3 observations · 10 transcripts analyzed · Explore the evidence before applying a pattern.
OBSERVATION 1
High-view videos have lower engagement rates
In the full sample, the videos with the highest views (e.g., 7665823355194363167, 50.8 million views, engagement rate 0.27%; 7654341587425692942, 32.1 million views, engagement rate 0.15%; 7655732127413112095, 33.3 million views, engagement rate 0.25%) have much lower engagement rates than videos with lower views in the sample (e.g., 7680996287877008670, 220,000 views, engagement rate 22.6%). This may be because high-view videos attract a large number of non-target audiences or passive viewers, resulting in relatively lower proportions of likes, comments, and shares.
Read evidence & explanation
In the full sample, the videos with the highest views (e.g., 7665823355194363167, 50.8 million views, engagement rate 0.27%; 7654341587425692942, 32.1 million views, engagement rate 0.15%; 7655732127413112095, 33.3 million views, engagement rate 0.25%) have much lower engagement rates than videos with lower views in the sample (e.g., 7680996287877008670, 220,000 views, engagement rate 22.6%). This may be because high-view videos attract a large number of non-target audiences or passive viewers, resulting in relatively lower proportions of likes, comments, and shares.
OBSERVATION 2
Video duration is associated with views
Among the videos analyzed in the transcripts, longer videos (e.g., 7665760124878802207, duration 2111 seconds, 15.9 million views; 7655732127413112095, duration 2461 seconds, 33.3 million views) achieved high views, while shorter videos (e.g., 7654716883307990302, duration 32 seconds, 1.3 million views) had relatively lower views. This may indicate that long videos (such as interviews) can provide in-depth content, attracting viewers to watch fully and promoting dissemination.
Read evidence & explanation
Among the videos analyzed in the transcripts, longer videos (e.g., 7665760124878802207, duration 2111 seconds, 15.9 million views; 7655732127413112095, duration 2461 seconds, 33.3 million views) achieved high views, while shorter videos (e.g., 7654716883307990302, duration 32 seconds, 1.3 million views) had relatively lower views. This may indicate that long videos (such as interviews) can provide in-depth content, attracting viewers to watch fully and promoting dissemination.
OBSERVATION 3
Content type affects dissemination effectiveness
In the videos analyzed in the transcripts, content types are diverse, including Dream Team draft (7670293981833612574, 1.4 million views), personal stories (7657262375246007583, 23.7 million views), etc. Personal story videos (e.g., 7657262375246007583) achieved high views, while interactive content (e.g., 7670293981833612574) had lower views but higher engagement rates (4.8%). This indicates that different types of content may attract audiences of different sizes, and personal stories may be more likely to resonate widely.
Read evidence & explanation
In the videos analyzed in the transcripts, content types are diverse, including Dream Team draft (7670293981833612574, 1.4 million views), personal stories (7657262375246007583, 23.7 million views), etc. Personal story videos (e.g., 7657262375246007583) achieved high views, while interactive content (e.g., 7670293981833612574) had lower views but higher engagement rates (4.8%). This indicates that different types of content may attract audiences of different sizes, and personal stories may be more likely to resonate widely.
Full overview & research limitations
This report is based on a complete metric sample of 50 videos, of which 10 videos underwent transcript analysis (Top 10 selection bias). The view counts of videos in the sample range from 11,300 to 50.8 million, and the publication time spans about 2 months. The analysis found that high-view videos typically have lower engagement rates (sum of likes, comments, and shares divided by views), while low-view videos may have higher engagement rates, reflecting the cumulative effect of views. By comparing successful videos (high views) with typical videos (medium views), we identified several patterns, including the negative correlation between engagement rate and views, the relationship between video duration and views, and the impact of content type (such as interviews, comment singing, personal stories) on dissemination. However, since transcript analysis only covers the Top 10 videos (sorted by views), and views accumulate over time, these patterns are only observations and do not constitute causal proof. Experimental suggestions should be interpreted with caution.
This analysis is based on a metric sample of 50 videos, of which only 10 videos underwent transcript analysis, and these 10 videos are the Top 10 with the highest views, introducing selection bias. Transcript analysis only covers high-view videos and may not represent the content characteristics of all videos. Additionally, views accumulate over time, and videos have different publication ages, making view comparisons unfair; the negative correlation between engagement rate and views may reflect a cumulative effect rather than content quality. All patterns are observations and do not constitute causal proof. The visual and audio content of the videos was not analyzed; only text transcripts and metrics were used, so the visual appeal or audio quality of the videos cannot be inferred. Experimental suggestions are only testing directions and do not guarantee results.
From profile to creator intelligence
Compare performance with transcript evidence
Import recent public videos, rank their performance, and study hooks, calls to action, and recurring topics. Analyze the full sample to compare top performers with typical videos.
Turn patterns into experiments
Reports link back to source videos and explain coverage limitations. Views change over time; patterns suggest experiments, not proof of what caused growth.