Added by Erik West on July 13, 2011
A research report issued by search giant Google provides a rare look into their efforts to improve search results, including understanding users’ intent.
The report details how Google evaluates a user’s search terms and ranks the search results using both on-page and off-page signals or features. It’s clear that page titles, headings, and URL depth play a major role, as shown in any search result. The report highlights the importance of the proximity of search terms, with search terms that are closer together given more weight than search terms that are further apart from each other. The report also discusses that the position of the user’s search terms on the page play a role in determining relevance, with a preference for content that has search terms closer to the beginning of content rather than towards the middle or end.
An interesting finding is how Goole analyzes patterns across the web for properties that include a lot of outgoing links, poor sentence structure, and the presence of offensive terms. Google takes advantage of the diversity of content on the internet to enable it to statistically determine the average length of good sentences, ratio of visible keywords compared to keywords in meta or alt text, plus the number of outgoing links compared to the number of words on the page. The last factor, the number of outgoing links is important because, the report states, webmasters have control over which sites they link to, and links to penalized sites can have a negative effect on the linking site, in terms of statistical analysis.
The report highlights how the search engine evaluates off-page signals like linking sites and anchor text to counter the effects of on-page keyword stuffing. Linking sites’ reputations are based on their own PageRank. PageRank is Google’s assigned score that simulates the actions of a typical web server and their probability of choosing a particular page over another page within a search result. Search engine marketing firms closely monitor their clients’ sites PageRank.
The report, available from Google Research, is one of the few that provides an overview of the search industry and its general direction.