When the Facebook Traffic Goes Away

In newsrooms in six countries scattered across the globe, alarm bells started to go off over the weekend: Something very strange was happening to the newsrooms’ posts on Facebook. Instead of appearing in the News Feeds of people following them on the social network, the posts were appearing in a new, separate section of the site, termed Explore Feed.

Facebook in Slovakia, Sri Lanka, Serbia, Bolivia, Guatemala, and Cambodia had begun to function differently, separating out posts from people and posts from pages.

On Saturday, Filip Struhárik, a journalist at the Bratislava-based newspaper Denník N, published a warning on Medium that spread quickly among social-media managers and Facebook observers. “Pages are seeing dramatic drops in organic reach. Reach of several asked Facebook pages fell on Thursday and Friday by two-thirds compared to previous days,” he wrote. “Sixty biggest Slovak media pages have 4 times fewer interactions (likes, comments, shares) since the test. It looks like the effect in Guatemala and Cambodia is the same.”

Several headlines have since focused on another part of the test: that pages could pay to appear in the main news feed. Mashable called it “a nightmare” pay-to-play scenario.

Needless to say, publishers were worried. In response to Struhárik’s story, the head of Facebook’s News Feed, Adam Mosseri, responded to him on Twitter. “This image reflects a test in Slovakia, Sri Lanka, Serbia, Bolivia, Guatemala and Cambodia,” Mosseri wrote. “It’s not global and there are no plans to be.”

When Struhárik followed up to ask how long the test would take—“days, weeks, months”—Mosseri replied, “Likely months as it can take that long for people to adapt, but we’ll be looking to improve the experience in the meantime.”

Publishers outside these six countries could breathe a sigh of relief. The effect of this kind of drop in traffic, particularly in the fourth quarter, when many (American) publishers have sold through a higher percentage of their “inventory,” would be devastating.

But for those inside those countries, newsrooms were, to put it gently, freaking out.

Marko Miletić works for the Serbian site Mašina, which saw its Facebook pages reach 58 percent fewer users last week as the test began to roll out, and 72 percent fewer interactions. “We don’t know how long this test will last but it can influence informational and political pluralism (at least the amount that existed on social network),” he wrote to me. That’s because, as he sees it, Facebook’s organic tools are all that smaller media publishers and “grassroots political initiatives” have. They can’t afford to pay for distribution on Facebook by “boosting” posts.

“Small (usually political-alternative) media or citizens’ initiatives, activist campaigns, solidarity actions, etc. that don't have enough money for sponsoring posts will have hard time to reach people through Facebook, which is their main tool of dispersion,” Miletić said.

Other publishers are trying to put a happier face on the changes, or at least withholding immediate judgment.

“It’s too early to say anything definitive about the impact this is having on our traffic and reach,” Jenni Reid, the web editor at The Phnom Penh Post, told me. “The two feeds still don’t seem to be fully separated yet for some people here in Cambodia, but so far it doesn’t look positive.”

Reid, who was a social-media editor for The Economist before her current position, said that it was hard to imagine how the Facebook change might positively affect their number of readers. “In its current format, it seems hard for me to see how the Explore Feed could be anything other than bad news for publishers,” she said. “It seems a strange move given that Facebook has been trying to build bridges with news organizations since the start of the year through the Facebook Journalism Project.”

These changes are significant for the broader media ecosystem in Cambodia, Reid said. “Last year, Facebook edged ahead of television as the number-one source of news for Cambodians according to one survey. Post Khmer, the Khmer-language Facebook page for the Phnom Penh Post, has the fourth-most likes in the country, and seven out of 10 of the most popular Facebook pages here are news websites or newspapers,” she told me. “That’s striking compared to, say, the United States, where there isn’t even a news publisher in the top 50 most popular pages among Facebook users.”

Reading between the lines, it’s clear: Cambodia’s news infrastructure experienced a radical change, overnight. And none of the editors I was able to contact, or anyone that they knew, had heard from Facebook about the change before it happened. They just walked into work one day and everything was different.

It’s possible that, in the long term, separating page posts out from people posts would be a good idea. News Feed has serious problems. Maybe the split feed would be a superior experience. But to see how it would be before making a huge change, one can imagine Facebook would want to do some major testing.

From Facebook’s perspective, the company has to be allowed to try out new versions of its software. It can’t be asked to keep its tools static because publishers have gotten used to them. And some tests might need to be disruptive to get to a future, better Facebook. This iterative process is, in fact, how Facebook has built the product that so many people use for an average of more than 50 minutes per day.

But Facebook did not simply end up controlling news distribution in countries across the world. They strategically entered the market, much as any company would, as part of their own competitive battle with other internet companies. Some responsibility must come with the deliberate rerouting of the public sphere through Facebook’s servers and ad networks. Right?

None of this is to get at the content of the test. Vox’s Matt Yglesias argued that the change could be good. “Facebook-induced traffic boom just devalues page views,” he tweeted. Total ad spend in the country is fairly fixed, so a decreasing number of page views should lead to rising advertising prices for publishers over time, he said. Certainly, when Facebook started sending more traffic to publishers, we saw the reverse: Ad rates declined as there was simply more supply. The reverse could be happen (although I can’t imagine there is a digital-media sales team out there who wishes they had less inventory to sell).

In any case, if you’re a publisher in one of the six affected countries, that must be cold comfort.

Facebook has been unusually forthcoming about this test after it was initiated, which is good. Mosseri responded on Twitter and followed up with a blog post that partially explained their rationale. “The goal of this test is to understand if people prefer to have separate places for personal and public content,” he wrote.

But Facebook’s response has been strangely blind to the bind that it has put whole countries’ worth of publishers in. Mosseri’s response to a Slovakian journalist was to assure him that there were no plans for the test to go global.

Facebook executives like to argue that people don’t understand the company’s staggering scale, but the test shows how that very scale might estrange it from how other people see the world.

Are all the people in the test areas some tiny percentage of Facebook’s user base (say, 1 percent) who want to see more posts from their family and friends, or are they the citizens of six sovereign nations who have come to rely on Facebook as a crucial part of their news-distribution infrastructure?

They’re both. And in this case, their news publishers are simply stuck reckoning with internet-company power, desperately tweeting.

I put these questions to Facebook: Were media organizations notified of, briefed about, or consulted with before the change was made? How were these countries selected? What deliberations were made about the possible trade-offs between what you’d learn and the probable deleterious effects on the media business and information ecosystem in these places?

Facebook has not responded.