The Washington PostDemocracy Dies in Darkness

Trump ‘shadow ban’ tweet: A F.A.Q.

President  Trump attends a change-of-command ceremony at the U.S. Coast Guard headquarters in Washington on June 1. (Jacquelyn Martin/AP)

There are Americans, I believe, who are largely insulated from the daily partisan back-and-forth that plays out on social media and cable news. Americans for whom the constant burbling anger that the Internet keeps in sight is at least pushed to the corner of their eyes. People for whom tweets like this one from the president of the United States must seem baffling.

Only about 1 in 10 American adults even gets news from Twitter, according to Pew Research Center data, though more than half hear a lot about President Trump’s tweets, in particular. So there is probably a substantial portion of the country who comes across this tweet, reads it and thinks, “What?”

Allow us.

What is “shadow banning”?

The phrase “shadow banning” is used in this case to describe a pattern in which certain Twitter accounts aren’t offered as suggestions when a user searches for them on Twitter.

Normally, if you go to the search bar on and type someone’s name, you’ll get a list of suggested accounts. Type “Trump,” for example and, as of writing, you get the president, the first lady and Ivanka Trump, among others. I get other options, too, like @trumphop, a bot I created that retweets old Trump tweets on the same day and time that he originally tweeted them. Which is to say: The results are customized.

So it’s not a ban?

It’s not a ban, hence the vague qualifier “shadow.” One of those who was being “shadow banned” was Republican National Committee Chairwoman Ronna McDaniel (who now shows up in a search). Had you gone to McDaniel’s Twitter page directly, you could have seen all her tweets and interacted as you wished. It was apparently just that some people weren’t showing up in Twitter’s search tool. Had you Googled “Ronna McDaniel Twitter” (certainly a more common way of finding her account), you would have been pointed to her Twitter page.

Why are prominent Republicans being targeted?

Well, the premise of your question is iffy. Instead, let’s start by asking why conservatives might have noticed this behavior.

The idea that conservatives were being muted or buried on social media platforms is a long-standing one. In 2016, Gizmodo reported that conservative news sites were regularly removed from that platform’s trending topics column, kicking off a firestorm. That was not the first example of perceived bias against conservatives, but it was a significant one. Since then, there has  been a tendency on the right to explore ways in which social media companies might be limiting their use of the platforms.

Put another way: There were conservatives who were either looking for or eager to embrace examples of social media platforms apparently stifling or sidelining their ideological compatriots.

So answer the question: Why are prominent Republicans being targeted?

It’s not clear that they are.

This debate was kicked off by a report at Vice News that explored allegations first raised by conservatives on Twitter. That article outlines the noticed behavior in the search tool and examples of those affected.

(We’ll note that the term “prominent Republicans” is a bit misleading. Is Donald Trump Jr.’s spokesman a prominent Republican?)

It also quotes New York Law School professor Ari Ezra Waldman, who testified about social media filtering before a House committee in April.

“This isn’t evidence of a pattern of anti-conservative bias, since some Republicans still appear and some don’t,” he said. “This just appears to be a cluster of conservatives who have been affected.”

We can go a step further. It’s not clear whether there is a similar cluster of people who can loosely be grouped as members of other political ideologies who are similarly affected. (There’s some indication that there might be.) This became a subject of discussion among conservatives, and a number of conservatives checked to see whether they were affected. It’s as if you went to a family reunion and noticed that you could all roll your tongues even though random passersby can’t. That doesn’t necessarily mean that this is specific to members of your family — and there may be a reason that your family has that in common.

What does Twitter say is happening?

Earlier this month, The Washington Post reported that Twitter was taking steps to clean up fake accounts and otherwise scale back negative activity on its platform. (To try, in other words, to turn down the heat on that constantly burbling anger.)

From our report:

“Rather than merely assessing the content of individual tweets, the company began studying thousands of behavioral signals, such as whether users tweet at large numbers of accounts they don’t follow, how often they are blocked by people they interact with, whether they have created many accounts from a single IP address, or whether they follow other accounts that are tagged as spam or bots.”
“Sometimes the company suspends the accounts. But Twitter also limits the reach of certain tweets by placing them lower in the stream of messages, sometimes referred to as ‘shadow banning,’ because the user may not know they are being demoted.”

Responding specifically to the Vice News article, Twitter’s Kayvon Beykpour linked the behavior to that change.

“In May, we started using behavioral signals and machine learning to reduce people’s ability to detract from healthy public conversation on Twitter,” he wrote you-know-where. “This approach looks at account behavior & interactions with other accounts that violate our rules.”

“It’s important to note,” he added, “that these behavior signals are not binary, and they are one of many other signals that factor into ranking. To be clear, our behavioral ranking doesn’t make judgments based on political views or the substance of tweets.”

So what does that tell us?

The suggestion from Twitter, then, is that automated processes look at, among other things, the responses certain accounts get to their messages. If you are blocked a lot or tweet at a lot of accounts you don’t follow, for example, Twitter raises an eyebrow. Reading between the lines of Beykpour’s tweets, it’s those users who are being flagged by the site’s “quality filter” and, therefore, may be downplayed in search.

In response to Gizmodo’s report about Facebook downplaying conservative sites in its trending topics, it was pointed out that many of the conservative sites being excluded from the trending topics were ones that often took an aggressive approach to politics and a casual approach to accuracy. The most popular site on social media among conservatives during 2016 was Breitbart. On Facebook, the third-most-popular site was Gateway Pundit, a site that regularly peddles untrue claims.

The outcry that followed the Gizmodo report had another effect, according to some observers: making Facebook balk at cracking down on iffy “news” sites that promoted Trump before the 2016 election. When Facebook eliminated human moderators, fake news with a conservative slant quickly appeared on the trending topics bar.

The implication, then, is that there may be patterns of behavior that are common among those who noticed the “shadow ban” which played a role in how Twitter treated them.

If this practice were targeting conservatives, would it be illegal as Trump claims?

It’s hard to see how it would be.

Trump likes to equate “bad” with “illegal.” In the past he has accused Hillary Clinton, Sen. Ted Cruz (R-Tex.), the Republican National Committee, Barack Obama, the State Department, Donna Brazile, former attorney general Loretta E. Lynch, Sen. Mark R. Warner (D-Va.), James B. Comey and the FBI of breaking the law. The boy sees a wolf.

In this case, though, Twitter is a private platform that can ban anyone it wants. It may be bad for business, but it’s not a violation of the law.

Why is the president tweeting about this, anyway?

Trump’s political instinct, which served him well during the 2016 campaign, is to embrace the concerns and frustrations of the vocal conservative base. This issue in particular has multiple points of appeal: It is a point of frustration among his conservative base (including his son Donald Trump Jr.), it positions his supporters as victims and it centers on technology companies.

That last point is important. There’s an ongoing effort — at times deliberate, at times not — to police the behavior of media companies to shape how platforms treat conservatives. Last month, The Post reported on a series of meetings and dinners between Twitter and Facebook and prominent conservatives and conservative groups aimed, on the companies’ end, at building better relationships with the right.

In response to the emergence of the “shadow ban” issue, Twitter CEO Jack Dorsey tweeted that it “suffices to say we have a lot more work to do to earn people’s trust on how we work.”

The problem for Twitter is that there’s a political utility to criticizing how the company treats conservative users. Positioning the company as being willing to silence certain voices (even in the search tool) positions those voices as endangered and important. There’s a reward to claiming to be targeted by a social media company that would encourage someone to assert in bad faith that some quirk in the code is actually a secret effort to keep them quiet.

That problem is harder to fix.