Tuesday, May 12, 2009

Enabling Rich Snippets in Custom Search

Posted by: Rajat Mukherjee, Group Product Manager and Nicholas Weininger, Software Engineer

Today, we announced Rich Snippets for Google web search, a new presentation of snippets that highlight structured data embedded in webpages. Rich Snippets provide summary information, including important page-specific attributes, to help you quickly identify the relevance of your results. Experiments on Google have shown that people find this additional data valuable -- if they see useful and relevant information from a webpage, they are more likely to click through to it. Our web search team is currently experimenting with a limited set of attributes for reviews and user profiles that webmasters can provide through in-line markup in their webpages by using open standards such as microformats or RDFa.

Since Custom Search sits on top of the Google web search platform, we're enabling Rich Snippets for custom search engines too. In fact, Custom Search provides support for richer snippets; you can define your own custom attributes that we'll index and return with your custom search results. In addition to microformats and RDFa, you can also provide custom metadata within your webpages via special markup called PageMaps. A PageMap identifies specific attributes that are recognized and preserved by Google at index time, and returned along with search results for presentation.

So, if you're using Custom Search or Site Search on your website, you can now control further how your content appears in search results. You can showcase key information -- such as image thumbnails, summaries, ratings, and prices -- in your result snippets if you provide the appropriate markup on your pages. As in the example shown below, displaying specific attributes can make the snippet more useful for shoppers.


At this time, Rich Snippets attribute information for Custom Search is only returned in XML (via <PageMap> tags), so you can use your own customized presentation controls. We're looking to make this easier for you in the future. Indexing of the rich snippets information can have unspecified latency, as some pages are indexed and refreshed more frequently than others, and PageMap attributes may not be indexed from all webpages.

As an illustration, the webpage featured in the following example provides custom information about an image thumbnail that is displayed in the rich snippet of the result along with date, author and category information.


A sample PageMap for the above example defines custom attributes that are encoded in DataObject tags via the following markup:
<!--
<PageMap>
<DataObject type="image">
<Attribute name="image_src" value="http://www.nytimes.com/images/2009/03/14/business/14vend.751.jpg"/>
</DataObject>
<DataObject type="publication">
<Attribute name="author" value="John Tagliabue"/>
<Attribute name="date" value="March 14, 2009"/>
<Attribute name="category" value="Business/World Business"/>
</DataObject>
</PageMap>
-->

If you are getting results back via XML, then the custom attributes are returned in the results within the PageMap tag, as shown below. You can parse the DataObjects within the PageMap tag and provide customized presentation of the relevant attributes.

<r n="1">
<u> http://www.xyz.com/business/vending_machine.html </u>
...
<t> In Italy, a Vending Machine Even Makes the <b>Pizza</b> </t>
...
<s>The European vending machine industry has annual sales of about #33 billion, much of it coming from factories and offices.</s>
...
<PageMap>
<DataObject type="image">
<Attribute name="image_src" value="http://www.nytimes.com/images/2009/03/14/business/14vend.751.jpg"/>
</DataObject>
<DataObject type="publication">
<Attribute name="author" value="John Tagliabue"/>
<Attribute name="date" value="March 14, 2009"/>
<Attribute name="category" value="Business/World Business"/>
</DataObject>
</PageMap>
...
</r>

Let us know if you have feedback about custom Rich Snippets. For more information, join us for a discussion on Advanced Custom Search Configuration at Google I/O on May 27.