If you’re like most people, the first thing you check when you log into Twitter—or Facebook, or Instagram—is the little number that tells you how many new notifications await you. As you skim those notifications, you might take note of how many replies, retweets, or favorites your latest tweets received. When you look at who interacted with you, an unfamiliar name or face might send you to their bio page to see how many followers they have. It continues in your feed, where each tweet comes time-stamped and pre-endorsed with a certain number of prior engagements by other users.
Now imagine all of that is gone. No numbers anywhere; no metadata to guide your assessments of self-worth, your estimation of a given tweet’s virality, or your judgments of other users. You’re left with nothing to go on but names, avatars, and the content itself.
It’s jarring. It’s confusing. It might make you yearn for the little numbers’ return. It will force you to reckon with just how heavily you relied on them and why they were so important to you. And then, at last, it might leave you more at peace—free to evaluate tweets and people and yourself by your own subjective lights, rather than according to the rules of the platform’s never-ending popularity contest.
That’s what the Twitter Demetricator does.
It’s a tool developed by Ben Grosser, an artist and professor of new media at the University of Illinois at Urbana-Champaign. He launched it last week as a sequel of sorts to the Facebook Demetricator, which he first published in 2012 as an artwork that invited people to consider how visible metrics influence their use of social media. (Grosser published a study on this in 2014.) It is most popular as a Chrome browser extension, though it’s also available on other platforms. You can download it and try it for yourself here.
The Demetricator was conceived as an art project and a social experiment. But at a time when Twitter and Facebook are rethinking their role in the world, it doubles as a potent critique of their platforms—one that reveals just how deeply ingrained are the features that make them so addictive, performative, and exploitable.
The sense that people are fed up with Facebook, Twitter, and other social platforms has never been stronger. The Atlantic’s Alexis Madrigal recently singled out “shareability” as the culprit in his feeling that Twitter is increasingly an exercise in condescension and outrage, and he got a friend to write a script that would effectively turn off retweets. The New York Times’ Farhad Manjoo—previously known for his prolific tweeting—unplugged from the platform and rediscovered newspapers. Facebook acknowledged in December that certain ways of using it can be bad for users’ mental well-being. And Twitter CEO Jack Dorsey last week admitted that his company “didn’t fully predict or understand the real-world negative consequences” of the service it had created. Both platforms have been accused of undermining civil discourse and allowing Russian agents to interfere in foreign elections. The mechanisms of virality, and the metrics that underpin them, are part of what allowed bots and trolls to amplify divisive messages.
For individual users trying to reshape their social media experience, Madrigal’s trick of turning off retweets holds some appeal, as does Manjoo’s decision to turn off social media notifications. But you might be surprised at how drastically you can change your Facebook or Twitter experience just by hiding a few little numbers.
Indeed, it’s almost impossible to comprehend just how central metrics are to the Twitter experience until you install Demetricator. Only when I tried it did I realize that my eyes were instinctively flicking to a tweet’s retweet and favorite counters before I even processed the tweet itself. Only when I tried Demetricator did I understand how much I relied on those signals to evaluate a tweet—not only its popularity or reach, but its value.
Without really meaning to, I’d been glossing over tweets that had relatively few likes and paying extra attention to those that had many. I had even subconsciously developed a sort of multiplier for various Twitter users based on the size of their followings, so that a tweet by a relatively obscure user that garnered 10 likes would stand out in my feed more than one by a famous user that got 100 likes. And in threads with lots of replies, I had been checking like counts as though they were an official scorecard of who was “winning” the conversation.
Absurdly shallow as that might sound, Grosser assured me I’m not alone in this. “People with Demetricator installed quickly realize that they were very competitive around these numbers,” Grosser told me. “I post a photo, somebody else posts a photo: ‘Why did they get more likes than I did?’ And they also realize that they feel compulsive around the numbers: You’re checking your notifications first thing, but checking them means you lose that little number, and the only way to get it back is to enter more content into the system.”
But the numbers’ impact runs deeper than just how people perceive content on Twitter and other social platforms. Grosser argues that it shapes, in profound ways, the content they create. “Part of what the metrics teach us is what to post in the first place,” he said. “Users create rules for themselves about what to write, how to act, what to say based on what the numbers show them. And they do this without realizing it.”
The moment Grosser said this, I realized it was true of me. Part of my internal calculation every time I post something on social media is: “Will people like this?” And, absent the ability to survey all of my friends or followers, I’d been relying almost exclusively on engagement metrics to answer that question. On occasion, I’ve even deleted tweets that received few favorites or retweets, on the assumption that something must have been wrong with them. David Zweig, writing about the Demetricator for the New Yorker last week, likewise observed that the numbers had been “guiding and constraining” his behavior, a process he described as the “robotization of users.”
That’s not to say that the social media experience suddenly becomes serene and fulfilling the moment you install Grosser’s tool. On the contrary, I found tweeting without the aid of metrics to be more anxiety-inducing, at least at first. I knew the metrics were still out there, and everyone could see them but me. Meanwhile, I found myself having to read each tweet in my feed more carefully, without the metrics to signal at a glance which ones were likely to be noteworthy and which ones might be skippable. Reading your Twitter feed without popularity metrics is like watching a network TV comedy without the laugh track.
And in some cases, the metadata is the message. Think of “the ratio,” in which a disproportionate number of replies to a given tweet suggests that the joke is on its author. Or a famous tweet by @dril, which is noteworthy only given the context that it’s viewed by many as seminal to the history of “weird Twitter.” Without the metrics, his viral “no” from 2008 is just another “no.” Or, more straightforwardly: It’s hard to fully grasp the impact of some Parkland survivors’ tweets without knowing that they’ve been retweeted hundreds of thousands of times.
That helps to explain why some users of Demetricator don’t stick with it very long: It makes using social media feel a bit like playing a game with one hand tied behind your back. That said, it’s very much worth trying, even if only to realize the extent to which social media really is a game—specifically, a popularity contest.
“Visible metrics don’t just draw our attention, but I argue that we are compelled by this deeply ingrained desire for more—a desire to make them larger simply because we can see them—based in an evolutionarily developed need for esteem,” Grosser said. “It intersects with capitalism in the way that capitalism treats value as a quantifiable thing, and there’s this endless need for growth.”
That might help explain why, for all their public soul-searching, companies like Facebook and Twitter have given no indication that they’d consider pulling back on their gamification of the user experience.
Twitter’s Dorsey was credited by many, including me, for the apparent humility of his recent tweetstorm pledging to rethink the service in fundamental ways, and soliciting ideas from outside experts. But it wasn’t lost on Grosser that Dorsey immediately turned to metrics as the path to a solution. The company’s request for proposals makes it clear from the outset that it isn’t interested in any kind of fix that can’t be readily quantified. In doing so, Grosser said, “Twitter presupposes the mechanism by which it would understand its own problem.”
In other words: If the root of Facebook and Twitter’s dysfunction lies in their commitment to quantifying value, then no metric or combination of metrics will ever fix them. As long as engagement metrics determine what people create, what people see, and how all of it is perceived, conversation on social media will never be “healthy” in the way Dorsey says he wants it to be. (To make this about Facebook instead of Twitter, just substitute “meaningful interactions” and “time well spent” for “healthy conversations.”) It’s a sobering thought.
Grosser said he isn’t necessarily arguing Facebook and Twitter ought to remove all their visible metrics. After all, he notes, both would still run on invisible metrics, which fuel the software behind both their feeds and their targeted ads. “I don’t propose Demetricator as a solution,” he said. “I propose it as an indicator of a problem.” That said, he added that he thinks that de-emphasizing certain metrics is at least “something we should be considering in any big rethinking or redesign” of social media systems.
After our long conversation about the perversity of quantifying social value, it seemed a bit rude to ask Grosser if he kept track of how many people had installed the Demetricator. I did it anyway, because reporters are trained to be rude. He said he has stopped keeping track of the number of people actively using Facebook Demetricator and knows only that it’s “between 10,000 and 100,000.” And as for the week-old Twitter Demetricator? He laughed, then paused for a moment. “On Chrome, it’s up somewhere in the 1,000 range,” he told me. “But I’d prefer you didn’t say it.”