Why Facebook’s survey about trust won’t make or break the media
Last Friday, Facebook announced two key changes to the News Feed: one, it would reduce the amount of news in the News Feed from roughly 5 percent to 4 percent; and two, it would begin taking into account how trusted a publisher is when ranking it in the feed. Trust levels are to be determined by a survey, and today BuzzFeed’s Alex Kantrowitz got ahold of it. The survey consists of two questions:
Do you recognize the following websites?
How much do you trust each of these domains?
– A lot
– Not at all
Journalists noted that this was not a particularly comprehensive survey. ”People could game this survey,” said the FT’s Hannah Kuchler. ”I’ve filled out more robust surveys at fast food restaurants,” said Recode’s Rani Molla.
Bloomberg’s Sarah Frier had follow-up questions: “This is kind of like a brand awareness survey, like an advertiser would run,” she tweeted. “Trust in news is much more complicated. How well-sourced is the article? Are other sites verifying it? Is it news or analysis?”
The anxiety here is that survey results could be inaccurate, leading to a mis-ranking of publishers that favors the most partisan sites and exacerbates Facebook’s negative effects on democracy. For journalists, there is a secondary, existential fear: that publishers who fare poorly in the survey will see their traffic collapse, leading to declining revenues and eventually layoffs.
This is the best you can come up with?
A two-question survey does little to assuage these fears. To some, it suggests a lack of thoughtfulness on Facebook’s part. This is the best you can come up with?
But there’s a missing piece of the puzzle here, and it’s this: no one knows how important the survey rankings will be to the distribution of news on Facebook. Here’s what News Feed chief Adam Mosseri said last week about the company’s methodology: “We surveyed a diverse and representative sample of people using Facebook across the US to gauge their familiarity with, and trust in, various different sources of news. This data will help to inform ranking in News Feed.”
The same post mentions two other signals the company plans to, in its words, “prioritize”: whether the content is considered “informative” — something else Facebook attempts to determine with user surveys — and whether the content is relevant to you locally.
Facebook did not (and likely won’t ever) say how those factors interrelate. Is one significantly more important than the others, or are they all equal? It also won’t tell us how those priorities intersect with other known signals that Facebook takes into account when determining where a story falls in your feed: how many people click on it, how long they spend reading after they click, whether they share or comment on it, and so on.
As Mosseri tweeted today: “How we incorporate survey data is every bit as important as the specific questions we ask.” I asked him if he could say any more. “Trust is one among many signals,” he said, “but only applies for publishers for which we have enough data, so it doesn’t yet affect most publishers.”
“This isn’t a simple vote.”
He went on: “The other important thing to understand is this isn’t a simple vote. We are not just valuing more publishers that a lot of people trust, but rather valuing more publishers that a lot of different types of people (based on reading habits) trust.”
I asked how much effect a trust score might have on how often their stories pop up in the News Feed. “How trustworthy people feel a publisher is a actually a big signal for publishers for which we have data, and not a signal at all for publishers for which we don’t,” Mosseri said. “It’s an important signal with limited coverage.”
That coverage will expand over time. Meanwhile, the most important thing I take from these comments is that trust is just one of many signals — and how those signals interact is unknowable. Facebook will talk to us directionally about where the News Feed is going, but when it comes to the most relevant particulars, it is stubbornly silent.
There are competitive reasons to do this, of course, and public-relations ones as well. The publication of the survey questions today led to an immediate round of second-guessing. Most people I’ve met have strong opinions about how their own News Feed could be ranked better. The more Facebook discloses, the more doubt it faces.
Until recently, none of this has mattered. Time spent in the News Feed increased, and Facebook’s profits soared. Then came 2016 and the reckoning that followed. Time spent in the feed is now reportedly declining. And the more consequential we perceive the News Feed to be, the more we rankle at the perpetual mystery surrounding its construction.
Maybe publishers’ trust scores will determine whether they live or die. But I’d be very surprised. It seems more likely that there will be a handful of modest winners, taking a slightly larger share of a shrinking universe of clicks. The rest will see the trend highlighted by publisher-analytics firm Chartbeat in a blog post today: “Since October 2017,” the company wrote, “we have seen an almost 15 percent decline in Facebook referrals across our client base.”
All the trust in the world is unlikely to bring those referrals back.
Go deeper with Facebook
This editorial is taken from The Interface, Casey Newton’s daily newsletter about social media and democracy Subscribe now
Next Up In Tech
- YouTube is asking its promoted artists not to insult the company
- Google is working on bringing AR to Chrome with downloadable 3D objects
- Stripe will end bitcoin support because customers aren’t interested
- Terrifying baby robot has crawled right out of my dirt nightmares and into my dirt heart
- Vine successor may arrive this summer with focus on community building
- It’s 2018 and Alesis has debuted a new wireless keytar