The uncomfortable starting point
AI video is very good at producing a first impression. Skin glows. Bottles float. Ingredients become cinematic. None of that proves a person will trust the message, click for details, or buy the product.
Purple Mangos will treat every video as a testable piece of communication—not as proof that an ad worked. We will publish the creative brief, the production choices, and the result when enough data exists to mean something.
What we will test
Each experiment starts with one real shopper question. For K-beauty, that might be texture, routine order, finish, skin-type fit, or the difference between two versions of a product. The video must answer that question before it earns the right to entertain.
- One product and one audience per test
- Three hooks built from the same verified product facts
- A consistent landing page so the video is the main variable
- Clicks and qualified viewing behavior before purchase claims
- Clear labels whenever visuals, voice, or demonstrations are AI-generated
What we will not pretend to know
We will not call a product a favorite before using it. We will not turn a manufacturer claim into a personal result. We will not present an AI face, hand, texture, or before-and-after as documentary evidence.
When a post contains an affiliate link, the commercial relationship will be disclosed. A commission may tell us that a piece of content converted; it does not turn that product into the best choice for everyone.
The scorecard
Early tests will be judged on four levels: attention, comprehension, intent, and purchase. A strong view count with weak retention is an attention failure. A high click rate with fast exits is likely a promise mismatch. Sales matter, but they are the end of a chain—not the only signal worth learning from.
This is the working thesis behind Purple Mangos: show the process, keep the evidence attached, and let useful creative beat empty spectacle.