/guide/guide-v2" >Install guide for V 2.0</a>

Each to point out the version it was designed for and maybe the year/month/date.

Hint: it may be helpful to point out the differences at the top of the individual pages. So customers can easily see why they need different guides ans search engines get additional information an unique content for each page lowering your risk of "near duplicate content" issues.

This way you make sure that the most recent guide is always present on example.com/product/guide/ and will served as result for generic searches like "install guide product". Search engines will recognize that content on this page gets updated frequently and also will recognize that sub-items of this pages exist and serve them for more specific search terms like "install guide Product V 2.0". By optimizing the pages for the specific versions they represent you make sure clients find their version by searching for it.

Answer 2

I've seen cases like yours when browsing technical documentation like Django docs.

There aren't any specific tag to add in this case but you can use the following strategies to hint Google that you prefer the v3 link (the latest version):

Links without versions

• When you refer to this content outside the v3 scope, use links without versions that automatically redirects to the latest version page, e.g.:

  example.com/install-guide -- redirects to the latest --> example.com/v3/install-guide

When you have a newer version (v4) it will redirect to example.com/v4/install-guide.

Alerts with links to latest version

• always show the available version of your documentation so the user can easily decide which one to use and if they visit an older version you can show an alert that they are not on the latest version and add a link to the link without the version in the url, Google will know that those older version pages are referring to the latest one, implicitely giving them more weight to decide which one to select. In this case the latest version always will be referenced by the other ones.

When Django deprecates a version they clearly shows it to the user, in your case you can also add the latest version link in the alert.

Fresh links

• fresh links (newer content, or newer versions in your case) will naturally get more traction than older ones.

Answer 3

I have not seen any proven technical way that would tell a search engine that a certain version of a guide or page is newer other than the date as norcal johnny mentioned.

However, I have noticed that in many cases the older version will include a warning at the top saying, go look over there for the newer version. A bit of a nag if you are using the version you are looking at now, but excellent if you are trying to evaluate a product and want to make sure you look at the latest.

The first website where I saw that practice clearly occurring was the W3C. For example, this document has a "latest version" at the top: https://www.w3.org/TR/html401/

Now, I have noticed it at MSDN, where they have a link named "Recommended version". https://msdn.microsoft.com/en-us/library/78zh94ax%28v=vs.71%29.aspx

Either way, I have not seen any kind of technical HTTP or HTML that could be used to indicate such. There is a link tag with the rel="alternate" that may work, but from what I understand that only really works for languages or other varying versions of the exact same document, not various different versions.

There is one other technology that Google tried with and I am pretty sure they are still using it, it is http://schema.org which happens to have a https://schema.org/softwareVersion tag. So assuming Google did implement that specific tag, then it should understand your version scheme. Following Seb's insight, having a /guide page, you could then have links to each version. The link would be tagged with schema to include that softwareVersion information. However, that version only accepts Text making me think that's a description ("Best Version Ever!") of the software and not an actual version (1.0, 2.3, 3.7.1...) But that I think would probably be the best you can do.

Answer 4

Whenever I search any FB API information, it return me latest information with latest API URL, and they are not using any redirection neither canonical link tag that point to latest information, what they simply do is, they start pointing to latest information everywhere, So Google will crawl that more often and will pass more juicyrank, that is the only signal you can also use for your website. You have control over your links so use that one.

Don't use noindex, 301 redirect and canonical link tag on old pages that point to lastest URL, that will make your webpage dead in eye of Google search engine, I mean Google will remove whatever initial benefits that are given to that webpage. I have asked recently on Google official hangout and they said they do not give any initial PR value for noindex tag, so I think, it will same apply to canonical link tag. and When you do 301 redirection then all PR will pass to new URL, and whatever initial PR is given to that old webpage will be remove because it is no longer available. Google give some value on those webpages which are live.

You don't really need to change your markup. eComerce sites display same kind of webpage on two different URL to target two country, and what they change is only currencry price. For example, for USA it's dollar and for India it's rupees, and rest of all things are same, and still that kind of website did not face any Panda penalty.

Answer 5

You can use Page Dates for sorting.

Google estimates a page date based on features of the page such as dates in the title and URL.

Custom Search allows you to use this date to sort, bias and range restrict results by using a special metadata key of date. This estimated date can be used in all operators that use the &sort= URL parameter, including "Sort by Attribute", "Bias by Attribute", "Restrict to Range".

For example:

For a listing of the current codes visit the Google Custom Search documentation.