We built this straight from DTC Midas’ thread (linked above) and it was fucking insane. It basically turned into its own engine for spitting out winners, giving us more data on what worked, and scaling image ads like crazy.
⚙️ How We Set It Up
- Columns directly copied from DTC Midas’ setup
- Priority levels:
- Critical → major sales / promos
- High → fast iteration on a winner
- Medium → general testing ideas
- Our addition:
- Whitelisting Page Column → this was sauce
🔑 Whitelisting Iterations
This was our edge. Let’s say we had a winner on the brand page → advertorial.
- We’d remake it for the MD page (authority style) → even bigger winner.
- Then remake again for the personal whitelisting page → another angle hitting.
- Same concept, same copy → just reframed for the persona running the page.
That system printed winners stupid easy.
📚 Feeding Claude
- Every brief + output → uploaded into Claude project.
- Claude learned what winners looked like, what flopped, and got sharper.
- Editors had creative freedom on design, but briefs locked in purpose + intent.
- When a winner hit → marked High Priority in Monday → team moved instantly.
👏 Team Ops
- Slack → I always congratulated wins.
- Kept team urgent + motivated without burning them.
- Everyone knew when something hit, and new briefs flowed instantly.
💣 Winning Styles
What printed the most winners for us:
- Lead Gen style ads → clickbait / news style / aggressive scroll-stoppers.
- Not full black-hat, but crafted to drive dirt cheap CPC from unaware audiences.
- Funnel: cheap click → listicle/advertorial → walk them into purchase.
Example:
- One ad spent $8K/day
- CPC ~ $0.20
- CVR only ~ 2%
- Still crushed because traffic was dirt cheap and scale carried conversions.
With Monday + whitelisting + lead gen style + Claude analysis → we were spitting out new winners every other day. Upload, test, scale.