Rage Against the Algorithm
War has been much in the news lately, so it’s top of mind.
The recent revolt over recent comments by Twitter’s CFO about Facebook-style feed filtering feels like preparation for war. It’s a war we don’t want, but we are prepared to fight. If the battle is joined, it’ll be a war to save Twitter from Twitter, Inc.
We know that users (on all platforms) dislike change. And because so much of the Twitter experience was created and in a sense owned by users — retweets, hashtags, @ replies — Twitter users are “Twitter” in a way that Facebook users are not really “Facebook.” With this sense of ownership comes a deep (and sometimes reflexive and ill-advised) aversion to change.
But if the emerging revolt against algorithmic feed curation feels different, that’s because it is. Filtering isn’t an evolution of Twitter to match the needs of the user base, like so many of its other changes were. Teens who sign their selfies with an endless stream of hashtags don’t know or care that the hashtag was conceived of as a tool for the digerati to keep track of themselves at conferences circa 2007. The medium has needed to evolve to meet the needs of a new set of users, and it’s been Twitter (a.k.a. the users), not Twitter, Inc. who have evolved it. The early adopters revolted against features like “New RTs” because they undermined the quirkiness of Old Twitter. But by making it easier for users to use the service, Twitter, Inc. was smartly adapting to the needs of its new users in a time of rapid growth.
Curation is not like that. Curation is Twitter becoming Not Twitter. And if it moves forward, users will have to declare war to save Twitter, Inc. from itself.
Don’t Hate on Algorithms
Much of the criticism thus far has centered around a reflexive distaste for The Algorithm, as experienced every day via Facebook’s News Feed. And it’s true, The Algorithm has many flaws. It favors dumb “takes” over meaningful news. It doesn’t deliver bad news, because no one Likes bad news. It suppresses diversity of opinion. It doesn’t cover stories like Ferguson. It is easily gamed.
But there is a good reason Facebook is run by The Algorithm, and it has everything to do with how Facebook as a network works.
Why do we connect on Facebook? Because we have parents, siblings, high school friends, and former co-workers we want with whom we want to recognize some form of connection. Once that friend request is accepted, it feels awkward to break it. Unlike Twitter, both parties mutually agree to be connected to one another. It’s a generally recognized as a bigger deal to be Facebook friends than to follow someone on Twitter, and with this comes certain compromises.
Just because someone is our friend — or our “friend” — does not mean we think they will always have interesting things to say. In the spring of 2009, Mark Zuckerberg went through a Copy Twitter phase and made an unfiltered “real-time” News Feed the default. At the time, I and others were syndicating our Twitter feeds to Facebook. With algorithmic filtering turned off, a number of my Facebook friends began complaining about the volume of content. Something that was completely natural on Twitter didn’t feel right on Facebook.
If Facebook were unfiltered, you would probably want to blow your brains out even more than you do when visiting today.
Why do we connect on Twitter? Because we do think the other person has something interesting to say, whether we’re real-life or Facebook friends with them or not. And if they’re not as interested in us as we are in them, that’s cool — we can follow even if they don’t follow back, and maybe they’ll one of our witty remarks in their Mentions one day, and there will be some back and forth without a formal Friend relationship.
Because Twitter relationships are based on interests, unfollowing is easy without World War III breaking out, and it’s very easy to curate the right feed for us unencumbered by social niceties. This means the “curation” problem is solved on Twitter in a way it’s not — or ever will be — on Facebook. And this is why — even after seven years on the platform — I find Twitter to be a vastly more satisfying experience than Facebook.
Let Facebook Be Facebook, and Twitter Be Twitter
The right answer to all of this feels like: Let Facebook be Facebook, because we still have Twitter. There is no reason both networks ought to fill exactly the same role. They have evolved into their current state for a reason.
On Facebook, where interpersonal connections matter more, content in general is deprioritized, while it is king on Twitter.
However, when your old friend is having a baby, or just got a new job, it makes sense to look for this in the place where you keep your real-life relationships, so The Algorithm does its job when it puts this on top of News Feed.
And when you want gritty, unfiltered access to Hard News, you go to Twitter instead.
This is what has hardcore Twitter users terrified about even the remotest threat of a filtered feed. With Twitter under the spell of The Algorithm, the natural division of labor that’s taken shape between the two platforms will be broken, and there will be no place left to hide from The Algorithm. What used to a nice breathing space will turn into the suffocation chamber Facebook can be sometimes, where certain types of content just don’t ever get shown.
What Twitter’s Algorithm Might Look Like
A Twitter algorithm wouldn’t turn it into Facebook overnight. In fact, it’s likely that the basic inclinations of Twitter users would largely remain the same with or without an algorithm. It is the impulse to conform in front of our friends, and the ever-present “Like” button that makes The Algorithm act as it does on Facebook. Twitter users would still be snarky and ironic, and more attuned to the news media than they are on Facebook. A Twitter algorithm might very well still push stories like Ferguson to the top. Because follower circles can tend to be larger than Facebook friend lists, it’s also likely that a Twitter algorithm would be more accurate, and less likely to mistakenly penalize the right content than Facebook. This is not why I worry about it.
Twitter already has a reward mechanism for interesting content in the form of the retweet. News can spread democratically from social circle to social circle, and do so quite quickly. As a watcher of minute-by-minute Twitter data, I am consistently impressed by how well this system works. This is how quickly news spread after the arrests of two reporters in Ferguson:
When significant news breaks, it takes between 8 and 10 minutes from the first tweet for it to reach peak tweets-per-minute (TPM). This happens consistently, almost without fail. During the World Cup, Twitter was able to monitor significant variations in tweets-per-second during penalty kick shootouts.
If everyone’s tweets were guaranteed only limited distribution, the real-time alert nature of Twitter would still continue to work, but prophylactically. The Algorithm might take a minute or two to sort out “Is this tweet really in the top 15% tweeted from this account? Should I show it to a few more people to test it out?”
And this the numb of the problem for me. This is a basic violation of Twitter’s promise to its users: unfettered real-time communication. A tweet of mine may interest no one, but it at least gets a fair shot by getting shown to everyone who is on at the time. With equal opportunity distribution, we can know with more certainty whether a message reached its maximum potential. Because it’s public by default, it can get distributed to people I don’t know much more easily than Facebook. And unlike Facebook, where I can usually hear the gears working (show it to 5 people, 1 likes it, show it to 15 more, 3 like it, show it to 100 more, etc.) viral growth on Twitter is meritocratic and real-time.
There is room for curation in the Twitter universe. Retweets are already part of it, if only additive to the core Twitter experience. The Discover tab is also a rich source of information, if sanitized and corporate, and lacking in the unique voices I normally receive in the main feed. I also use Nuzzel to pinpoint which stories have been shared the most by my follows (and Facebook friends too).
We’ve Reached Peak Twitter — And That’s Okay
This is not what investors want to hear, but here it goes: Twitter IPO’d too late for public market investors to see significant upside growth. The market has been perturbed by the notion that Twitter’s MAUs have remained stagnant, and the company is casting about for ways to respond. But the fact is that both Twitter and Facebook have likely achieved maximum saturation in the United States, and eight and ten years in respectively, there probably isn’t a market in the world where either is seen as the Next Big Thing, not with the spread of single-purpose social apps and phone contacts as the new social graph.
In this dynamic, Twitter is tapped out at around one fourth the size of Facebook, at least as far as the United States goes. Facebook was conceived of as a directory and Twitter as a new form of communication. It’s much easier to be listed in a directory than be expected to communicate. Even if people do communicate on Facebook, it is its stature as the ultimate social directory that keeps it the 800 lb. gorilla of social. And while directories require at least majority participation, communications channels can succeed without the same network effects. There are plenty of successful communications platforms even smaller than Twitter (i.e. Reddit).
This means that aside from squeezing more revenue out of its existing user base, there are no easy answers for Twitter, and this has led to somewhat strange insider speculation about what its future might hold. First, it was de-emphasizing the @ reply and the hashtag (though the latter is huge on still more emergent platforms like Instagram or Twitter-owned Vine). Now it’s a possible trial balloon about curation. At first, the Twitter’s problem was that new users weren’t following anyone and dropping off the platform as a result; now it’s that feeds are too cluttered. Twitter’s problem may actually be that it doesn’t know what its problem really is — or more maddeningly still, that it already works perfectly already.
