Properly formatted, timed to the second, and written knowing the first three seconds decide everything else.

The hook is not the first sentence, it is the first image plus the first four words. We write both and mark the timing.
Every script is read aloud before delivery. If the presenter runs out of breath, the sentence is wrong.
Word count maps to runtime at roughly 150 words per minute for ads, 170 for explainers. We deliver to the second.
If a line needs two visuals to make sense, it is two lines. Editors should never have to guess.