Virtually every digital publisher has seen two big trends in recent years: Traffic is up sharply, and mobile traffic is up spectacularly, often overtaking the desktop web. In both cases, the cause can be distilled down to a single word: Facebook.

But what Facebook giveth, Facebook can taketh away. (Just ask Zynga.) And 2015 is going to be the year that it starts doing exactly that.

Facebook has two natural constituencies: its advertisers and its users. It has no particular use for third-party publishers, except insofar as links to those publishers make its users happy, and keep them coming back for more such links. But Facebook is now a mobile company — most of its users are mobile, most of its revenue comes from mobile, and the proportions coming from the desktop are certain to continue to shrink. And in a mobile world, links to third-party websites are, to use a technical term, shit.

What’s more, Facebook’s algorithm is already working overtime on trying to slim down a virtually infinite range of possible News Feed posts to a much smaller number. A significant chunk of the NewsFeed is already ads, so in order to make it into the News Feed if you’re not an ad, you need to be really, really good. Like, one close friend announcing her engagement, or a video of another friend pouring a bucket of ice water over her head, or a long and hilarious comment thread on a third friend’s status update. What’s not really, really good? A link to some random website which has a user experience which Facebook can’t control, and which is probably suboptimal on mobile.

In 2015, then, the winners of the Facebook attention lottery are going to be more videos, as well as genuinely native, in-app content from advertisers. The losers are going to be external websites who have become reliant on the Facebook traffic firehose. That traffic is going to start falling, in 2015, for the first time. And the repercussions are likely to be huge.