Facebook suggests videos to users for several reasons. The main goals behind video recommendations are to keep users engaged on the platform, serve relevant content, and maximize ad revenue.
User Engagement
One of the primary reasons Facebook suggests videos is to keep users engaged on their platform for longer periods of time. The longer users stay on Facebook, the more ads they see and the more data Facebook can collect on them. Suggesting engaging videos based on users’ interests and behavior patterns helps keep their attention.
Research has shown that people spend more time on platforms when viewing video content compared to text or images alone. A study by Socialbakers found users spend over 16 minutes on average watching video on Facebook, compared to less than 6 minutes spent otherwise. Facebook wants to capitalize on this by populating users’ feeds with plenty of suggested videos to capture their sustained attention.
Videos auto-play in feeds once scrolled into view, immediately grabbing the user’s focus. The eye-catching motion of video gives Facebook an advantage over other content in garnering user engagement. Scrolling through text-heavy posts feels laborious in comparison to effortlessly watching a series of videos.
Facebook is able to monitor users’ engagement with suggested videos based on metrics like completion rates and watch times. The platform’s algorithms learn an individual’s preferences to recommend the types of videos most likely to resonate and thoroughly engage each unique user.
Relevant Content
Facebook also suggests videos in an attempt to serve users content relevant to their tastes and interests. The platform wants to provide each user with an experience personalized to them.
Their algorithms analyze pages and accounts a user follows, groups they are members of, past viewing behavior, and more. Based on these signals, it recommends videos likely to align with the individual’s demonstrated preferences.
For example, if a user frequently interacts with football-related pages and posts, the algorithm will pick up on this interest and suggest more football video content. It learns as it goes based on what content garners engagement from the user.
Fb also allows users to explicitly provide inputs that inform video suggestions. Features like Watch Lists let users indicate TV shows and sports teams they enjoy following. Users can heart or react to videos they find interesting. All these actions teach the algorithm what video content resonates with that particular user.
Maximize Ad Revenue
At its core, Facebook is an advertising platform. Its primary goal is to make money by serving ads. Videos allow more advertising opportunities compared to plain text or image posts.
Videos on Facebook can have multiple ads inserted at the beginning, middle, or end. The longer a video’s runtime, the more ads it can accommodate. Facebook Watch shows averagely have 3.5 ads inserted in a 30-minute episode. More suggested videos for users to watch equals increased ad inventory and revenue.
Videos also lend themselves to new interactive ad formats. For example, ads can invite viewers to engage further by installing a game app or taking a brand survey. The immersive experience of video makes users more prone to interact with these ad units. Higher engagement directly translates to more money made.
Further, videos provide a branding opportunity that text-based posts don’t allow. Advertisers have creative control over their video ads, customizing sights, sounds, and messaging to promote their brand. Associated metrics like completion rates give insight into how effectively the video captured audience attention.
In summary, serving users more suggested videos allows Facebook to:
- Capture longer and more focused attention spans
- Provide personalized content catered to individuals
- Insert multiple ads and new interactive ad formats
- Offer branding potential not afforded by static posts
All of these factors ultimately maximize ad revenue.
How are Video Suggestions Determined?
Facebook uses complex machine learning algorithms to determine which videos to suggest each user. Many signals get taken into account during this process.
Key Factors In Video Suggestions
- Pages and accounts followed – Videos from pages and accounts a user follows are prioritized
- Groups joined – Relevant videos from groups a user is a member of are suggested
- Watch history – Videos similar to those a user has previously watched will be recommended
- Friends’ activity – Videos friends have recently watched or engaged with may be suggested
- Location – Videos pertaining to a user’s city or country may get preference
- Keywords – Scanning text associated with a video to match relevant keywords to the user’s interests
- Offline activity – Factor in activities outside of Facebook that indicate interests
These signals combine to curate an optimized, personalized feed of suggested videos for each individual user.
Optimization Process
Facebook uses an iterative process to refine and optimize how video suggestions are determined over time:
- Predict – Make predictions about which videos will best engage each user based on available signals
- Display – Serve the predicted optimal video suggestions to the user in their feed
- Measure – Track how the user interacts with the suggested videos by monitoring engagement metrics
- Refine – Improve the prediction model based on measured outcomes to increase relevance
- Repeat – Continuously iterate through this process to get smarter over time
With each iteration, Facebook’s algorithms become more adept at suggesting videos individual users will enjoy watching and engaging with.
Controlling What Videos Are Suggested
Users do have some control over what videos get suggested to them on Facebook:
- They can indicate topics of interest through Watch Lists
- Adjust ad topic preferences in Ad Settings
- Follow Pages and Groups that share content they care about
- Actively like/react to videos they find interesting
- Click “Not Interested” on suggested videos that miss the mark
However, users have limited influence relative to Facebook’s algorithms that work behind the scenes. Many users feel recommended videos are still not relevant despite their inputs. The platform prioritizes its own optimization goals above user preferences in some cases.
Controversy Around Facebook’s Video Suggestions
Facebook’s algorithms for determining video suggestions have come under scrutiny for a few reasons:
Maximizing Engagement Over User Well-Being
Critics argue Facebook prioritizes its engagement and revenue goals too strongly over users’ well-being. The algorithms work extremely well at determining our deepest interests and recommending increasingly specific content.
But for sensitive topics like politics, this can push users further into echo chambers and extremes. The quest to keep eyes glued to the screen supersedes concerns about polarizing or radicalizing users.
Promoting Misinformation
To maximize engagement, Facebook’s algorithms often promote controversial, clickbait, and hyper-partisan video content. Recommending these attention-grabbing videos above informative quality sources helps their metrics.
But this gives a platform to misinformation meant more to work people up than convey truth. Facebook has received particular criticism for enabling the spread of misinformation regarding elections and COVID-19.
Lack of Transparency
The inner workings of Facebook’s complex proprietary algorithms are highly opaque. The company reveals little publicly about how exactly they determine suggested videos.
Without transparency, it is impossible to audit the system for fairness and accuracy. Facebook has total control over curating reality for billions of users. This tremendous centralized power worries critics.
In summary, while Facebook’s algorithms allow an optimized video experience, concerns remain over user well-being, misinformation, and transparency.
The Future of Facebook’s Video Suggestions
Here are a few ways we may see Facebook’s approach to suggested videos evolve moving forward:
More Personalization
Facebook will likely keep improving its ability to hyper-personalize video suggestions for each user. As AI and machine learning progress, the algorithms will continuously get smarter.
They will further optimize to each user’s niche interests by incorporating more signals like biometrics and environmental context.
Different Recommendation Criteria
Facebook may tweak its algorithms to recommend videos based on criteria beyond engagement, like informational value or objectivity. This could help address criticism around echo chambers and misinformation.
More Transparency
There may be public pressure for Facebook to reveal more about how their algorithms determine suggested videos. This could help hold them accountable and ease concerns over centralized power.
New Ad Options
More interactive and immersive video ad formats are likely to emerge. For example, ads that allow viewers to instantly shop featured products or try on makeup virtually.
Facebook will find new ways to monetize video content to justify ongoing heavy investments in recommendation algorithms.
Conclusion
Facebook’s algorithms suggest videos with the dual goal of maximizing user engagement and ad revenue. They take various signals about each user as input to serve hyper-relevant video recommendations in feeds.
The results are an addictive and personalized video experience. But concerns remain around the impacts on user well-being and the spread of misinformation due to lack of transparency.
Moving forward, we can expect Facebook’s approach to suggested videos to involve more advanced personalization, some changes addressing criticism, increased transparency, and new interactive ad formats.