Yahoo algorithm predicts what people want to read, increases home page clicks

Fast Company

Yahoo attributes a huge increase in clicks on stories in its "Today" module on the home page to a sophisticated algorithm that predicts what users want to read. The algorithm, called CORE (Content Optimization and Relevance Engine), generates 45,000 versions of the module every five minutes and chooses which ones to display to users based on their gender and age (obtained if they have registered for a Yahoo account) and previous clicks on Yahoo. CORE figures out which story packages are attractive to which users by testing them before a select group of users in each demographic. The algorithm has made Yahoo editors savvier, they say, about what content will appeal to readers; editors can override the automated recommendations to promote an important story, even if it isn't predicted to generated a lot of interest. CORE will be rolled out to Yahoo News next.

  • Steve Myers

    Steve Myers was the managing editor of Poynter.org until August 2012, when he became the deputy managing editor and senior staff writer for The Lens, a nonprofit investigative news site in New Orleans.

Comments

Related News

Email IconGroup 3Facebook IconLinkedIn IconsearchGroupTwitter IconGroup 2YouTube Icon