before-after-transformation-videolisted

# Before-and-After Transformation Transformation is the highest-retention format in short-form. The viewer sees a problem state, knows a resolution is coming, and stays for it. That's a complete narrative in six seconds. It's also the format most likely to produce a false claim, because the model will happily generate a result the product doesn't deliver. ## Why it holds attention The format works because it front-loads a promise. The "before" frame is the hook — it establishes stakes and implies the payoff without any copy at all. Nothing else in short-form does that as efficiently. Which means the **before state is the more important half**, and it's the half people rush. ## Generate as two beats, not one Asking one generation to morph a dirty surface into a clean one produces the characteristic melting transition that reads as AI immediately. ``` Beat 1 "before" 3-4s the problem state, static or slow motion Beat 2 "action" 4-5s the product being applied Beat 3 "after" 3-4s the result state ``` Three separate generations, cut together. The cut does the transformation, which is both more convincing and more controllable than a generated morph. **Match the framing exactly across beats.** Same camera position, same distance, same angle, same light. That's what makes the comparison legible — a before and after from different angles proves nothing and looks evasive. ``` FRAMING = ("Fixed camera, 45 degrees above, subject centred, " "soft window