🧠

Understanding the Algorithm

How YouTube Recommendation Works

The YouTube algorithm's goal is to "show the most satisfying videos to viewers." It operates with different weights across 4 surfaces: Home feed, Suggested (next video), Search, and Shorts feed.

Step-by-Step

1

Impression: Thumbnail displayed on viewer screen

2

CTR: Percentage of people who clicked after seeing thumbnail (4-10% average)

3

Watch Time: How long viewers watched (target 40%+)

4

Satisfaction signals: Likes, comments, shares, "not interested" ratio

Pros

  • Enables data-driven decisions
  • Identify growth bottlenecks

Cons

  • Algorithm updates frequently
  • Over-optimization reduces creativity

Use Cases

Analyzing and improving per-video performance Optimizing upload timing Content strategy planning

References