Amazon.com on this week took the wraps off its search engine, A9.com, which combines Web search with personalization features based on user profiles.
A9, an Amazon search subsidiary established 11 months ago in Palo Alto, CA, relies on Google for its Web search results. Also, from the same A9 results page, searchers can access information from Google Images; Amazon's Search Inside the Book feature; the Internet Movie Database; and reference resources from GuruNet. A diary feature lets users make notes about sites they visit. A9 displays Google paid search listings.
A9 also added personalization features. Along with cosmetic changes, like the ability for users to alter the site's font, color scheme and language, A9 takes steps in tailoring search results based on user search and consumer behavior. The search engine requires users to sign in with their Amazon user name and password. It stores users' search history, lets them bookmark searches and tracks and records which results were clicked. A9 encourages users to download its toolbar, which further builds profiles by recording all Web behavior.
A9 also suggests Web sites to users based on their surfing history. The recommendation feature, called Discover, is labeled a test by A9. The technology is similar to Amazon's collaborative filtering that allows it to recommend books and videos to consumers based on their previous shopping habits. A9 said it would make changes to its feature based on user feedback.
Search engines have tabbed personalized search as a goal, yet few features have been developed as they tread lightly around concerns that they are collecting too much personal information. Google in late March rolled out a test of a personalized search engine that tailors search results based on a profile users build. Google Personalized, however, does not correlate search data with personal information.
Yahoo has moved carefully in personalization, despite having personal information for 146 million registered users, in deference to potential privacy concerns.
A9 goes much further, potentially tying together search data, Web browsing history as well as shopping and personally identifiable information. It collects data on all sites visited and all searches done, which A9 can use to further personalize results. The information is stored on A9's servers, not on a user's computer.