Wrong API endpoint, not a block: how a YouTube scraping error cost three months

A development team spent three months believing YouTube blocked their datacenter IPs after a probe of the internal API returned zero comments, which they mistakenly interpreted as a block. The real cause was a flawed testing method: YouTube's /youtubei/v1/next endpoint never includes comments in its initial watch payload, as comments require a second request using a continuation token. The team's single-step probe consistently returned empty comment data across all IP tiers — direct, datacenter, and residential — making the false negative appear highly credible. An outdated internal note about IP-class blocking on a different endpoint had been incorrectly generalised to cover all YouTube API surfaces, suppressing further investigation. Once the correct two-step process was applied, all three IP tiers successfully returned comment data, revealing no block existed and exposing one of the largest unserved scraping demand clusters the team had encountered.