How to fix Twitter
Massive outcry recently over Twitter’s idea to become even more like Facebook in the hopes of becoming financially more like Facebook.
Copying Facebook
Advertising revenue has become less valuable over the years, while personal data has become even more valuable. But laws such as the UK’s Data Protection Act forbid companies from retaining more data on a customer than is necessary for their business purposes.
Facebook’s solution to this is was to modify their site to make retaining personal data necessary for the site’s core functionality of displaying the user timeline. It remembers which posts you “Like”, how long you spend reading each post, which people you interact with most, and so on.
It then uses all this data to re-order the posts on the user’s timeline algorithmically, instead of chronologically, which was controversial at the time because it allows Facebook to decide what content you should see, potentially creating a censorship or propaganda system in what used to be an open peer-to-peer communications system.
Twitter is less junky
While it does solve the problem of Facebook timelines full of endless junk content, it creates alternative problems. Content is no longer in any meaningful order. Important messages from people you don’t talk to often will never appear on your timeline. Facebook becomes an echo chamber of people who agree with you.
Twitter avoids much of the Facebook problem already, by design. Tweets are shorter than Facebook posts, so you can read more messages with less effort. Twitter follows are non-reciprocal and accounts are generally public, so following an account signals “I want to read this content stream” instead of “I know this person” or “I like this brand”. There is no real name policy or expectation that your connections map existing real-world relationships.
This means Twitter users only follow accounts that they want to read. It doesn’t suffer so strongly from Facebook’s junk timeline problem because Twitter isn’t a social network; it’s a real-time aggregator of content streams, some of which happen to be social connections.
Content is not users
The Google+ project failed to understand this when they attempted to merge with Youtube. Youtube accounts are content streams, not people. Some people will follow an account because it’s run by their friend, but the most significant accounts are popular outside of the owner’s immediate social group because of what the content is, not who the creator is, even if the creator’s presence is a significant part of that work.
Google+ wanted to force Youtube to become more like a social network to be more like Facebook, in a sort of cargo cult attempt to be as popular as Facebook. In doing so, they tried to redefine a Youtube account from a content stream to a person, which was a complete failure because they’re totally different things. Youtube succeeds precisely because an account is not a person. It was nothing more than a crude attempt to pull users from a popular site to an unpopular one.
Twitter is not Facebook
Now we see Twitter at risk of similar mistakes. First, Twitter “Favorites” have been redefined as “Likes”, quite probably for no reason than Facebook does it. This isn’t a huge change, but with recent talk of algorithmically ordering timelines, there’s considerable concern that Twitter is aping Facebook in a panic for ways to grow financially and please investors.
Non-chronological timelines are the big issue. Part of Twitter’s identity is that it will tell you what’s happening now, rather than just give a summary of the best stories of the week. Calculating what content a person likes best is difficult to do effectively, and in a one-dimensional timeline it means erasing other more current tweets from accounts that the user intentionally subscribed to for their content.
Privacy, censorship and usability
Another concern is that if Twitter’s goal is not merely to copy Facebook but to make it necessary for Twitter to collect personal data to function correctly, users may be forced or nagged into using the non-chronological timeline against their own preference, as they are on Facebook. This also poses issues with regard to privacy, censorship and surveillance.
For many users, the user interface issues are more important than privacy, and scrambling the user’s timeline while hiding content they wanted to see is a big problem. It’s also an issue for the content creators Twitter relies on, who can no longer rely on their content reaching their followers.
Even casual Twitter users may find that the system is effectively muting them from their friends, which is a direct violation of the purpose of the system. It impedes the core function of Twitter.
How Twitter finds problems
Twitter succeeds by doing what its users like. That means keeping features that they enjoy, and only adding new ones that don’t detract from the core functionality. Changes to Twitter must only be made to solve problems the users have or improve the user experience.
Twitter has a few current user issues. One is that you naturally miss out on content that happened when you were away from Twitter, normally because you were asleep, busy, or otherwise occupied. Another is that some users follow a huge number of accounts, sometimes by engaging in “like for like” schemes for a sense of popularity, or by copying follow habits prevalent on social networks.
Some just find there’s a lot of good content on Twitter and want to follow it all. In any case, the problem is that some people have too much Twitter.
What Twitter shouldn’t do
The problem with algorithmic timeline as a solution is that Twitter isn’t Facebook. Twitter is real-time; algorithmic timeline destroys that. And Twitter has a generally good signal to noise ratio (no expectation to follow accounts whose content you don’t like, no social disincentive against unfollowing accounts later), so the cost of hiding content is much worse than Facebook.
Twitter users are right to fear the algorithmic timeline. It would be even worse for Twitter than it was for Facebook.
What Twitter can do
Twitter should create a new “catch-up” tab, complementing the live Twitter experience users know now with a secondary timeline that summarizes recent Twitter. Unless users are forced or pushed into making the catch-up tab their primary Twitter experience, this new feature will solve a lot of problems without the risk of alienating Twitter’s existing userbase by destroying core functionality.
Some third-party Twitter clients already offer the ability for users to follow multiple streams at once, such as the standard timeline on one column, notifications on another, and a search in a third column. Those users could easily accept a catch-up column.
Mobile users, web users and others could easily check the catch-up tab once when they wake up, or after any period of time away from Twitter.
Twitter should also ask itself whether it could sell premium accounts to large numbers of users as a direct income source, and what additional new features would be reasonable for such an account.
Update (August 2025): Following its acquisition by Elon Musk, Twitter (now X) actually did switch to selling premium accounts as a major revenue source. Contrary to my prediction, the institution of the algorithmic timeline as the default did nothing to dissuade most users, although some turned to superstitious magic words as a solution. The most recent user exodus, primarily to Bluesky, had nothing to do with site functionality, but rather political disagreement with the site’s owner.