By Dario Cavallaro · Published · Updated
How the YouTube Algorithm Works
YouTube personalises recommendations across Home, Up Next and the Shorts feed. Search responds to a query and should be considered separately. The public guidance describes several signals; it does not give creators a universal score for every surface.
The surfaces, briefly
- Home. Recommendations primarily use the viewer’s watch history.
- Up Next. The video currently being watched is a main recommendation signal.
- Search. Consider the query and whether your title, description and video answer it.
- Shorts. A personalised feed for short-form video.
The signals that matter
YouTube describes watch and search history, subscriptions, reactions, feedback controls and satisfaction surveys. Different features use signals differently. For your own review, compare whether the title earns interest with whether the video serves the viewer. Retention and click-through are useful diagnostics, not a guaranteed reach formula.
What it means for how you make videos
Give the title and thumbnail an accurate promise, then deliver it in the video. A longer explanation may need chapters and examples; a Short may work better with one point. When repurposing, review each adaptation as its own piece, suited to its own surface. These are production choices to test, not guaranteed ranking rewards.
The honest version
Make the subject clear, earn the opening honestly and use your results to improve the next video. Keep audience fit, viewing experience and your actual goal together in the review. You cannot infer a hidden ranking weight from one upload.
Sources and scope
Reviewed on 4 October 2026. YouTube: how recommendations work.
Frequently asked questions
Does YouTube rank videos by views or by satisfaction?
YouTube uses several signals, including viewing history and satisfaction surveys, to personalise recommendations. It does not publish one score or a universal ranking of views, retention and satisfaction for every surface.
How is the YouTube Shorts algorithm different from long-form?
The Shorts feed is a personalised surface. Assess how viewers respond to your Shorts separately from long-form uploads; a metric that helps diagnose one format is not a universal ranking weight.
Do thumbnails and titles affect the YouTube algorithm?
Titles and thumbnails help viewers decide what to watch. Compare clicks with what happens after the opening, and keep the promise accurate. A high click-through rate alone does not guarantee recommendations.