How to turn one short into ten new scripts

A creator I work with asked me last week how to turn one winning short into ten new ones. He'd read a thread on it. The thread said generate ten variations of the topic. He'd done it. He had ten variations. They were all bad.

The reason they were all bad: he was generating ten ideas, not ten angles. Ten ideas is a lot of ideas. Most of them will be weak, because good ideas are scarce. Ten angles on one good idea is much easier and much more useful, because angles are just rotation around something that already worked.

Why "ten new ideas" is the wrong frame

If you start with a winning short and ask "give me ten more video ideas like this," your brain or your AI tool will generate ten topics. Each topic needs its own hook, its own structure, its own validation. Nine of the ten will be worse than the original because the original was already the strongest hook in that space.

If instead you ask "what are ten ways to come at this same insight," you keep the structural validation the original gave you, and you only vary the dimension that's making it new. That's a lot less work and a lot more reliable.

The unit of repurposing isn't the topic. It's the angle.

The five angles I keep coming back to

I have a list of five rotation axes I cycle through. Not five angles total — five types of angle, each of which produces multiple scripts.

Audience swap. Same insight, different audience. The calorie-tracking insight that worked for general fitness audiences can be reshaped for new moms, for people in their 50s, for athletes, for people on GLP-1 drugs. Each one is a different script because the entry context is different, even though the core insight is identical.

Contrarian flip. Take the conclusion of the original and argue the other side as a hook, then route to a more nuanced version of the original conclusion. "Everyone says X. I used to say X. Here's why I changed my mind." The original conclusion stays intact; the script just makes the journey more interesting.

Time scale. The original gives the 30-second version. Make a 15-second version that just delivers the punchline cold. Make a 60-second version that explains the mechanism more carefully. Make a long-form version for YouTube or a newsletter. These are not the same script in different lengths — the proportion of setup, mechanism, and payoff changes radically.

Format swap. The original was a talking-head walkthrough. Make a voiceover-over-b-roll version. Make a screen-recording version. Make a stitched reaction to another video. Same insight, different stylistic vehicle. Each pulls a different sub-audience.

Specificity zoom. Take one sentence of the original and treat it as the entire next video. The original said "your sleep architecture matters more than your protein." Make a script entirely about deep sleep architecture. Then another about REM. Then another about the room-temperature thing. You're not running out of material — you're using the material the original couldn't cover.

Five axes. Pick any two and you have ten scripts. Pick all five and you have material for a month.

The shareability check

Not all ten are worth publishing. Before recording, run each script through one question: would anyone share this without me asking them to?

The audience swap ones tend to pass — they're shareable to specific people. ("Hey new mom friend, this one's for you.")

The contrarian flips pass when the framing is genuinely new, not just contrarian for show. If your contrarian take is a take you didn't actually hold a year ago, it lands. If it's a take you invented for the video, the viewer can tell.

The time scale ones split. The 15-second cold-payoff version is highly shareable. The 60-second careful version is more for replay and save than share — still useful, but a different signal.

The format swaps depend on whether the new format adds something. A voiceover-over-b-roll version of a talking-head script is just a remix. A screen-recording version of a fitness script is incoherent. Be honest about which swaps actually fit.

The specificity zooms almost always pass. They're standalone takeaways, which is the most shareable shape there is.

Drop any that fail the check. You'll usually end up with six to eight from your initial ten. That's fine — six to eight scripts from one winning insight is a month of content.

The trap I see most often

Creators get attached to all ten of their initial variations because they did the work. They publish all of them. Three do okay. Seven die.

The seven that die didn't just waste a slot. They trained the algorithm to think the channel is mid. The algorithm doesn't grade on a curve; it grades on absolute completion and replay. Seven mid videos pull down the channel's reach for the three good ones.

It's the same lesson as with hooks (see the post on five hook patterns): it's easier to delete than to fix. Generate ten. Keep six. Ship six. Don't shop the cutting room floor.

What the workflow looks like, condensed

Drop a winning short into a structural extractor or do it manually with a pen. Get the beat map. Apply the five angle axes mentally, picking two that fit your channel best. Write seven to ten scripts. Shareability-check them. Keep what survives.

That's a 90-minute process for one winning short → six shippable scripts. Or it's the workflow ShortRemix is built to produce — same five angles, same shareability scoring, same shape of output. Either way, the unit you're working in is angles, not ideas.