CREATIVE TESTING SYSTEM FOR DTC BRANDS
Creative stops being a guessing game when it runs on a system.
Five years in DTC performance marketing, mostly beauty, health, and wellness. My take on creative: test one thing at a time, judge the concept before the edit, and let clear rules decide what scales. That’s why every ad runs through the same brief template, test matrix, and naming convention.
.png)
Meta • Google • AppLovin • Vibe • MNTN • Triple Whale • Shopify • Shopify Collabs • Insense • Trybe • ShopMy • LTK • Social Snowball • Motion • Billo •
THE PROBLEM
DTC brands launch a lot of creative and learn very little from it.
WHY IT MATTERS
Creative is now the main lever.
When creators, editors, and media buyers each work from their own assumptions, every result is a mystery. A win could be the angle, the hook, the offer, or the edit, and you can’t repeat what you can’t explain.
Meta’s advice has shifted from narrow targeting to creative variety. The algorithm finds the audience, so the ad has to do the work. A system helps you find strong concepts faster, and keep finding them.
The system I built, from brief to scale
1. Brief
I brief every ad the same way: hook, angle, format, proof, and offer. That way creators know the goal before they pick up a camera.
2. Test
One variable per test. If I change the hook, the angle, format, and offer stay put, so I know exactly why something worked.
3. Track
I put the test details right in the ad name: the angle, the format, who made it. So I can pull any report and see results by concept without sorting ads by hand.
3. Graduate
Before a test starts, I decide what a winner looks like. Winners get scaled, near-misses get iterated, and everything else gets cut.
HOW IT WORKS IN PRACTICE
Built for volume, Easy to read
Concept first, execution second
I test each angle with a few different creators and edits before I make the call. One bad edit shouldn’t kill a good idea.
Build on what wins
When a concept works, my next briefs are variations on it: a new hook, a new format, or different proof.
Reporting that mostly does itself
My ad names are set up so I can see spend, sales, and CPA by concept in minutes. No digging, and I can see which type of creative is actually working.
One pipeline for every source
In-house edits, UGC creators, and influencer or affiliate content all follow the same brief and naming. That keeps everything from getting messy.
How I get up to speed on your brand.
Onboarding shouldn’t take weeks. I’ve learned a lot of products fast, and this is how I get up to speed on yours.
I read your reviews, comments, and support questions to hear how customers really talk about the product. The best hooks are usually already in there.
1
2
3
4
Learn the brand
Audit the account
I look at how the account is set up and what has already won or flopped, grouped by concept so the patterns are easy to see.
Build the roadmap
You get a test plan and the first batch of briefs, with one variable per test and the rules for what scales set up front.
Launch and learn
Ads go live and I report back in plain language: what worked, what didn’t, and what we test next.
What's changing going into 2027
Diversify
Meta's own advice
What I do: brief for different ideas.
~5%
Of Meta ads are winners
What I do: plan for volume.
25.4%
Avg hook rate, Meta video
What I do: test hooks first.
61%
No meaningful AI results yet
What I do: AI for ideas, humans for taste.
*Heading into Q4: plan for 2 to 3x your normal creative cadence around BFCM.
Paid media changes constantly, so I've never stopped learning. I built my foundation in my first two years as a media buyer, then kept adding new channels, platform updates, and better testing setups. I worked up to Performance Marketing Manager, running Meta, Google, AppLovin, and CTV.