Going on a date with Google – Dublin Core and More

It’s all about the first date… and this post covers some of the devious lengths Google will go to to get it.

Why does Google want to get a date out of you?

Time is generally accepted as an essential ingredient in the calculations of relevancy in SERPs. Some events will be promoted in SERPs because they are news, others, such as recurring events, may well appear in results at appropriate times. Updates and edits to site content can help revive an older story. All these examples are dependent upon the date attributed to the article, the origin and the changes made. The minty freshness update back in November 2011 caused noticeable impact to around 10% searches.

To try to benefit from these changes, many site owners looked for ways to touch up older pages and content to make it at least look more current. It is the equivalent to software folks in Unix world reaching for the ‘touch’ command to change the ‘last modified dates’ on all their files to ‘get things noticed’.

 

But what is Google looking for?

With the drive to discover fresh content and attempts to improve the understanding of timelines, Google will try to ‘datestamp’ everything it finds… and it can make mistakes when generating a datestamp for your content that impacts your relevancy to search, especially if ‘date order’ is selected as an advanced search option.

 

Unique IDs as Dates

Many e-Com sites use unique identifiers in page titles in order to reduce duplications and create unique, informative pages. Readers are likely to be familiar with GWMT diagnostics section on HTML suggestions:

Non Informative Title Tags and duplications

 

In order to reduce the number of duplicated page titles, a typical workaround would be to use the following format for product details pages:

{brand} {product} {key feature} – {unique identifier}

The purpose of the unique ID is to ensure that varieties of the same product (e.g. sizes, colours) each have a unique page title. [You may have noticed GWMT also reports 'noninformative' title tags... strongly suggesting that the title remains a key source of information for Google's indexing algorithms]

For an average SEO making page title unique is just the start. Introducing an ID can guarantee uniqueness and introduce useful information (win-win!), but it can have unintended consequences (oops!).

The purpose of this post is to cover just one of these unintended consequences – misinterpreting a product ID for a datestamp.

The search result below shows an example of Google misinterpreting a product ID and using it as a datestamp in results:

Product ID used as date stampThe unique identifier 11171977 has been interpreted as a date. Now this listing carries the datestamp: 17 Nov 1977 which has been taken from the product ID (11-17-1977)

That’s inconvenient if you’re trying to sell this house, it looks like its been on the market since 1977 !

Here’s another example (pink camera)- this time it’s the model number that is being treated as a date:

Pink Camera product ID as date stampThe DSC-T70 8.1 in the title is being treated as if it means 1st August 1970 !

Dates in Content

When an article is posted on a blog, there is often a time-stamp associated with the post and prominently displayed at the top of the page. Google finds this on the page and the association is made, in this case, quite sensibly assuming your server time is correct and you aren’t manually fiddling the dates.

Less obviously, if you have a table or other date source within your content, there is a chance Google will take that date to be an appropriate datestamp. For instance, if you reference the last time a certain event happened, but fail to state the current date, then the entire article could be catalogued as the time of the referenced event rather than the current one.

Be careful to check your content for date references and help Google find an appropriate date for your content.

If you see a strange datestamp in SERPs, then look at these on-page elements for clues as to the cause of the mix-up:

a) Footer for software version numbers

b) Tables of reference data

c) References to dates within text (ensure the first date referenced is the date you’d like to have associated with the post)

d) Telephone numbers

 

Avoiding a bad first date:

Here are a few ways you can help make your first date with search engines more memorable:

a) Place the preferred date between the article’s title and the article’s text in a separate line of HTML, and remove all other date references from the body of the article to avoid confusions.

b) Clearly mark up the datestamp you want associated with the article using the Dublin Core meta tag:

<code><meta name="DC.date.issued" content="YYYY-MM-DD"></code>

or more completely:

<code><link rel="schema.DCTERMS" href="http://purl.org/dc/terms/" />
<meta name="DCTERMS.created" content="YYYY-MM-DD" />
<meta name="DCTERMS.modified" content="YYYY-MM-DD" />
</code>

c) Use the <publication_date> tag inside news sitemap xml files

d) If using OpenGraph, then add the article:published_time, article:modified_time tags

e) Using microformats and schema mark-up to ensure telephone numbers are parsed correctly (see Schema.org for these definitions), ie. not treated as dates.

f) Modify the way product IDs are displayed. Introduce characters or spaces to break up the ID so it becomes less ‘date-like’, yet retains uniqueness.

If erroneous Google datestamps have caused you problems with search relevancy, please let me know !

 

Enhanced by Zemanta

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>