Facebook will rank content in 2023 – Feeds, Stories, Reels, and more

Facebook has updated the resource that explains how it ranks its content on its website including Feeds, Stories, and Reels.

Meta announced a number of new features and tools to improve transparency on its social network. This will help marketers.

Why we care. Advertisers and marketers need to understand how Facebook ranks their content to make the best decisions for their campaigns.

What is new? Meta has published several new features to help clarify its ranking factors through its newsroom.

• System Card: Facebook created 14 system cards for marketers to understand how Facebook uses AI in order to rank content and create feeds tailored to individual users. These cards explain how users can customize what they see.
• “Why Am I Seeing This?”: In the next few weeks, Meta will expand this feature to Facebook Reels. This feature allows users to see how previous site activity has affected the content that AI currently considers relevant for them.
• “Show More, Show Less”: Facebook is planning to make the feature, currently available in Feeds, Videos and Reels through the three-dot menu, more prominent.
• Meta’s Content LibraryI: Facebook is planning to launch a new set of tools called Meta’s Content Library API and in the coming weeks. The new library will include data from posts, pages and groups on the social network site.

System Cards

Facebook’s resource center has been updated with the addition of system cards. This system is made up of 14 cards.

• FeedFacebook calculates relevance scores for 500 posts, and ranks them in order of decreasing relevance. The feed is designed to display a variety in content, so a user should not see more than one video post in a single row.
• Feed Ranked Comments:AI has ranked comments according to what it believes will be the most relevant for each user. This is done by looking at factors like how popular comments are, and if they were published by someone within their network.
• Feed recommendations:AI determines what content is most likely to be engaging by analyzing factors like groups that the user has recently joined, and posts that they have liked. This information is then used to determine what content to recommend (e.g. posts, reels or live videos).
• ReelsAI determines which reels to serve and in what order, based on what the user is likely to be most interested in. These predictions are made by looking at factors like accounts that the user has liked, followed or engaged with recently.
• Stories The AI system will automatically show Stories from pages or people based on what it predicts the user is most likely to find interesting. The system applies rules that ensure users receive a mix of stories with varying content.
• People you may know: AI tries to identify people of interest based on factors like people in the same group as the user or those who are friends.
• Video When users interact with Facebook Video and view it, one of its AI systems will deliver a variety of video types to match their preferences. The Video tab contains this content. This content could include music, games, reels or shows. These are creators who may have content that users find interesting.
• Marketplace : One of the AI systems that underlies Facebook’s Marketplace feed recommends listings relevant to the user when they interact with Facebook. Users can, for example, see listings in categories like pet supplies, sporting goods and home goods. The feeds of users may also contain other recommendations such as content and sellers that they might be interested in.
• NotificationsAI selects which notifications to send, and then ranks them in order of relevance to the user. In the meantime, notifications that have been viewed previously are displayed in order of when they were viewed.
• Search:AI gives each search result, based on factors like content type, a score that reflects how relevant it is for the user. The results will be ranked by relevance based on the scoring.
• Groups Feed :AI determines automatically which posts will appear in the feed for Groups, and in what sequence, by scoring content based on relevance.
• Individual group feed:AI predicts which content users will engage with most and ranks it according relevance in their individual groups feed. Relevancy factors are based on what users recently liked, followed or engaged with.
• Suggested group: Facebook AI uses factors like groups in which a user’s friends may be members and topics that are related to products the user has recently interacted with to identify groups of interest.
• Pages you may like: AI suggests pages to follow, based on the pages that a user’s friends recently liked or those pages which might be related to posts and products the user recently interacted with.

Meta’s Content Library API

Meta’s new API and Content Library is another major update to Facebook’s Resource center. The database will include data from:

• Public Posts
• Pages
• Groups
• The following are some examples of the events that you can attend

The library will allow users to search, filter and explore on a graphical interface or via a programmatic API.

This tool is only available to researchers who are qualified and affiliated with academic or research institutions, and who are pursuing research in areas of public interest, scientific research, or both. Researchers will have to apply to access this data.

Personalizing user experience

Facebook has confirmed that it will not only provide greater transparency about its ranking factors but also give users tools to regain control over the content they view – such as the “Why Am I Seeing This?” feature. feature.

Facebook’s tools allow users to customize their experience and decide what they want to see. Facebook users can change their preferences by going to Settings and then clicking on Feed Preferences.

What has Facebook said? Nick Clegg shared on Meta’s digital newsroom how AI ranks content and will make it easier for users in the future to control their experience. He said:

• “[Our AI] systems increase the likelihood that you will see posts relevant to and interesting to yourself. We are also testing new controls, making them more accessible and making it clearer on how to better control the content you see in our apps. We’re also providing more information to experts, so they can better analyze our systems.
• Our AI systems can predict the value of a piece content to you so that we can present it sooner. Our systems can predict that you’ll share a particular post, for example. This is an indication that you find the post interesting.
• As you can imagine, there is no perfect way to determine whether or not a particular post will be valuable for you. We use a variety of predictions to help us get the best content. Some are based on behaviors, and others on feedback from users.
• We hope that by introducing our products to researchers at an early stage of the development process, they will provide constructive feedback so we can build the best tools possible to meet their requirements.

Deep diveYou will find a detailed explanation on the blog of the AI that is behind the content recommendations. Visit Meta’s Transparency Center for more information about how AI makes predictions using signals.

The post Search Engine Land : How Facebook will rank content in 2023 – Feeds, Stories, Reels, and more first appeared on Search Engineland .