YouTube Channel Analytics

Any public channel's recent videos, ranked by what they actually did. Shorts counted separately, outliers marked.

What YouTube Studio will not show you

Studio is good, and it only works on channels you own. The moment you want to look at somebody else — the channel you are competing with, the one you want to learn from, the one you are thinking of sponsoring — you are back to scrolling their grid and guessing.

This reads the last fifty uploads of any public channel and puts them in order of what they did, which is the view nobody gets for free anywhere else.

The two columns that matter, and why

• Views per day is the one that survives age. Fifty uploads can span three years, and raw views reward whatever was published first. A video from last month at 4,000 views a day is outperforming one from 2023 with a bigger total.
• Multiples of the median shows how far each upload sits above or below that channel's typical recent video. A video at two times the median or more did not do well because the channel exists; something specific about it landed. That is the one worth looking at twice, and it is marked in the list.

Medians rather than averages, throughout. One video that went unusually far would otherwise lift the average and make every ordinary upload afterwards look like a failure.

⚠️ Shorts are counted separately, and that is not a detail

YouTube's Data API has no isShort field and is not going to grow one, so every tool that separates them is inferring it from something else. Most infer it from length alone, which stopped working in 2024: a Short can now run up to three minutes, so a rule like “under a minute” quietly files half of a channel's Shorts with its long videos.

This page reads the shape as well. The API returns the dimensions of a video's embed, and a Short comes back vertical (800×1422) where an ordinary video comes back wide (800×450). A video counts as a Short when it is vertical or square and short enough for the rules that applied when it was published — one minute before October 2024, three minutes after, because YouTube did not convert the back catalogue.

⭐ And where that is still not enough to be sure, the page says so instead of guessing: those videos are counted as format unclear and left out of both medians. A slightly smaller sample beats a median with the wrong videos in it. On one channel we tested, the old length-only rule had the median of the long videos wrong by 118%.

If a channel runs both, the page also tells you which format is carrying it. That comparison is often the single most useful line on the screen, and it is one Studio does not put in front of you either.

Where the numbers come from

The official YouTube Data API, the same source YouTube serves to everyone: views, likes, comments, durations and publish dates as YouTube reports them. Not estimates, not a proprietary score out of a hundred. Where a channel hides a number, we leave it blank rather than filling the hole with a guess.

Where to go from here

Questions