Facebook gages attention to bypass lackluster “Likes”
Facebook doesn’t need you to “Like” something any more to know you’re interested: just by taking the time to read, the ever-seeing algorithm is learning. The social networking behemoth has announced the latest iteration of its News Feed system, no longer tracking just “Likes” and comments, but comparing the relative attention each post garners to figure out what users might really be intrigued by.
Traditionally, Facebook used explicit measures to work out whether a person was interested or not in something in their News Feed. If they clicked “Like”, or left a comment, the algorithm would assume they’d like to see more stories similar to those in future.
Those that went without such clicks, however, were presumed to be less relevant, and so similar stories would show up less frequently over time.
Turns out, Facebook’s research suggests, “just because someone didn’t like, comment or share a story in their News Feed doesn’t mean it wasn’t meaningful to them.”
Instead Facebook is looking at how long each story grips someone’s interest, basing algorithmic learning on that amount of silent attention.
Crucially, it’s the relative reading time that is being looked at. For instance, if you look at a video post for fifteen seconds, that could be because you’re truly interested in it – even if you still don’t go on to “Like” or comment on it – or it might just be because your connection is slow and buffering takes a long time.
So, Facebook’s engineers are using comparative measures: how long you looked at each story in relation to how long you looked at the rest of the page.
The changes will be rolled out seamlessly over the coming weeks, Facebook says, and it shouldn’t be the sort of thing you necessarily identify as happening. The goal, as always, Facebook says, is to simply put more content that you care about at the top of your page whenever you visit, and of course hopefully keep you coming back for more.