Of the panels I sat through at the recent Search Engine Strategies (SES) conference in San Jose, the most relevant to the space in which I work was the panel on News Search.  Although the panel is a consistent presence at the SES events, this particular panel had a Google representative, Maile Ohye, talk about how Google News indexes, groups and ranks news content.

When it comes to ranking, some of the more important variables include:

  • Recency – Fresher stories are going to be ranked higher.
  • Originality – No brainer.
  • Locality – I was unclear how this would impact relevancy unless GN was considering the IP of the searcher when displaying results.
  • Quality of source – I’ve written about how Google News considers quality of source in the past.  Their 2003 patent application (which was recently approved) explains in depth the variables that make up “quality of source” in Google’s eyes.
  • Aggregate editorial interest – This is of particular interest.  I equate this somewhat to content clustering.  Basically, when Google News is analyzing a news story, it is going to run some content analysis script on the article.  GN is then going to compare the content in the article to the aggregate index.  If the article’s content shares an affinity with prominent content appearing in the aggregate then it will be ranked higher.

I don’t recall Ohye specifically mentioning this but I find it hard to believe that they don’t look carefully at click-through rates of stories as well. Ohye finished up by running through some basic best practices for news search, which the TopRank blog does a good job explaining.