How social platforms learned to stop worrying and love the algorithm

The trend: As social media platforms continue their never-ending jockeying for user attention, personalized, algorithmically recommended content is making a comeback—and it’s been accelerated by TikTok.

How we got here: Five years ago, when the algorithm was discussed in media and marketing circles, it was often with anger or annoyance. Platforms had the power to prioritize or deprioritize certain types of content, seemingly on a whim, and control entire revenue streams in the process.

  • In 2018, Facebook’s announcement that it was tweaking its algorithm to show fewer publisher posts sent shock waves through the media industry. Meanwhile, YouTube’s algorithm became a byword for radicalization and misinformation.

What changed? TikTok is arguably the first major social platform to prioritize algorithmically recommended posts over those from accounts that a user follows. Users can follow zero accounts and still have an endless feed of content to scroll through on the algorithm-driven “For You” page.

  • TikTok explains its algorithm uses interactions like comments and likes, video info like hashtags and captions, and device settings like language preferences to curate a user’s feed.

That’s a radical departure from the follower-driven model used by most social platforms—and it has paid off. In late September, TikTok hit 1 billion monthly active users, reaching that milestone years ahead of other social platforms.

  • It’s also been hugely successful in driving time spent on the app. We estimate that US TikTok users spent 39 minutes a day on the platform in 2020—the highest time spent figure of any social network we measure—and will maintain its lead through 2023.

Why it matters: That success has led to other social platforms taking renewed interest in incorporating recommended content—and for TikTok parent ByteDance to take its algorithmic offerings off-platform.

  • Instagram and Snapchat both rolled out their own TikTok competitors, Reels and Spotlight, respectively, complete with fully algorithm-driven feeds.
  • YouTube engagement has been primarily driven by recommendations for a while, but in November it began testing, among select users, opening the app directly to its 100% algorithmic Shorts feed.
  • In June, Instagram also began adding recommended content on its main feed by way of “Suggested Posts,” meaning algorithmically recommended content is now present in all parts of the app.
  • Twitter began testing “Suggested Follows” in October of last year, and it also tested algorithm-driven content on the now-defunct Fleets.

Meanwhile, ByteDance has been building on its algorithm’s success off-platform.

  • In June, it launched a new enterprise software as a service (Saas) division called BytePlus. One of BytePlus’ offerings is a product called “Recommend,” which lets businesses apply the algorithm that powers TikTok to other ventures—for example, to uncover user preferences and feed them new product suggestions.

What’s the catch? Some platforms—Facebook in particular—are still scaling back, likely due to immense criticism over how algorithms contribute to misinformation and political divisiveness.

  • In April 2021, Facebook rolled out a host of new user controls, most notably the ability to switch to a reverse-chronological News Feed.
  • Instagram also announced in December that it's working on bringing back “a version” of its chronological feed, per The Verge.
  • Still, a reverse-chronological feed stops Facebook and Instagram from algorithmically selecting content from friends, Groups, or Pages that a user follows. But it won’t stop them from algorithmically recommending new content, such as “Suggested for you” posts or “People you may know” carousels.

The bottom line: Despite lingering criticism on the potential negative consequences of personalized feeds, TikTok’s success has cemented algorithms’ position on social media as something to revere instead of fear, with marketers even turning to algorithms to inform their activity off social media platforms.

"Behind the Numbers" Podcast