TikTok reveals how its video recommendation feed works

Chinese social media platform TikTok has revealed how its video recommendation feed works in a bid to be more transparent with users.

The popular app, which lets users to make short lip-syncing clips to share with their followers, has detailed the workings of the For You feed, one of its ‘defining features’.

For You is powered by an algorithm that weighs up a number of metrics, including interactions and account settings, to deliver a stream of videos unique to each user.

TikTok, which is owned by Chinese company Bytedance, acknowledged uncertainty around how recommendations are delivered to a user’s feed.

Western fears have emerged that the app’s owner could be sharing users’ data with the Chinese government, which could potentially force the app to share data under Chinese law. 

US Senate and US military staff have already been banned from using the app, while Reddit co-founder Steve Huffman previously called TikTok ‘fundamentally parasitic’ and ‘spyware on your phone’. 

As part of efforts to build up trust and allay concerns about its alleged links with the Chinese government, it’s published a detailed description of the video recommendation process.

It follows a report in The Wall Street Journal this week that some TikTok users were praising China’s leader Xi Jinping to see if it would boost their posts.  

‘TikTok’s mission is to inspire creativity and bring joy – we’re building a global community where you can create and share authentically, discover the world, and connect with others,’ the platform said in a blog post.

‘The For You feed is one of the defining features of the TikTok platform, but we know there are questions about how recommendations are delivered to your feed.

‘For You is powered by your feedback – the system is designed to continuously improve, correct, and learn from your own engagement with the platform to produce personalised recommendations that we hope inspire creativity and bring joy.’

When TikTok users open the app, they are presented with the For You feed – a stream of videos recommended to users based on their interests.

Key to the feature is a recommendation system just like the ones used for search engines, streaming services and other social media platforms to help people have a more personalised experience.

‘In general, these systems suggest content after taking into account user preferences as expressed through interactions with the app, like posting a comment or following an account,’ said TikTok in the post.

‘These signals help the recommendation system gauge the content you like as well as the content you’d prefer to skip.’

The system takes into account user input – the videos a user likes or shares, accounts they follow, comments they post and content they create – much like Facebook or Instagram. No surprises yet. 

But the algorithm also considers device and account settings like your language preference, country setting, and device type, as well as video information such as captions, sounds and hashtags associated with the content a user likes.

All these factors are processed by the recommendation system and weighted based on their value to each individual TikTok user.

The algorithm also uses a number of subtle ‘signals’ from how a user interacted with a video to decide whether to drop a video of a similar nature into the feed.

These are separated into ‘strong’ signals and ‘weak’ signals.

For example, a strong indicator of interest would be a user watching a longer video from beginning to end. 

This would receive greater weight than a weak indicator, such as whether both the creator and viewer of the video are in the same country.

Videos are then ranked to determine the likelihood of a user’s interest in a piece of content, and delivered to each unique For You feed.

TikTok said neither follower count nor whether the account has had previous high-performing videos are direct factors in the recommendation system.

In other words, the algorithm doesn’t prioritise videos from accounts with higher followers when deciding what video to drop into each user’s feed.

When someone opens up a TikTok account, the platform invites the new user to select categories of interest, like pets or travel, to help deliver appropriate content.

Their first set of likes, comments and replays will ‘initiate an early round of recommendations’ as the machine learning system learns more about their tastes.

And just like any machine learning system, the more a user interacts, the more the algorithm picks up tips on what to recommended.

To keep For You interesting and varied, the recommendation system also adds diverse types of content that a user may not have expressed an interest in.

Users can tap ‘Not Interested’ to indicate that they don’t care for a particular video.

Videos that have just been uploaded or are under review, such as those showing ‘excessively gruesome or shocking’ content, and spam videos seeking to artificially increase traffic, also may be ineligible for recommendation into anyone’s For You feed.

The platform, which allows users as young as 13 to create an account, has previously come under fire for circulating explicit content. 

At the end of April, it introduced a minimum age requirement of 16 to use its direct messaging feature, to help ensure families can ‘trust the service’ during coronavirus lockdown.