We checked 115 faceless YouTube channels — 47% had stopped posting

The clearest split was not the niche and not the subscriber count. It was how many videos the channel had ever published. Channels sitting under 15 videos were mostly dormant. Channels past that point were posting every few days.

The numbers

Median time since the last upload across the whole sample: 18 days.

The wall is somewhere around video 15

YouTube's public RSS feed lists a channel's most recent videos and stops at fifteen. That gave us a free way to split the sample: a channel whose feed is full has published at least fifteen videos, and a channel whose feed is short has published only that many, ever.

Three days. Not three weeks. Once a channel is past that first stretch, it is not posting occasionally — it is posting more or less continuously.

Two honest caveats. This is correlation, not a mechanism: a channel that was always going to survive naturally also accumulates more videos, so the count is partly an effect rather than a cause. And fifteen is an artefact of where the feed cuts off, not a magic number. The real finding is the gap between the two groups, not the exact threshold.

Why we think it has this shape

We run three faceless channels ourselves. Over 150 published videos in ten weeks, around 50,000 views on YouTube. The first ten came easily, because the ideas were the fun part and the backlog was full. What ran out was not ideas.

Making one narrated short by hand is an evening: find the story, check it, write it, record or generate the voice, source the visuals, cut it, caption it, write the title. Do that once and it is a hobby. Do it every night for three weeks and it is a second job you did not apply for. The channels in the dormant half did not stop because the niche was wrong. They stopped because the tenth evening in a row is where enthusiasm stops carrying it.

That is also why the dormant channels in our sample look so healthy from the outside. Several had genuinely good videos with tens of thousands of views and then nothing. One geography channel promised a new country every day, got 25,000 views on one short, and stopped eighteen months ago.

What we would tell someone starting one

Decide the rhythm before the niche. Pick something you can keep for sixty videos, not something that sounds good for six. Two a week you actually publish beats daily you abandon in March.

Batch the boring half. The part that kills channels is not writing, it is the assembly line after the writing. Whatever you can prepare in bulk — scripts, voices, thumbnails — prepare in bulk, so a bad evening costs you nothing.

Count videos, not subscribers, for the first two months. Subscribers arrive late and unevenly. Video count is the only number you control, and it is the one that predicted survival in this sample.

Assume the wall is coming. Somewhere in the first twenty videos there is an evening when you do not want to do it. Everyone gets that evening. Deciding in advance what you will do on it is worth more than another week of niche research.

How we collected this

The sample is 120 channels drawn at random from 535 faceless and shorts channels we had collected from YouTube's own channel search, using faceless, shorts and individual genre terms as the queries. For each one we resolved the channel ID from its public page and read the last publish date from YouTube's public RSS feed. Three handles did not resolve and two channels had no uploads at all, leaving 115. No API, no scraping of private data, no login. Measured 8 September 2026.

The script that does this is open source: github.com/plotclip/channel-activity. Point it at a list of handles and it prints the same table.

Corrected 8 September 2026. The first version of this piece counted one upload too many per channel. YouTube's feed carries a channel-level publish date — when the channel was created — alongside one date per video, and our first pass counted them together. That put the threshold at sixteen instead of fifteen, and it wrongly treated two channels with no uploads at all as having published on the day they were created. The numbers above are the corrected ones and the published script has the fix.

What this sample is not. It is not a random sample of all YouTube channels, and it is not a random sample of all faceless channels. These are channels that surfaced in YouTube's own channel search for faceless, shorts and genre terms, which skews toward channels that are at least discoverable. Subscriber counts vary widely across the sample, from single digits to hundreds of thousands, so this is not exclusively a study of beginners.

If you want to replicate it, the method is short: resolve the handle, fetch youtube.com/feeds/videos.xml?channel_id=..., and read the newest published element from inside an entry — not from the feed root, which is the channel's own creation date.

Questions we got asked

Does a dormant channel still earn views? Yes, for a while. Several of the silent channels in the sample have shorts with five figures of views that are presumably still being served. Dormancy is not deletion. But nothing in this sample suggested a channel grows while silent.

Is 15 videos a real threshold? No. It is where YouTube's public feed happens to stop, which is what made the split measurable for free. The finding is that the two groups behave very differently, not that video 15 is a gate.

Does posting daily matter more than posting well? This data cannot answer that — it only measures whether channels posted at all. What it does show is that the surviving group posts on a median gap of three days, so whatever they are doing, they are doing it often.

How many faceless channels are abandoned? In this sample, 47% had gone more than a month without publishing and 20% had been silent for over a year.