YouTube Outlier Finder

How this YouTube outlier finder works

An outlier is a video that performs unusually well relative to its own channel. A video with 50,000 views could be exceptional for one creator and ordinary for another. This tool looks at the channel’s public uploads to give that number context. It does not compare a small creator directly with a large creator, and it does not use subscriber count as a substitute for normal video performance.

Paste a public YouTube video URL, a channel’s @handle, or a /channel/ URL. A video link checks that specific upload and finds other standout videos from its channel. A channel link starts with its recent uploads. Old /c/ channel links are not supported; copy a public video link from that channel instead. You do not need to connect your YouTube account.

Read a bounded sample

We request the latest 50 entries in the channel’s upload list, then retrieve the available public video details. This is not a scan of the channel’s entire history. Private, deleted, and unavailable uploads can reduce the number returned. The results show how many public uploads were read, how many met the filters, and how many had enough comparison data.

The first version considers videos published between 7 and 180 days ago and longer than three minutes. It excludes identifiable livestreams and premieres. This duration restriction keeps the first version focused; it is not a claim that duration alone perfectly identifies Shorts. A linked video outside this window receives an explanation instead of an unsupported ratio.

Compare nearby uploads

For each candidate, the comparison group contains other eligible uploads from the same channel, published within 60 days of the candidate and between half and twice its duration. We select up to 20 nearest uploads by publication date. At least 10 are required. The candidate never contributes to its own baseline, and ordinary low-view videos are not removed simply because their results are low.

These filters match timing and duration, not subject matter. If a channel recently changed its audience or content format, inspect the comparison list before relying on the result. A gaming tutorial and a channel announcement might have similar lengths but very different audiences.

Understand the reference ratio

The calculation is current video views ÷ median views of the comparison group. For example, 64,000 views against a median of 8,000 gives an 8× reference ratio. This is an illustrative calculation, not a real creator result. The median is the middle view count after sorting, or the average of the two middle counts for an even-sized group.

A ratio of at least 3× places a video in the candidate list. This is Thumbrix’s screening threshold, not an official YouTube classification or a universal standard for success. Expand “View comparison” to see the underlying uploads, dates, views, and calculation. If fewer than 10 comparisons exist, or their median is zero, we explain why no ratio is available.

All figures compare current cumulative views. They do not compare videos at the same number of days after publication. An older upload has had longer to accumulate views. We do not reconstruct historical growth, calculate current hourly velocity, infer CTR, or claim that a thumbnail caused the performance. Public data cannot establish those conclusions.

Turn a reference into your own thumbnail

Choose “Use for my video” when you find an example relevant to your audience. Enter your own video topic and what viewers should remember. You can keep or remove the source title and link before continuing. Your own facts should drive the creative brief; the example provides context rather than a title to copy.

Continue to Thumbrix’s existing creative planner and review the brief before starting. Planning three directions costs 1 Credit. Image generation costs 5 Credits per image after any available free uses. Selecting a reference, filling the brief, and moving to the planner do not start a paid operation. The detector itself does not call an AI model or consume AI Credits.

For more research context, read how to find outlier videos on YouTube. For composition and readability, see how to make a good YouTube thumbnail.

YouTube outlier finder questions

Is the lookup free?

Yes. Each network can make three fresh lookups per UTC day. Cached results remain available within the short-term request limit. Public results may be reused for up to one hour; the retrieval time is shown alongside the channel. Upstream failures return the fresh-lookup allowance. AI planning and image generation are separate, explicitly priced actions.

Why are there no standout videos?

Either no comparable video reached the 3× threshold, or the sample had too few comparable uploads. These are different outcomes, and the result explains which applies. A channel with many short videos, infrequent uploads, or a recent format change may have too little usable data. Try another channel or a specific public video URL.

Does the tool analyze the video or thumbnail?

No. Detection uses public YouTube metadata, view counts, and arithmetic. It does not download the video, analyze its audio, or visually score its thumbnail. The reference ratio is calculated by Thumbrix and is not a metric supplied or endorsed by YouTube. Review the source video yourself before treating it as relevant to your own project.