What to post next, from your own numbers
The best guide to your next ten posts is already sitting in your own account, and almost nobody reads it, because every tool in this category points outward at other people's winners instead. Your catalogue contains a small number of posts that beat your own median by a wide margin. Those are your outliers. Whatever they share — a topic, an entry point, a format, an assumed audience — is evidence about your channel specifically, which is worth more than any general trend because it already survived contact with your audience. Here is how to extract it by hand, and what a tool can honestly add.
Free creator diagnostic
Run a free YouTube channel audit on your own channel
Paste your channel handle and get a free read of the bottleneck holding back your Shorts, uploads, or channel positioning. No signup and no card for the first read.
Why outward-facing outlier tools keep disappointing you
Most outlier finders do one job: they surface videos that beat their own channel's usual reach, across other people's accounts. That is genuinely useful for topic discovery, and it is a different question from the one this page answers.
When you copy a format from someone else's breakout, you are importing the visible parts — structure, pacing, a title shape — without the invisible ones. The invisible ones are usually what carried it: an audience already trained on that creator, a subject that account had established authority in, and a first batch of viewers chosen based on that account's history. Take the format, leave the context, and you get a video that looks like the winner and performs like your median.
This is the honest reason people report copying a proven format and flopping. Not that the tool lied — that the causal ingredient did not travel.
A read on your own outliers has the opposite property. The context is already yours.
The method: a table you can build in thirty minutes
No tool needed. Open a spreadsheet.
- List your last 25 to 30 posts. Skip anything under two weeks old for long-form or 72 hours for short-form; unfinished distribution will mislead you.
- Record views for each, plus subscribers or followers gained if the platform reports it per post.
- Compute your median views — the middle value. Not the average.
- Mark every post above 2× your median. Those are your outliers. Most people find between two and five.
- Tag every post on five attributes: topic, format, who it assumes you already know (regulars only, or anyone), first-frame legibility, and length.
- Read the tags on the outliers as a group. You are looking for an attribute that shows up in most of your winners and few of the rest.
The output is one sentence: *"my posts that travel are the ones that ___"*. That sentence is your brief for the next ten.
What the attribute usually turns out to be
After you have done this a few times, patterns recur. In rough order of frequency:
Assumed context. The winners are the ones a stranger could understand. This is the single most common finding, and it is uncomfortable because the posts that assume context are usually the ones your regulars like most.
Entry point, not subject. Same topic, but the winner opened on the result and the flop opened on the setup.
Specificity. Narrow, concrete posts often beat broad ones, because a specific promise is legible in two seconds and a broad one is not. "How I fixed X" over "tips for Y".
A repeatable series shape. Posts that were episode three of something outperform one-offs, because the audience recognises the format before parsing the content.
Nothing at all. Sometimes the winners share no attribute, which is a real result: your variance is chance-dominated and your effort belongs in raising the floor rather than in chasing a pattern. Why one video blows up and the next dies covers what that means.
Turning the attribute into your next ten posts
A rule that has held up better than "just make more of what works": eight and two.
Eight posts that carry the winning attribute. Not eight remakes of the same video. Eight different subjects, each built with the property your outliers share. If your winners are the ones a stranger can follow, then eight posts where nothing requires prior context.
Two deliberate bets that break the pattern. A format you have not tried, a subject one step outside your lane, a different length. These are not filler. Without them your catalogue narrows toward a single audience and you stop discovering anything, which is the mechanism behind a lot of stalled channels.
Then re-run the table. The attribute changes as your audience changes, so this is a loop rather than a one-time exercise.
Common failure modes of this method
Reading one outlier as a pattern. One post is one observation, and outliers are partly luck. Wait until you have three above 2× median before restructuring anything.
Tagging after you already know the answer. Tag every post before you look at which ones won, or you will find the attribute you expected.
Using the average instead of the median. A single breakout drags the mean up so far that genuine outliers stop showing as outliers.
Ignoring the follower or subscriber column. Views and recruits are different outcomes. Your highest-view post and your highest-recruiting post are frequently not the same post, and if growth is your goal the second one matters more.
What tools can honestly do here
Disclosed interest: we sell creator diagnostics, so read this as an interested party being explicit about scope.
What genuinely needs a tool: watching your own videos with fresh eyes. You cannot judge your own first two seconds, because you already know what the post is about. That blindness is the real reason the manual method sometimes stalls — people tag "first-frame legibility" from memory rather than from viewing.
What our tools do, precisely:
- Channel X-Ray is the own-catalogue read. It works out your typical performance against your own proven high, scores which topic lanes pay back and which drag, puts your best and worst videos side by side, and names one bottleneck instead of returning a dashboard. It reads YouTube and Instagram. Where the data does not support a conclusion, it says so rather than filling the space.
- Reel IQ is the per-video read, before you publish. Hook, pacing, visual change, caption clarity, cover frame, retention risk.
- Viral Radar is the outward tool — topic search across other creators' Shorts and Reels, surfacing videos outrunning their own channel's usual reach. It is deliberately not grounded in your posting history, so it answers "what is working in this topic" rather than "what works for me". Useful, different question, and worth knowing which one you are asking. It costs credits after the signup grant.
What no tool can do: tell you which of your outliers was luck. Three winners sharing an attribute is evidence. One winner is a story.
If you want the free version of the catalogue read with no account, the free YouTube channel audit and the free Instagram Reel analyzer both run a lighter pass and neither asks for a card.
"What should I post next if I want my next ten to move the channel?"
Moving a channel and getting views are different targets, and the table above answers both if you record both columns.
If views are the goal, weight toward the attribute your high-view outliers share.
If growth is the goal, weight toward the attribute your high-recruiting posts share, which is very often a series shape or a clearly stated point of view rather than the broadest topic. A post that reaches a million uninterested people moves nothing.
And if the two columns disagree sharply — your biggest posts recruit nobody — that is itself the finding, and the fix is a channel-packaging problem rather than a topic-selection one. On Instagram that is unpacked in views not converting to followers; on YouTube, in views to subscriber ratio.
Common questions
How do I figure out which of my own videos I should make more of?
List your last 25 to 30 posts with their view counts, compute the median, and mark every post above twice that median. Those are your outliers. Tag every post on topic, format, how much prior context it assumes, first-frame legibility and length, then look at which attribute your outliers share and the rest do not. That shared attribute is your brief for the next ten posts.
Is there a tool that finds the outliers inside my own uploads rather than other people's?
Most outlier finders point outward at other creators by design. The own-catalogue read is a different job: comparing your posts against your own median rather than against anyone else's. We build a channel diagnostic that does it, which is a disclosed interest, and you can also do it by hand in a spreadsheet in about thirty minutes. The manual version teaches you your own baseline, which the automated version cannot.
I copied a format straight from an outlier tool and my video still flopped. Why?
Because you imported the visible parts and not the invisible ones. Someone else's breakout carried an audience already trained on that creator, established authority in that subject, and a first test batch chosen from that account's history. The format was rarely the causal ingredient. Reading your own outliers avoids this, since the context is already yours.
What should I post next if I want my next ten videos to actually move the channel?
Split them eight and two. Eight different subjects that all carry the attribute your own outliers share, and two deliberate bets that break the pattern so your catalogue keeps discovering. And be clear which outcome you are optimising: the posts that get the most views and the posts that recruit the most followers are frequently not the same posts, and only the second kind moves a channel.
How many videos do I need before this method works?
About 25 to 30 posts, and at least three that beat twice your median. One outlier is one observation and is partly luck, so restructuring a channel around a single hit usually backfires. If your winners share no attribute at all, that is a genuine result: your variance is chance-dominated, and the useful work is raising your floor rather than chasing a pattern that is not there.
Is Viral Radar the same as looking at my own best posts?
No, and it is worth keeping the two apart. Viral Radar searches topics across other creators' Shorts and Reels and surfaces videos outrunning their own channel's usual reach. It is not grounded in your posting history, so it answers what is working in a topic rather than what works for you. It also costs credits after the signup grant. The own-catalogue question is what Channel X-Ray and the spreadsheet method above are for.
Free creator diagnostic
Run a free YouTube channel audit on your own channel
Paste your channel handle and get a free read of the bottleneck holding back your Shorts, uploads, or channel positioning. No signup and no card for the first read.