improvements to its search-ranking algorithms, codenamed "Panda," have just rolled out in all languages except Chinese, Japanese and Korean. The changes are intended to reduce the effect of "content farms," sites that churn out lots of low-quality content to skew search results in their favor. Reports show that Panda appears to be working; the biggest content farms have shown marked decreases in traffic since Panda first launched in the U.S.Google's recent
The changes began in February for the U.S. market and expanded internationally, though still only for English results, in April. Google reports that the algorithm changes affected almost 12% of English queries and the inclusion of new languages will affect 6-9% of queries worldwide.
Google's challenge is to use its endless stream of data to determine abstract values like quality and relevance. The goal is "to give people the most relevant answers to their queries as quickly as possible." Relevance can be a rather subjective quality. The process of developing the Panda algorithm improvements incorporated some manual feedback from Chrome users who utilized a Personal Blocklist extension to their browsers, but the solutions are mostly based on automated scores.
This hasn't improved results across the board. Some sites in the travel industry, for instance, have been experiencing much worse search results than they did before the change. But Google has outlined an extensive list of criteria defining its vision of quality web content, and it concentrates on providing value for the user. Sites or industries built on gaming search optimization might have to revise their approach.
To learn more about how Google defines quality and builds it into its search algorithms, check out this fascinating in-depth panel discussion with Amit Singhal (author of today's announcement), Matt Cutts and Ben Gomes, some of Google's top search people: