The next film started with the last one.
We didn't begin each month with a blank page. We looked at what people watched, where they left, which hooks held them and what the account's performance told us to try next.
My part: I led creative strategy and direction, and built the system connecting performance insight, scripts and production. The media-buying decisions were not mine.
Different stories for different people.
We tried different openings, situations and tones across markets. These are four frames from the work; I share the full films privately on request.
Four frames, four different ways into a story. The full films are available privately.
Ask to see the films ↗Each round gave the next one a brief.
This is not a straight production line. The work went out, produced evidence, and came back to the team as a better starting point.
Review results and analyse the films frame by frame.
Look at hooks, viewing hold and which stories worked for different audiences.
Combine those findings with relevant trends. A script assistant helps; the team selects and shapes the scripts.
AI artists make the approved directions. The new films then generate the next report.
More work, with a reason behind it.
The account averaged about 150 films a month and exceeded 250 in some months. Over three months, reported ROAS reached three times its starting level. My team contributed the creative system and the films; we did not run media buying.