Why your prompts give you forty versions of the same answer
The five-part shape behind a prompt that produces something usable — including the step almost everyone skips — with the same request written badly and then properly.
Why the first 1-3 seconds decide everything, hook patterns that consistently work, and how to A/B test hooks on a small budget.
Nothing else in a short-form video matters if the hook doesn't work — a strong hook is the single highest-leverage thing you can improve about your scripts.
On Shorts, Reels, and TikTok, the cost of scrolling away is close to zero, and viewers make that decision almost instantly. A hook's job isn't to be clever — it's to make continuing to watch feel more interesting than scrolling to the next video.
The strongest hooks usually combine two of these rather than relying on just one.
Generic questions ("Did you know...?"), slow scene-setting before the actual hook, and hooks that oversell what the video actually delivers all tend to underperform — the last one specifically damages retention on the rest of the video once viewers feel misled.
You don't need paid ad spend to test hooks — publish the same core content with different opening lines across multiple Shorts/Reels over a few weeks and compare average view duration on the first few seconds specifically, which most platform analytics dashboards expose directly.
Should the hook match the thumbnail's promise exactly?
Yes — a mismatch between what the thumbnail/cover promises and what the hook actually delivers is one of the fastest ways to spike early drop-off, since it reads as a bait-and-switch the moment the video starts.