Cheers films more than most brands ship: 150 raw pieces in a single week, 75 to 100 TikToks across seven accounts, and founder-led science that sells. The bottleneck isn't output, it's learning. Videos launch and nobody has time to ask why they worked, and your creative performance lead writes every brief by hand, week by week. What gets installed is the system behind your Winner's Playbook: every ad tagged so cut, double down, and explore come straight from the data, briefs drafted for your team to approve, a bank of every clip you film, and training that gets your editors building on the same system. Three months, built alongside you, in place before holiday party season.
From our call on October 28. Seven self-taught editors in Latin America cutting both filmed and AI footage, Brooks leading the science, an in-house creator plus Houston actors, and a creative performance lead writing every brief by hand. Two pieces of the call this proposal is built around:
"We're cranking out so much content. Once that video is done, it's launched. We don't take the time to go and see why it performed. And if we do, it's all manual."
That's the gap the Winner's Playbook is meant to close, and it can't close by hand at your volume. The read on what to cut, double down on, and explore should be waiting for you, not something someone builds on a Friday.
"She's racking her head because it's so much work coming up with new concepts ahead of time. She does it week by week. Setting up something like this would make her life a million times better."
Every brief for Brooks, your in-house creator, and your actors starts from a blank page today. The system drafts them from what's working and what's never been tested, and she becomes the editor of the briefs, not the author of every one.
Every ad gets tagged by persona, angle, and format, so a winner tells you why it won: the person, the promise, or the way it was shot. Then cut, double down, and explore stop being calls and become a weekly read: 60% iterations on what's working, 20% reskins of winners, 20% new ICPs mined from Reddit and TikTok.
Your top 20 live ads are all founder-led video, one to three and a half minutes, almost all on the science of the morning after. That's a strong combination. The data tells you how deep it goes and what's next.
When we pulled the full history for a protein brand, gut-health ads that had done 3 to 4x two years earlier were sitting untouched. The first new batch built on them performed right away.
With 289 ads live and hundreds more behind them, Cheers almost certainly has the same buried winners. Gemini and Claude watch and transcribe every ad, so the analysis covers what Brooks says, not just what the thumbnail shows.
Every clip you film gets sliced, transcribed, and tagged into a bank your editors can search. New hooks on winning founder videos, mashups, and reskins become a request, not a re-shoot.
Your editors keep the taste. The system takes the grunt work, and the training teaches them to edit with Claude instead of trial and error.
Four questions, four straight answers.
Month 2 is built around them. Working sessions with your editor manager and the team, booked for their hours, on the skills we use every day: cutting from raw footage by text, captions timed to the word, B-roll pinned to what's being said, reskins and mashups on demand. They stop guessing and start building on what's already proven.
That's exactly why we composite instead of generate. Your real product photos, real fonts, and real logo are placed in code, at the right size, on top of the scene. The jar is never drawn by AI, so it never looks AI. Same approach for the premium brands we run today.
Three months. Weekly working calls in month 1, bi-weekly after, and Slack throughout. Everything is built in your environment on your Claude Max, connected to the tools you already use like Atria, so it's yours when we're done.
Your approved claims list gets built into the brief and copy checks, so nothing new goes out saying something you can't say. Edutainment stays science-led and on-claim by default. Your team still signs off before anything ships.
Three tranches. Foundation lands first because nothing downstream works without it. Production takes the brief grind off your creative lead. Video and Editors gets the team building on the system.
Weekly calls in the build month, bi-weekly after, once your install is running and just needs tuning. Slack-embedded throughout. I keep it to a handful of consulting clients a quarter, so the work stays deep, not templated.
Heavier in month 1 because that's where the build happens. Lower after because your install runs itself with light calibration once it's built. 3-month minimum committed up front, no lock-in beyond it.
So nobody's surprised in month 2. If any of these become important, we scope them separately.
The sprint pace, assuming we sign this week. Each item below is what's true at the end of that step.
SOW countersigned, M1 invoice paid. Slack channel created, kickoff booked for next week.
Set up on your Claude Max, Atria and Meta connected live on the call. The full ad history pull kicks off and runs unattended overnight.
Persona, angle, format, and talent on every ad. Cut, double down, and explore flagged. Buried winners and untested ICPs on the table.
Brief drafts for Brooks, your creator, and your actors on the 60/20/20 mix, compliance-checked. Your creative lead reviews instead of writing from scratch, ahead of holiday party season. We move to bi-weekly + Slack.
Footage bank built from your raw clips. Working sessions with your editing team: real-footage cuts, new hooks on winners, editing with Claude Code.
Landing pages for winning angles, composited statics for paid and email, feedback loop running. Month 3 closes with a handoff doc and a "running it without me" checklist.
Everything's laid out above. If it's a go, I send the SOW + Stripe link the same day and we kick off next week, with briefs flowing before holiday party season.
