Bot: find the videos that broke the average

A video with 500 thousand views on a channel of 5 million subscribers is a normal video; one with 200 thousand on a channel of 20 thousand is a signal. This bot looks for the second kind: videos that broke the average of their own channel, which is where the replicable idea lives. ## By the numbers - It compares against the **channel's own baseline**, not absolute views. Without that you only find big channels. - A video is an outlier when it passes **3 times** its baseline. Below 2 times it is usually the algorithm, not the idea. - Start with **3 to 5 channels** you already follow, so you can judge whether the findings are real. ## The prompt ``` Set up a new bot that researches YouTube. Guide me to connect YouTube and then configure it: 1. In the niches I give you, look for outlier videos: the ones that performed far above normal FOR THEIR CHANNEL. 2. Send me a list with: title, channel, views versus that channel's baseline (give me the multiple), and what that video did differently. 3. In "what it did differently" be concrete: the hook of the first 5 seconds, the title, the thumbnail, the format, or the topic. Do not tell me "good content". 4. Mark which ones are replicable for my Hispanic audience and which are not. Ask me: my niches, 3 to 5 channels I already follow, and how often I want the list. Run ONE search so I can verify that the outliers are real before saving it. ``` ## What to change before using it - **The multiple.** 3 times the baseline works for small niches; in big niches raise it to 5 or you fill up with noise. - **The language filter.** Adding "mark the replicable ones for a Hispanic audience" is what makes it useful here: half of the English hooks do not transfer. - **The frequency.** Weekly is enough. Daily gives you the same videos with one more day of views. ## Frequently asked questions **Is this for copying videos?** No. It is for finding the pattern (the hook, the format, the angle) and applying it to your topic. Copying someone else's video does not work and it also ages badly. **Why against the baseline and not by views?** Because absolute views only reward big channels. The baseline shows you where an idea beat what the channel normally achieves, and that is the replicable part. **Does it work for Spanish-language content?** Yes, and there is less competition there. Ask for Spanish-language niches explicitly, because by default it will bring you what is best measured, which is in English. ## Credit Adapted from [Youtube Outlier Finder](https://botdirectory.ai/bots/youtube-outlier-finder/) by `petergyang`, from the [botdirectory.ai](https://botdirectory.ai) directory (MIT, Copyright 2026 Inbox Zero Inc.). Our version fixes the multiple (3 times the baseline) that the original left undefined, demands concreteness in the "what it did differently" instead of accepting vagueness, and adds the replicability marking for a Hispanic audience.