In the case of the new breed of viral content sites taking over the internet, the controller is Facebook. Sites like Buzzfeed and Upworthy -- and a growing number of imitators -- have all but perfected the science of producing stories that Facebook users can't resist sharing with their friends and "liking." With 1.2 billion users, Facebook offers a platform big enough to build an entire business on: Upworthy, which said it had 87 million unique visitors in November, gets more than half its traffic from the social network, while Buzzfeed, with 133 million uniques, counts on it for about one-third of referrals.
It would therefore be potentially devastating to these sites -- and to the entire category of what I'll non-judgmentally label virality mills -- if Facebook did anything to make it more difficult for them to disseminate their content through its users' newsfeeds. But some people think that's exactly what it's trying to do.
Last week, Facebook said it was making new changes to the algorithm that governs what shows up in users' streams. The changes are meant to ensure users see "high quality articles" instead of just "the latest meme."
The Washington Post's Ezra Klein was among those who saw the move as as response to the rising tide of virality mills. "As a genre, this stuff is flooding into Facebook so fast, and it's so much more effective at getting shared than anything that came before it, that it seems almost certain that there'll eventually be a correction in the algorithm to keep it from taking over news feeds entirely," Klein wrote.
Peter Kafka of AllThingsD went a step farther, quoting sources who called the update "Facebook's Panda." Panda was the name of an algorithm update Google implemented in 2011 to improve the quality of search results. It was widely seen as a response to the tactics of so-called content farms like Demand Media, which were using armies of freelancers to produce cheap content in response to fluctuations in web search trends. Asked directly whether Facebook is trying to do something similar, only with Buzzfeed and Upworthy standing in for Demand and Associated Content, Newsfeed product manager Lars Backstrom told Kafka, "We don't have any sort of specific enemies or targets."
I'll expand on Kafka's history lesson. When Google implemented Panda, it, too, resisted attempts to characterize it as "a dagger aimed at the heart of companies like Demand," as Mathew Ingram put it at the time. Google's reticence allowed Demand CEO Richard Rosenblatt to claim his company wasn't a target, and even to intimate that Google saw Demand's output as the sort of high-quality content Panda was meant to promote.
Events quickly proved otherwise. These days, Rosenblatt is out as CEO and Demand's stock is trading at less than 25% of its 2011 peak level. Content farming is back to being a marginal publishing strategy.
Is that an overly dire parallel to suggest? After all, Backstrom tells Kafka the Newsfeed update will only have "a 10 to 20% impact." That doesn't sound like a lot -- until you consider that Google said Panda would affect about 12% of search results.
Whatever Facebook's true intentions, you can be certain that every publisher that relies on the platform for any substantial amount of its traffic will be carefully parsing every utterance coming out of Palo Alto for clues to what the new Newsfeed construes as "quality content."
With that in mind, here's something to think about: According to a post from August, Facebook's engineers began this process by surveying thousands of users to solicit their views on what defines quality. That's something of a departure for a company that typically bases its decisions on the firmer data of user behavior -- what people actually do versus what they say they think.
Implicit in the decision to take a survey was the understanding that just because people will read something doesn't mean they think it's good. We may follow a hyperlink because it's right there in front of us as the headline has been A/B tested to the point of irresistibility, just as we may continue to eat Pringles until the sleeve is empty, while in both cases wishing for something more nutritious to save us from our own compulsions. "Content that gets clicked on is high quality by definition," says Lightspeed Ventures partner Jeremy Liew, but Facebook's actions suggest it thinks otherwise.
"[T]he surveys are not necessarily the truth," Backstrom says. "But it's just as naive to treat every single click as having the same value."
Ironically, perhaps, Buzzfeed, and by extension the entire virality mill genre, has its roots in a similar discrepancy. Its founder, Jonah Peretti, has often talked about the difference between the content people seek out through search versus the content they consume via social. People use search to find information they'd rather not talk about with friends -- diet pills, celebrity nipple slips, etc. -- and social to publicize their consumption of content that they think reflects well on them -- eg. "11 Incredibly Powerful Letters From History."
How big is the delta between the kinds of content we want to be seen as consuming and the content we actually like to consume? The answer to that question may determine who benefits from Facebook's recent moves and who loses out.
