YouTube keyword search is broken: Why per-video analysis misses 70% of trends (bdccomenets.xyz)
by anon | permalink
24/ 99 viralityNiche, low traction
edit & rescore →
10 HN points · front-page probability 31%
p10 · 2p90 · 306
The model already found titles that score higher. Try one.
I run bdccomenets. We classify YouTube comments into a fixed taxonomy (content idea / audience question / criticism / praise) with weight derived from likes + replies, then cluster recurring requests across all uploads on a channel — not per-video. This post is the reasoning behind the design: why per-video paste tools miss patterns, why keyword search ("feature", "wish", "please") catches ~30% of real requests, and why a fixed taxonomy beats free-form tagging once a team grows past one person. Would love feedback on the taxonomy and edge cases I'm not seeing.
ForesynWanna keep in touch?
Built this solo over a weekend. Soft-launching before the HN post on Monday. If you scored a draft and the prediction either nailed it or whiffed, I want to know.
DM @crimeacs on Telegram — fastest way to reach me
Connect on LinkedIn — Artemii Novoselov
Edit & re-score