Twitter—the version of it that I see in my media-heavy timeline, anyway—is in a tizzy over the future of Twitter. The site’s most ardent users fear that it’s on the verge of abandoning its most distinguishing feature: a timeline that shows every single tweet from everyone you follow, in order of how recently it was tweeted.
They fear, in other words, that Twitter is about to become a lot more like Facebook, which uses secret machine-learning algorithms to decide which posts its users see and which they don’t. Those algorithms are at once the most widely reviled feature of Facebook—the reason its news feed is choked with viral memes and baby photos—and the key to its staggering popularity.
For Twitter to adopt similar algorithms would be a major change in strategy. It’s a change that could vastly broaden its reach. But it would come at the cost of infuriating its most loyal users.
The proximate cause of this fear is a series of comments made by the company’s chief financial officer at a recent tech conference, reported on Wednesday in the Wall Street Journal. Here’s the paragraph that inspired the panic:
Twitter’s timeline is organized in reverse chronological order, a delivery system that has not changed since the product was created eight years ago and one that some early adopters consider sacred to the core Twitter experience. But this “isn’t the most relevant experience for a user,” [CFO Anthony] Noto said. Timely tweets can get buried at the bottom of the feed if the user doesn’t have the app open, for example. “Putting that content in front of the person at that moment in time is a way to organize that content better.”
Noto tried to cushion the blow a bit, saying the changes would be implemented carefully and incrementally. “Individual users are not going to wake up one day and find their timeline completely ranked by an algorithm,” he said. But that only reinforced the impression that users are going to wake up one day and find their timelines at least partially ranked by an algorithm.
In fact, in a small way, they already have: Last month Twitter began occasionally peppering users’ timelines with tweets from users they don’t follow. CEO Dick Costolo clarified this week that this is meant to happen only in a very specific circumstance: when a user refreshes her timeline at least twice in a row without finding any new tweets from people she does follow. Still, to those attuned to such changes, it feels like the first stumble down a slippery slope to Facebook-style filtering.
Fortunately, I don’t think the slope is as slippery as it seems.
“Filtering” is a loaded word in social media circles. It’s one thing for content to be highlighted, reorganized, or added to your feed. “Filtering” implies that some of it is actually being hidden, and that spooks people.
Facebook’s news feed does this. For years it has relied on sophisticated algorithms to display only those posts they deem most relevant to a given user. You can scroll down all you want—the Washington Post’s Tim Herrera recently did it for six hours straight—but the majority of new posts from your friends and the sites you follow will never appear.
Facebook’s reasoning is that if you saw every post from every one of your friends, the vast majority of them would simply bore and annoy you. And, ultimately, Facebook itself would bore and annoy you.
For most people, that’s absolutely true. It’s a function of Facebook’s dual role as a social news feed and a digital address book. You friend some people on Facebook because you’re interested in what they post, sure, but you also friend a lot of others just because you’re, well, friends. Or family. Or co-workers, or acquaintances, or whatever. The point is that just because you friend someone doesn’t mean you care to see his every status update from now until death do you part. That’s why Facebook engages in filtering, despite the relentless criticism it endures as a result.
Twitter is different. You can follow your friends, sure, but you’re not required to do so just because they follow you. A lot of people use it primarily to follow celebrities, comedians, and news outlets as opposed to, say, their 10th-grade classmates and second cousins. As a result, Twitter has been able to get away with showing people every tweet from everyone they follow without losing its appeal. The raw, unfiltered flow imbues the service with a sense of transparency and immediacy that Facebook can’t match.
The straightforward reverse-chronological order is ideal for newshounds who refresh their timelines continually throughout the day. Unfortunately for Twitter, newshounds form only a fraction of the populace. Most people would rather sip their news and commentary from a water bottle than slurp it from a fire hose. And when big news breaks, even the most avid tweeters can get frustrated with the noisy, repetitive, and fragmented flow of the collective conversation.
My raw Twitter feed is now nothing but people talking about filtered Twitter feeds. I’m starting to think it might not be the worst idea.— Neetzan Zimmerman (@neetzan) September 4, 2014
This is the dilemma that has loomed over Twitter ever since the company announced its plans to go public. With that decision, Twitter committed itself to maximizing shareholder value, and while there are a number of possible ways to achieve that, the most obvious is to enlarge its active user base. Facebook’s example suggests that algorithmic sorting could help it do that. For casual users, Twitter would be a lot less intimidating if it automatically displayed a bunch of stuff they were interested in as soon as they logged in. And it would give Twitter’s developers much more freedom to come up with ways to better organize conversations about specific topics.
So, yes, Twitter is bound to get at least a little more like Facebook as it tries to expand. But here’s the thing: Twitter already uses algorithms. Its timeline already is not strictly reverse-chronological, thanks to retweets, sponsored tweets, and those little blue lines that connect conversations among users. And, crucially, introducing algorithms to Twitter’s timeline need not mean that “a Facebook-style filtered feed is coming, whether you like it or not,” in the alarmist phrasing of Gigaom’s Mathew Ingram.
In fact, if you want to get technical about it, Twitter already has a Facebook-style filtered feed. It’s called the “Discover” tab, and it sits third from the left among the four tabs on Twitter.com, behind the default “Home” and the indispensable “Notifications.” On the latest version of the Twitter mobile app, it’s in second place.
At this point the tab seems to be neglected by the majority of Twitter’s users. That may be partly because its algorithms aren’t particularly good yet. (They seem to focus mainly on what others in your circle are sharing or favoriting, without much thought given to your own demonstrated preferences.) But it’s also partly just a function of how people use mobile apps: They tend to spend most of their time on the default home screen.
It’s easy to imagine Twitter refining, rebranding, and eventually moving the Discover tab up in the priority list, perhaps even to the default position for new users. If it does, however, I’d place a strong bet on the company retaining some version of its current reverse-chronological feed for power users, and making it easy to toggle between the two.
Another alternative—more likely, perhaps, given the recent changes and Noto’s comments—is that Twitter could continue to tweak the organization of the Home tab, including adding algorithmically suggested tweets to users’ timelines. But there are all sorts of ways it could achieve this without “Facebook-style filtering.”
For instance, it could add a small number of “in case you missed it” tweets to the top of your timeline each time you log in, while keeping the rest of it largely intact. It could occasionally alert you to tweets from people you don’t follow if it thinks you’ll be particularly interested in them. It could enlarge or otherwise highlight tweets from people you do follow that are getting a lot of favorites or retweets from others in your circles. It could work on better grouping related tweets according to their subject matter. It could even experiment with hiding some tweets from users whose tweets you haven’t engaged with in a long while, yet retain an option to easily display them again if you so desire.
The latter option might be fairly labeled “filtering,” but it would remain a far cry from Facebook-style filtering, in which the majority of your friends’ posts never show up at all. It could be more analogous to the spam filter in your email inbox. And as long as it could easily be toggled to the “off” position, it wouldn’t raise the same kinds of free-speech concerns as Facebook’s news feed, which was criticized in the wake of the Ferguson protests for algorithmically burying the story.
In short, Twitter can continue to tinker with its timeline in subtle, incremental ways without ever losing its underlying reverse-chronological structure. It can do this because the people you follow will always be a different group from the people you friend on Facebook. Really, it’s Facebook that’s hemmed in to a heavily filtered feed by the fact that it began as a social network rather than a place for status updates. Without the baggage of doubling as a 21st-century phonebook and photo album, Twitter is free to borrow from Facebook’s asynchronous, algorithmic playbook without abandoning its focus on real-time news and discussion.
What Twitter’s critics forget is that outsiders aren’t the only ones deeply attached to the service’s core structure. It’s also highly valued by many within the company, including at the highest levels. And so the company agonizes internally over every little change to its original functionality, trying to patch its biggest weaknesses without sacrificing its core strengths. Twitter needs to keep growing, but it also knows that becoming a Facebook clone would be a losing strategy in the long run.
Twitter isn’t going to stop feeding the newshounds—it just wants to throw a few more bones to everyone else.