Mining expert YouTube channels for ranking and selling playbooks

What you're doing and why

Before you invent a ranking or selling tactic, check whether someone with a bigger portfolio has already measured it. Three of the biggest rank-and-rent / local-SEO YouTube channels publish their entire method for free — including their dollar figures and their failures. We scraped all three channels (every video, every auto-caption), turned them into clean searchable transcripts, and distilled each into a playbook doc so the business runs on tested tactics, not guesses. Crucially, we cross-checked the three playbooks against each other: where two experts agree we adopt with confidence; where they contradict, we flag it and let our own formula doc decide.

The exact steps

List every video on a channel without downloading video files (this needs the yt-dlp tool and runs in the terminal — ask your agent to do it): yt-dlp --flat-playlist --print "%(id)s|%(title)s|%(duration)s" https://www.youtube.com/@CHANNEL
Download just the auto-captions (small text files, no video): ask your agent to run a resumable caption-download script over the list — ours is ippei-youtube/dl-parallel-generic.sh. Expect ~99% of videos to have captions; the missing few are music/shorts.
Clean the caption files into plain readable transcripts (strip timestamps and duplicated lines) — ours is a small Python script per channel folder.
Sort videos into topic batches (rank-and-rent, links, sales, niche research…) by title keywords, then read the important ones and note every tactic with the video id as the receipt.
Write the distilled playbook doc, and give every tactic one of three labels: agrees with what we already do (confidence boost), new (candidate to test on one site), or contradicts our rules (flag for the owner — never adopt silently).

What it looked like for us

@buildinpublic (The Edward Show): 1,250 videos → 1,248 transcripts, ~4.5M words. The gold is the guests: Darren Shaw's 2026 measured ranking factors (review recency and open hours now matter; NAP consistency and GBP posts measurably don't), Joy Hawkins (December 2025 core update wiped small-business sites but never touched the local pack), and Caleb Ulku (texting the review ask from the number the customer knows gets a 40–45% response vs ~5% for software).
@indexsy (Jackie Chou): 799 videos → 797 transcripts, ~2.1M words, 99 of them rank-and-rent. Exact receipts: a 350-citation starter pack costs ~$60; a parasite post on Outlook India ~$349; a European operator runs 500 sites at ~€50/site/month (€26k/mo); the cold email that closes is five sentences asking "do you have capacity for an extra 20–25 clients a month?" — with no website link and no mention of how the leads are generated.
Both playbooks: docs/handover/2026-07-28-buildinpublic-youtube-playbook.md and docs/handover/2026-07-28-indexsy-youtube-playbook.md, each with a contradiction ledger.

Gotchas

Experts contradict each other constantly. Indexsy blasts 300+ citations; Joy Hawkins says ~10 core ones are all that matter. Indexsy swears by exact-match domains; Darren Shaw calls them "mostly useless". Never adopt a tactic because ONE famous person said it — check it against your own rulebook (ours is docs/formula.md) and flag conflicts for the owner.
Some published tactics are illegal or account-ending (bought Google Business Profile verifications, paid review surges, CTR bots). The channels openly say "people have been sued for this". Write them down as rejected, with the reason, so the next person doesn't rediscover them as if they were new ideas.
Daily-vlog channels bury the tactics. Most Indexsy episodes are 60% revenue-report ritual before the tactic starts. Skim with keyword searches first; only deep-read the episodes whose titles promise a method.
Auto-captions mangle names and numbers. Any dollar figure or tool name that ends up in a playbook was re-checked against the raw transcript before we trusted it.
The most valuable tape is often the contrarian one. The "rank-and-rent is dead" episode (60–70% of clicks go to the map pack, ~25% to ads) did more to change our roadmap priorities than any how-to video.