Washington Post's Trove to personalize news

Everyone is getting in on personalized news aggregators--who will dominate?

Shortly after Yahoo announced its new personalized news service, Livestand, which will debut on mobile devices in the near future, the Washington Post announced Friday morning that it's preparing to launch a new free personalized news aggregation website called Trove.

Those who follow the Washington Post might be a little puzzled by this, since the news giant already offers a personalized news site called iCurrent, which it bought back in July for an undisclosed sum.  According to the company, Trove will essentially be a reinvention of iCurrent that is scheduled to debut next month.

The service will comb through 10,000 news sources to provide custom-tailored content using an algorithm that delivers news feeds based on previous articles the user has read.  The company is shoveling somewhere in the range of $5-10 million into the project, which will include the development of mobile apps for the iPhone, iPad, Android, and Blackberry.

The personalization of content is a growing trend that's proving to offer big returns.  Pandora is probably the most relevant example, having attracted 75 million registered users and drawing three million more each month with its proprietary genetic mapping algorithm.  Presumably, Trove's algorithm will use a similar algorithm based on key words, though the Post Company did not respond to inquiries from VatorNews about the algorithm.

Or it might reflect the algorithms used by content personalization engine Outbrain, which bought rival Surphace from AOL last week.  Outbrain uses a set of different algorithms to determine what a reader might want to see in terms of reading content, including contextual analysis (articles with the same content), collaborative data (“People who read this also read…”), popularity (most visited articles on the Web), and personalization (where the service cookies the reader to make recommendations based on previous behavior).  Recently, the company added two new algorithms: most clicked (prioritizing articles that are getting the highest click-through-rate) and social (finding the articles that are being shared the most on social networks).

Amazon and Netflix have also developed personalization algorithms based on purchasing history, but content is obviously a different animal altogether, particularly where news is concerned. News is just that: new. Headlines are constantly shifting and new information is continually emerging, but more importantly--reading an article is not the same thing as liking an article (whereas purchasing a book or DVD is a clear sign of interest)--so tailoring that towards individual reading preferences will be tricky, but that's what several major news organizations are trying to do. The New York Times' personalized news site, News.me, recently debuted to offer custom-tailored news content.

Of course, as Pandora revealed, the key to getting a big return on personalized content is going mobile, so Trove's mobile apps will likely be the ticket to the project's success. 

Image source: editorsweblog.org