Verified against ChatGPT · 2026-07-22
Write a long-form script structured around where viewers actually drop off
Builds a full script around a cold open, a stated open loop, and a fixed pattern-interrupt cadence, and treats a past video's real retention-graph dip as a structural constraint rather than writing generically "engaging" prose.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are a YouTube scriptwriter building a long-form script around the platform's actual retention mechanics, not around what merely reads well on the page. A script that is well-written prose but ignores where real audiences bail — the first 15 seconds, right after a slow intro, at the midpoint lull — loses exactly the viewers whose watch time the recommendation system weighs most heavily. VIDEO TOPIC Why your iPhone battery degrades faster in year one than year two TARGET LENGTH 9-11 minutes CHANNEL AND AUDIENCE Tech explainer channel, 140K subscribers, average viewer age 22-34, prior videos in this format run 8-12 minutes PAST RETENTION DATA Last video on a similar topic fell from 71% to 38% retention between 0:45 and 1:30 — that section was a slow lithium-ion-history preamble before reaching the actual answer KEY POINTS TO COVER Battery chemistry degrades faster under heavy fast-charging cycles; software throttling masks the decline from the user; leaving the phone at 100% charge overnight is the single biggest accelerant most viewers don't know about STRUCTURE RULES Open with the payoff or central tension of the video within the first three sentences — not a channel intro, not "hey guys welcome back," not scene-setting. State plainly, in the viewer's language, what they will know or be able to do by the end, and open at least one specific unanswered question — an open loop — that the rest of the script is structured to resolve, not to answer immediately. Place a pattern interrupt — a new visual beat, a tone shift, a new sub-question, a cut to a different setting — at minimum every 60 to 90 seconds; a script that runs three or more straight minutes in the same visual and vocal register is a script asking to be scrubbed past, regardless of how accurate its content is. Never place the single most interesting point at the very end as a "big reveal" unless the whole script is explicitly built as a countdown toward it with stakes restated along the way — a surprise nobody was reminded to wait for lands as filler, not payoff. If the retention data shows a specific drop-off timestamp from a past video, treat that as a structural constraint on this script: name what likely caused it — a slow setup, a tangent, a repeated point — and build this script's equivalent section deliberately shorter or restructured around a fresh hook, rather than assuming the past drop was a fluke. Resolve every open loop you open; a curiosity gap that lures a viewer in and never pays off reads as a bait-and-switch and drives an explicit "not interested" signal, not just a soft attrition tick. OUTPUT FORMAT 1. A cold-open hook (first 10-15 seconds), written as spoken lines, not a summary of what the hook should do. 2. A full script broken into labeled sections with a timestamp estimate for each, spoken lines throughout, [bracketed] notes for visual cuts or B-roll cues only where they matter to pacing. 3. A one-line note under each section naming the pattern-interrupt device used. 4. A list of every open loop opened in the hook or early sections, each paired with the exact section where it gets resolved.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
Audience retention — the percentage of viewers still watching at each timestamp, shown as a graph in YouTube Studio — is a direct input to what the recommendation system keeps surfacing a video for, which is why this prompt treats a past video's actual retention dip as a structural constraint rather than a footnote: a script that restructures the exact section type that caused a 33-point drop last time is targeting the real signal YouTube measures, not a generic instruction to "keep it engaging" that gives the model nothing concrete to act on. The fixed 60-to-90-second pattern-interrupt cadence exists because the steepest retention declines on most channels' own graphs cluster around long uncut stretches of the same visual and vocal register — a static talking-head shot with no cutaway, no tone change, no new visual for three-plus minutes — so forcing a beat change on a fixed interval is a direct countermeasure to the failure mode the data itself shows, not an arbitrary pacing rule. Requiring the hook to open a genuine open loop, and requiring every open loop to be explicitly resolved somewhere in the script, targets a specific and common failure of AI-drafted hooks: a curiosity-gap opener that never actually gets answered reads to a real viewer as a bait-and-switch, which produces an explicit negative signal — a "not interested" click or a dislike — that is measurably worse for a channel than a viewer who simply drifted off partway through, because negative explicit feedback suppresses future recommendations more directly than soft attrition does. The rule against burying the single best point as an unearned "big reveal" addresses a subtler version of the same problem: viewers who leave before the reveal never experience the payoff at all, so a script that saves everything for the end is optimizing for a viewer who, per the very retention data this prompt asks for, statistically will not still be there. Naming the pattern-interrupt device used in each section, rather than leaving pacing as an invisible design choice, also gives whoever storyboards or edits the video a concrete cue to shoot or cut for, instead of a script that reads fine on paper and only reveals its pacing problems once it is already filmed.
Verified against
ChatGPT GPT-5.1 · 2026-07-22
Claude Sonnet 4.6 · 2026-07-30
Changelog
- 2026-07-22 — Initial publish, verified against ChatGPT (GPT-5.1) and Claude (Sonnet 4.6).
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
