NOAH / CREATIVE STUDY

Fictional skincare spec · Meta hook comparison

Three hooks,
one asset set.

One fictional skincare clip set, cut three ways for a Meta hook comparison. The first two seconds change; the eight seconds after them are frame-identical.

Product close-up

The pump head at native pixels with a slow push, under a serif title.

Loading A.

Product first: the object does the stopping, and the title names the ritual before anyone speaks.

Presenter held

Her first frame, mouth closed, held for two seconds with one boxed question.

Loading B.

Presenter first: a face and a two-word question, then the line lands the moment the body starts.

Caption interrupt

A charcoal card with two lines of heavy type and nothing else in the frame.

Loading C.

Caption interrupt: a flat dark card in a soft ivory feed, naming the moment the ad is served.

Each file is 10 seconds, 9:16, 720 by 1280. Seconds 2 to 10 are the same presenter line, product finish, end card and Learn more pill in all three.

What each one would be judged on

Same audience, same measure.

One audience and one order of metrics for all three, so the rows can be compared cleanly. Nothing here has run; the columns show how the comparison would be set up.

Hook, angle, audience and measure for each variant
VariantHook, first two secondsAngleAudienceWhat would be measured
AProduct close-up: pump head at native resolution, slow push, "Make room for your evening ritual."The quiet ritualCold, broad3-second view rate first, then hold rate
BPresenter held: her closed-mouth first frame with the caption "Rough day?"A person tells you to pauseCold, broad3-second view rate first, then hold rate
CCaption interrupt: charcoal card, "Still scrolling at 9 pm?" then "Put the phone down."The 9 pm scrollCold, broad3-second view rate first, then hold rate

Fictional product and AI presenter. This is a spec built to show the method: no campaign, brand or performance data sits behind it. One presenter line exists, so the angles live in the openings and captions.