A competitor channel analysis prompt works when the model sees the full record, including the videos that flopped. Show it only the hits and it will explain success with traits that also describe the flops.
The honest version compares winners and losers on the same channel, then asks what actually separates them rather than what the channel always does.
Pull a stretch of uploads with their view counts and dates, not just the top ten. Include the quiet videos. Those are the control group that keeps the analysis from turning into flattery.
Note the format, length, and thumbnail style for each one so the model has something to compare beyond the raw numbers.
A trait that shows up in every video, hit or miss, is a habit, not a cause. The prompt should force the model to separate the two before it draws a conclusion.
If a pattern also describes the flops, it is not a reason for success. It is just something the channel always does.
Copying a competitor's format rarely transfers, because their audience differs from yours. What travels is the underlying structure: the hook shape, the pacing, the promise in the title.
Test one borrowed structure on your own audience before you rebuild your approach around it. Their result is a hypothesis for you, not a template.
Recheck the competitor every quarter. What worked for them last year may be a saturated move by now.
Feeling your way through a script works until it doesn't. Pattern extraction turns 'this channel feels good' into a repeatable, teachable formula.
One viral video teaches you almost nothing about a channel. Twenty ordinary videos reveal the structure that made it worth studying at all.
A blueprint you admire is useless until it becomes a limit you write inside. Here is how to convert observations into constraints that speed up drafting.
Analyze a channel, extract its structural DNA, and generate scripts engineered from patterns that actually work.
Analyze a Channel