Here's how TikTok algorithm works and why it beats competitors

It is quite literally the trillion-dollar question: in only a few short years, how did TikTok emerge from a specialized social network for lip-syncing teenagers to the most widely used app in the western world, threatening to completely unseat Facebook?

There are countless solutions, and TikTok owes its incredible success to a variety of wise decisions: A vast library of licensed music allowed teenagers to soundtrack their clips without worrying about copyright issues. A billion-dollar advertising campaign across Facebook and Instagram bought new users as quickly as Zuckerberg's company would send them over. Simple video creation tools blurred the line between creator and consumer far more than YouTube had ever managed.

However, the company's celebrated "For You Page," or FYP, and the algorithm that populates it, are the most effective tools TikTok has to attract customers and keep them interested.

The FYP screen opens by default for new users when they launch the application. Even if you don't follow a single other account, you'll still notice that it is immediately filled with an endless stream of small video selected from what is trending across the service. A Facebook or Twitter account with no friends or followers is a lonely, lifeless place, but TikTok is engaging from day one, giving the company an advantage over the competition.

But the company's secret ingredient is what occurs next. The selection of videos you see as you navigate through the FYP steadily starts to change, and eventually, according to the app's regular users, it starts to forecast which videos from other parts of the site will catch your attention.

On the surface, the business is disarmingly transparent about how that algorithm functions. In 2020, the company stated that "recommendations are based on a number of factors," including "user interactions such as the videos you like or share, accounts you follow, comments you post, and content you create; video information, which may include information like captions, sounds, and hashtags; [and] device and account settings like your language preference, country setting, and device type."

According to Chris Stokel-Walker, author of TikTok Boom, it is unclear how these numerous inputs are weighted and what specific characteristics cause any given video to appear in your feed. There is no recipe or secret formula behind it, one TikTok employee who is in charge of tracking what goes viral and why told me in my book. That the employee acknowledged that "I don't think even the algo team have the answer to it." Simply put, it's incredibly sophisticated.

However, according to Chris Stokel-Walker, author of TikTok Boom, how those numerous inputs are weighted and the exact criteria that determine which videos appear in your stream are opaque. "There is no recipe for it, no secret formula, as one TikTok employee in charge of attempting to track what goes viral and why told me in my book. Even the employee stated, "I don't think the algorithm team has an answer to that." It simply seems so smart.

One significant advancement is that, in contrast to past recommendation algorithms, TikTok doesn't merely wait for a user to give a video a thumbs up or content itself by evaluating the content a user selects to watch. Instead, it seems to be actively testing its own hypotheses by experimenting by displaying films that it believes would be entertaining and monitoring the reaction. According to Stokel-Walker, "it pushes the limits of your interests and tracks how you interact with those new films it seeds in your For You Page." "It might show you some videos about supercars if it believes you enjoy videos about Formula One."

According to Sascha Morgan-Evans, head of the TikTok studio at creative agency OK COOL, this experimentation is crucial to what the site offers creators and enables the service to quickly identify the general outlines of a particular viewer's interests. Each video that is uploaded to TikTok is shown to at least one user on the For You page. Based on how views add up, we've deduced that TikTok distributes each individual video to groups of users. With each successful round, one where the majority of users in a batch had a high number of favorable interactions with the video, the number of users in these batches grows.

This implies that there is a chance for worldwide renown for every user. You can immediately reach thousands or even millions of viewers even if you have no followers at all because your video will eventually appear on someone's For You Page if they are assessed to have interacted favorably. The pace of the movies also enables TikTok to quickly refine its data: Think about how many videos you watch on YouTube in an hour and the data that generates about you against how many you can watch on TikTok, advises Stokel-Walker.

But since the FYP isn't a magic trick, its failures can teach us just as much as its successes. The software is compulsive about collecting personal data, pleading for access to the contacts list, and tracking every uploaded video, as new users of the app will quickly discover. If you don't provide it those data points, it will have to personalize the feed as little as possible based on the wide geolocation and device information it has access to.

However, Stokel-Walker claims that when the algorithm is effective, TikTok appears to be nearly completely overpowered by its strength. Even messages telling users to put their phones down have been snuck in to those it believes are overly obsessed.

In one such message, TikTok star Gabe Erwin begs the viewer to "Go get some extra sleep, turn your phone off, do yourself that favor and have a lovely night" from the company's TikTokTips account to viewers reading through their feed for hours on end late at night. In an effort to curtail the most compulsive usage of its app, the business has also implemented additional "screen time" options, particularly for younger users. These features include turning off alerts after bedtime and enabling users to set a maximum time on the app each day.