YouTube
From 2022 to 2025, I conducted two major experiments with YouTube's AI-powered video recommendation algorithm to grow five YouTube channels. First, I utilized a Search Engine Optimization (SEO) tool called TubeBuddy to identify trending keywords users searched for on YouTube. I incorporated these keywords into the video titles and metadata, including descriptions, tags, playlists, end cards, and English subtitles. Additionally, I created thumbnail images using generative AI tools, ChatGPT and Dall-E. I provided ChatGPT the .srt subtitle file, poster, and film synopsis as input data to write detailed text prompts for Dall-E, an image-generating AI. As a result, the YouTube channels grew from zero to 50,000+ subscribers over the course of three years. This page documents the data that effectively engaged with the YouTube algorithm.
YouTube Metrics:
-
Impressions: The number of times the YouTube algorithm displayed the video thumbnails to viewers.
Figure 1. "Impressions by: Channel," Graph. Stash Movies on YouTube Content Manager. Accessed on Jan 15, 2025. YouTube. https://www.youtube.com/channel/UCuE6xnCgaG0LvEGAbvn8MEg.
The red vertical line marks the point of time I facilitated the SEO tool and thumbnail improvements on October 3rd, 2023. From that point onward, the YouTube algorithm began to favor our videos more than before. The enhanced impressions led to a significant increase in Click-Through Rate (CTR).
-
Click-Through Rate (CTR): Percentage of viewers who clicked a video after seeing its thumbnail and title.
The graph below illustrates the exponential increase in CTR for the video "Mr. Earth" after the SEO and thumbnail updates on October 10th, 2023.


Figure 2. "Impressions Click-Through Rate (CTR) by: Content," Graph. Stash Movies on YouTube Content Manager. Accessed on Jan 15, 2025. YouTube.
https://www.youtube.com/channel/UCuE6xnCgaG0LvEGAbvn8MEg.

Original Thumbnail (Movie Poster):
