1,284 long-form YouTube channels. 121,294 hours of expertise in 24 months. Of the 1,206 whose status we could measure, only 10 (0.83%) publish that work as text on a website they own. The Dark Archive shows where the rest of that knowledge goes, whether machines can read it, and what those 10 do differently.
$24.50 List $49
Founding price for the first 10 customers.
Who is it for?
- Creators deciding whether a website is worth building, and what to put on it first.
- Agencies and channel managers who pitch search, AI visibility or a site build: the numbers and the method to put in the deck.
- Anyone building for creators who needs a measured baseline instead of a guess.
What did the study find?
- Most channels own nothing to build on. 678 of the 1,284 channels (52.80%) have no website at all.
- The sites that exist are hard for machines to read. Of 528 owned sites we scored for machine readability, 0 scored above 70 out of 100. The median was 20.
- Crawlers are not the problem. 476 of those 528 sites (90.15%) let every AI crawler in. The gap is in what the sites publish.
- The unwritten archive is huge. In 24 months the channels published 121,294 hours of long-form video. The median channel: 72 videos, 48.8 hours.
What is inside the full report?
- The finding, with every denominator and the channels we could not measure
- Method: how each channel was tiered, scored and checked
- The visibility ladder, from Dark to Sovereign, with the number of channels on each rung
- Tier distribution by niche: which genres keep the least of their work in writing
- The machine-readability gap: the score’s components, crawler policies and llms.txt files as we found them
- The citation test: whether an AI answer engine cited channels on their own topics, and the results that surprised us
- Why the gap exists
- What the channels at the top do differently: anonymised case studies
- What the gap costs a creator
- Limitations: what this sample can and cannot show
- Method appendix: definitions and decision rules
Every figure carries its evidence tag: measured from the data, sourced from a saved outside page, or inferred, with what would prove the inference wrong. Quote any number in a pitch or a post and know exactly what stands behind it.
Built to be quoted: how we chose the channels
We evaluated 14,713 candidate channels across 8 niches and kept only the 1,284 that met all six criteria below (8.73%, about 1 in 11). The other 13,429 were excluded, each with its reason on record. Every number in the report is computed from this finished sample.
| Criterion | A channel qualifies when | Excluded |
|---|---|---|
| Subscriber band | 20,000 to 300,000 subscribers | 12,074 |
| Language | English is the primary language | 153 |
| Long-form | the median of the last 10 non-Short uploads runs at least 15 minutes | 445 |
| Active | at least 12 uploads in the last 365 days, and at least 1 in the last 60 | 535 |
| Script-heavy | fewer than half of the last 10 uploads are livestreams; music, gameplay, compilation and reaction channels had no separate check, the topic-based search was the only screen | 192 |
| Independent | not owned by a media company, a news outlet or a brand content team | 30 |
The eight niches:
- Video essay and cultural analysis
- History and documentary
- Science and engineering explainers
- True crime and investigation
- Business and case-study breakdowns
- Interview and conversation podcasts
- Technology deep dives
- Personal finance and economics
Questions before you buy
How is this different from the free notes?
The notes quote a handful of headline numbers. The report has every finding by tier and by niche, the citation test, the case studies and the full method.
Can I get a refund?
It is a digital download, so there is no refund once you have downloaded it. Until then, our refund policy applies.
I want my own channel’s numbers, not the whole study.
Start with the free Dark Matter Check, or get the full diagnosis of your channel with Report Pro.
Get the report
$24.50 List $49
Founding price for the first 10 customers. Instant PDF download right after payment, on the confirmation page and in your receipt email. No refund after download.