Refining Search Engine Queries: Clustering and User Intent


Understanding the way that search engines determine which pages to display in response to a user query can help to enlighten business owners and end users regarding the rankings and page views they receive from these search engines. A recent study conducted by Stanford University in conjunction with Google Inc [1]. explored the various ways that queries can be refined and related more closely to the intention of the end user. A variety of methods were examined and evaluated in order to determine the most effective ways to handle related queries and to provide the user with the information they need more effectively and rapidly.


Presenting a number of diverse results can provide the end user with a wide range of choices. This method assumes that the user will scan through these widely variant results to find the one that best meets his or her needs. However, in many cases the number of variants may be overwhelming for the end user and may not produce the correct results even when the most relevant topics are included in the results.


A more advanced method of ordering search results groups related pages together. For example, pages about apples would be clustered together in one group, while Apple computer pages would be included in a separate cluster on the results page. This provides the benefit of choice to the end user while still allowing a larger sampling of relevant pages to be displayed for each search engine query. Most advanced methods of ordering search engine results use some form of clustering to display those results to the user.

Multi-session methods

Sophisticated search engine query algorithms can take into consideration the previous search activities of the IP address or user in order to provide the most relevant results. By considering recent searches when returning query results, search engines can provide faster, more accurate service for their end users; this is especially true when those results are clustered into relevant groupings.

Putting it all together

Clustered search engine results are preferred by users, as they provide an optimal degree of diversity while maintaining a manageable list of potential results. By refining the search engine algorithm’s clustering method with information about recent searches and deriving information about the user’s intentions from those searches, query results can be tailored to the specific needs of the end user. This allows the most relevant and immediately necessary results to be ranked more highly and displayed more prominently than those that are less likely to render the needed information to the end user.

By clustering query results by general topic, search engine providers can better serve their end users and provide more accurate page rankings and query responses for all parties involved.


[1] Clustering Query Refinements by User Intent by Sadikov, E., Madhavan, J., Wang,L., Halevy,A.

Dan Petrovic, the managing director of DEJAN, is Australia’s best-known name in the field of search engine optimisation. Dan is a web author, innovator and a highly regarded search industry event speaker.

More Posts - Website