A step-by-step interactive guide to demystify how YouTube decides what videos to recommend to you
Start LearningYouTube starts by analyzing your watch history. It looks at:
This is why logging in improves recommendations - it allows YouTube to personalize suggestions based on your history.
The algorithm then finds videos that are similar to what you've watched, considering:
Video titles, descriptions, and metadata are analyzed for keywords.
"People who watched X also watched Y" patterns are identified.
YouTube uses natural language processing and deep learning to understand video content beyond just keywords.
For each potential recommendation, YouTube evaluates how users typically engage with the video:
The algorithm then personalizes recommendations based on your specific preferences:
Your current watch session influences what's recommended next.
Local trends and device type affect recommendations.
Age, gender, and other demographic data influence suggestions.
Subscriptions and notification settings impact what you see first.
The system combines all these factors to rank and display recommendations:
The glow effect indicates the video the algorithm predicts you're most likely to watch next based on all analyzed factors.
YouTube's algorithm evolves constantly. Subscribe to get notified about important updates and new strategies.
We respect your privacy. Unsubscribe at any time.