# 0.83%: how many YouTube channels keep a written archive By Tai Nguyen, ExDarkMatter. Published 2026-10-04; updated 2026-10-04. Canonical: https://exdarkmatter.com/notes/how-many-channels-keep-a-written-archive/ **Short answer:** Very few. In The Dark Archive, our study of 1,284 English-language long-form YouTube channels, we could measure the website status of 1,206. Only 10 of them, 0.83%, systematically republish their videos as text on a website they own. For 99.2% of the channels we measured, the knowledge in their videos has no written home that the creator controls. ## What did we measure? The Dark Archive looked at English-language YouTube channels that publish long-form, script-heavy work: video essays, history, science explainers, interview podcasts, business breakdowns and similar knowledge content. For each channel we followed every link that leads out of YouTube and asked one narrow question: does this creator keep a written version of their videos on a website they control? We placed every channel on a ladder of five rungs, defined in the [glossary](https://exdarkmatter.com/glossary/). At the bottom is Dark: no link out of YouTube at all. Next is Satellite: only rented profiles, such as social accounts, a link-in-bio page or Patreon. Beacon owns a domain but has nothing on it to index. Dormant has a blog or episode pages that went stale, or that are not built from the videos. At the top is Sovereign: a site that keeps publishing, where most posts come from the videos and each post stands on its own as an article. The rules were deterministic and applied from the top down, so the same channel always lands on the same rung. A channel whose own domain did not respond got no tier at all. That is why the share below is out of 1,206 channels and not out of the full sample of 1,284. ## How many channels reached the top of the ladder? Ten. Of the 1,206 channels whose status we could measure, 10 met the Sovereign rule. That is 0.83%. Put the other way round, 99.2% of the channels we measured do not keep a systematic written archive of their videos on a site they own. That is not the same as saying these creators have no website. Many of them do. Some keep a landing page that sends visitors straight back to YouTube. Some rely on the episode list their podcast host generates. Some started a blog and stopped. What is rare is one specific habit: turning each video into a page that a person can read, a search engine can index and an answer engine can quote. ## Why does a written archive matter if the videos are already public? The videos are public, and AI answer engines do cite YouTube. In [an Ahrefs study of Google's AI Overviews](https://ahrefs.com/blog/most-cited-domains-ai-overviews/), youtube.com was the most cited domain of all. So the case for a written archive is not that machines cannot find a video. It is narrower than that, and we think more useful: - **Not every engine cites video.** In [a citation experiment run by OtterlyAI](https://otterly.ai/blog/geo-experiment-video-to-blog/), the same stories were published as videos and as articles, and several assistants cited only the written versions. A creator with no text is missing from those answers, even on their own topic. - **A page can be quoted one idea at a time.** A long episode holds dozens of answers. A video page rarely surfaces as the source for any single one of them. An article with a clear heading and a direct answer under it can. - **The creator owns it.** A page on your own domain keeps working when a platform changes its rules, its recommendations or its terms. None of this guarantees traffic. Being cited and being clicked are different things, and we make no claim here about revenue from AI citations. ## Why do so few creators do it? Our reading of the evidence is that most creators know a website is possible, and many have tried one. The reasons they give for stopping are time and return. Writing up every video is slow work, and for a channel paid by advertising and sponsorship, a site adds no obvious income. The middle of the ladder looks like that story: creators who bought a domain and never filled it, and creators who started a blog and let it go quiet. There is also a quieter reason. A transcript is not an article. Pasting the words of a video onto a page produces something long, unstructured and hard to read, so the result rarely feels worth publishing. Turning a video into a page that answers questions takes editing, headings and links, and that is exactly the step that falls away when time runs short. The report weighs these explanations against each other and shows which way the evidence leans. ## What should a creator take from this? If you publish long-form work, the bar for standing out in text is low. Almost nobody in your niche is doing it systematically, so even a modest archive that grows every week puts you in a small group. A few practical points follow from the finding: - **Start with the videos people already search for.** Your back catalogue holds answers to questions that people ask every day. - **Write each page around one question.** Answer it in the first paragraph, then expand. That is the shape answer engines lift. - **Publish on a domain you own.** Link each page back to the moment in the video it came from, so readers and machines can check the source. - **Keep going.** A stale blog is the most common way a creator site ends up on the Dormant rung. ## Where are the full numbers? This note covers one finding. The full breakdown by rung and by niche, the method behind each rule and the anonymised case studies of the channels at the top are in [The Dark Archive report](https://exdarkmatter.com/report/). For a related finding about access, read [what blocking AI crawlers costs a creator's site](https://exdarkmatter.com/notes/what-blocking-ai-crawlers-costs/). To see where your own channel sits on the ladder, [run the free Dark Matter Check](https://exdarkmatter.com/check/). ## Sources - [Ahrefs: the most cited domains in Google's AI Overviews](https://ahrefs.com/blog/most-cited-domains-ai-overviews/), retrieved 2026-09-30 - [OtterlyAI: video-to-blog citation experiment](https://otterly.ai/blog/geo-experiment-video-to-blog/), retrieved 2026-09-30