If you’ve ever used TikTok then you’ll know that the For You Page, the company’s recommendation engine that feeds you new content, is arguably the most powerful thing about it. The problem is that no-one really knows how videos are selected for the For You Page, or how the For You Page recommends content to you when you’re new to the platform.
The basics about recommendation systems
From shopping to streaming to search engines, recommendation systems are all around us and according to various companies, they’re designed to help us have a more personalised experience.
Usually, these systems suggest content after taking into account your content preferences as expressed through interactions with the app, like posting a comment or following an account. These help the recommendation system gauge the content you like as well as the content you’d prefer to skip.
What factors contribute to For You?
On TikTok, the For You feed reflects preferences unique to each user. The system recommends content by ranking videos based on a combination of factors – starting from interests you express as a new user and adjusting for things you indicate you’re not interested in, too – to form your personalised For You feed.
Recommendations are based on a number of factors, including things like:
- User interactions such as the videos you like or share, accounts you follow, comments you post, and content you create.
- Video information, which might include details like captions, sounds, and hashtags.
- Device and account settings like your language preference, country setting, and device type. These factors are included to make sure the system is optimised for performance, but they receive lower weight in the recommendation system relative to other data points measured since users don’t actively express these as preferences.
TikTok says that all of these factors are processed by their recommendation system and weighted based on their value to you. A strong indicator of interest, such as whether you finish watching a longer video from beginning to end, would receive greater weight than a weak indicator, such as whether the video’s viewer and creator are both in the same country. Videos are then ranked to determine the likelihood of your interest in a piece of content, and delivered to each unique For You feed.
While a video is likely to receive more views if posted by an account that has more followers, by virtue of that account having built up a larger follower base, neither follower count nor whether the account has had previous high-performing videos are direct factors in the recommendation system.
Getting started
How can you possibly know what you like on TikTok when you’ve only just started on the app? To help kick things off TikTok invites new users to select categories of interest, like pets or travel, to help tailor recommendations to their preferences. This allows the app to develop an initial feed, and it will start to polish recommendations based on your interactions with an early set of videos.
For users who don’t select categories, the app starts by offering you a generalised feed of popular videos to get the ball rolling. Your first set of likes, comments, and replays will initiate an early round of recommendations as the system begins to learn more about your content tastes.
Finding more of what you’re interested in
Every new interaction helps the system learn about your interests and suggest content – so the best way to curate your For You feed is to use the app. Over time, your For You feed should increasingly be able to surface recommendations that are relevant to your interests.
Your For You feed isn’t only shaped by your engagement through the feed itself. When you decide to follow new accounts, for example, that action will help refine your recommendations too, as will exploring hashtags, sounds, effects, and trending topics on the Discover tab. All of these are ways to tailor your experience and invite new categories of content into your feed.
Seeing less of what you’re not interested in
TikTok is home to creators with many different interests and perspectives, and sometimes you may come across a video that isn’t quite to your taste. Just like you can long-press to add a video to your favourites, you can simply long-press on a video and tap “Not Interested” to indicate that you don’t care for a particular video. You can also choose to hide videos from a given creator or made with a certain sound or report a video that seems out of line with our guidelines. All these actions contribute to future recommendations in your For You feed.
Addressing the challenges of recommendation engines
One of the inherent challenges with recommendation engines is that they can inadvertently limit your experience – what is sometimes referred to as a “filter bubble.” By optimising for personalisation and relevance, there is a risk of presenting an increasingly homogenous stream of videos. TikTok says that this is a concern they take seriously as they maintain their recommendation system.
Interrupting repetitive patterns
To keep your For You feed interesting and varied, ThikTok’s recommendation system works to intersperse diverse types of content along with those you already know you love. For example, your For You feed generally won’t show two videos in a row made with the same sound or by the same creator. TikTok also doesn’t recommend duplicated content, content you’ve already seen before, or any content that’s considered spam. However, you might be recommended a video that’s been well received by other users who share similar interests.
Diversifying recommendations
Diversity is essential to maintaining a thriving global community, and it brings the many corners of TikTok closer together. To that end, sometimes you may come across a video in your feed that doesn’t appear to be relevant to your expressed interests or have amassed a huge number of likes. TikTok says that this is an important and intentional component of their approach to recommendation: bringing a diversity of videos into your For You feed gives you additional opportunities to stumble upon new content categories, discover new creators, and experience new perspectives and ideas as you scroll through your feed.
By offering different videos from time to time, the system is also able to get a better sense of what’s popular among a wider range of audiences to help provide other TikTok users a great experience, too. The company says that their goal is to find a balance between suggesting content that’s relevant to you while also helping you find content and creators that encourage you to explore experiences you might not otherwise see.
Safeguarding the viewing experience
TikTok says that their recommendation system is designed with safety as a consideration. Reviewed content found to depict things like graphic medical procedures or legal consumption of regulated goods, for example – which may be shocking if surfaced as a recommended video to a general audience that hasn’t opted in to such content – may not be eligible for recommendation. Similarly, videos that have just been uploaded or are under review, and spam content such as videos seeking to artificially increase traffic, also may be ineligible for recommendation into anyone’s For You feed.
Improving For You
“Developing and maintaining TikTok’s recommendation system is a continuous process as we work to refine accuracy, adjust models, and reassess the factors and weights that contribute to recommendations based on feedback from users, research, and data. We are committed to further research and investment as we work to build in even more protections against the engagement bias that can affect any recommendation system” says the company.
This work spans many teams – including product, safety, and security – whose work helps improve the relevance of the recommendation system and its accuracy in suggesting content and categories you’re more likely to enjoy.
Ultimately, your For You feed 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.
