Algorithm YouTube

What is the YouTube algorithm?

The YouTube algorithm is the platform’s recommendation system — it decides which videos each viewer sees on the homepage, in the Up Next suggestions and in the Shorts feed. It compares a viewer’s habits with those of similar audiences and ranks the videos they are most likely to watch and enjoy.

Recommendations appear in 5 places across YouTube: the homepage, Up Next suggestions next to the video you are watching, the Shorts feed, destination pages such as Shopping, Music and Movies & TV, and personalized shelves on channel pages.

How the YouTube algorithm works

The system learns from more than 80 billion signals every day. The 8 primary signals are:

• Watch history — the videos you have watched help YouTube pick better recommendations.

• Search history — what you search on YouTube influences future recommendations.

• Channel subscriptions — subscriptions hint at other videos you may like.

• Likes — liked videos help predict interest in similar content.

• Dislikes — tell the system what to avoid recommending.

• “Not interested” feedback — flags videos not to suggest again.

• “Don’t recommend channel” feedback — signals that a channel’s content did not land.

• Satisfaction surveys — video ratings that help the system understand satisfaction, not just watch time.

For creators, these signals fall into 2 groups: viewer personalization (what a specific person enjoys) and content performance — whether viewers click your video when it is shown, how long they watch and whether they engage. That is exactly what CTR and watch time measure.

Each surface weighs signals differently: Up Next relies mainly on the video you are currently watching, while homepage recommendations are built primarily on your watch history.

Why it matters for creators

The algorithm has 2 goals: help each viewer find videos they want to watch and maximize long-term viewer satisfaction. It analyzes a viewer’s profile in real time, considering device, time of day and past habits.

YouTube’s own advice is to focus on the audience, not the mechanics: instead of asking “does the algorithm like my content?”, ask “does my audience like this?”. Recommendations follow what viewers watch and enjoy.

The system also evaluates every new video on its own fresh performance data rather than your channel’s past results. A failed experiment here and there will not hold the channel back, but frequently posting content that does not resonate can affect overall channel performance over time.

How to work with the algorithm

YouTube’s official guidance for creators comes down to a few practices:

• Be consistent and recognizable. Regular uploads and a consistent title and thumbnail style help viewers instantly identify your videos. If your thumbnails are not earning clicks, start with our guide to what counts as a good CTR on YouTube.

• Know your audience. The Audience tab in YouTube Analytics shows preferred formats and which channels and content your viewers also watch.

• Help viewers watch more. Content series, playlists, end screens and clear calls to action surface more of your videos and generate positive interaction signals.

• Choose quality over quantity. Consistent quality beats high upload frequency, and the algorithm does not penalize you for taking breaks.

• Don’t overthink publish time. Upload time is not known to affect long-term performance; it matters mainly for Premieres and live streams, where you want your audience online.