Key takeaways
- Change one variable per test, in the order the variables move results: hook, then value prop, then format, then call to action.
- Five hooks against the same body teach you more than five completely new ads.
- Run it as a weekly sprint: review, brief, ship, repeat. At Shopify that produced 100+ ad variations a quarter.
- Call to action placement was the least useful thing we tested. The lifts were marginal.
- Judge winners on what you pay for, and read them again later. About 40% of our short-term winners had no long-term effect.
01The framework in two sentences
An ad creative testing framework is a fixed routine that changes one variable per test, in a set order, and judges each test on the outcome you pay for. Mine tests the hook first, then the value prop, then the format, then the call to action, on a weekly sprint.
I built it at Shopify, where I ran $20M+ a year in paid media and creative fatigued fast. Two smaller examples further down cover offers and angles, which are a different kind of test.
02What to test, in order
| Order | Variable | What changes | What I found |
|---|---|---|---|
| 1 | Hook | The opening line or first frame | Moves results most. Five hooks on one body beat five new ads |
| 2 | Value prop | The main reason to care | Worth testing once a hook is working |
| 3 | Format | Static, video or creator-style | Comes after the message is settled |
| 4 | Call to action | Wording and placement | Least useful. Lifts were marginal |
The reason for one variable at a time is simple. If you launch five new ads and one wins, you don't know whether it was the opening, the message, the format or luck. If you launch five hooks on the same body, the buyer and the value prop hold still, and you learn which angle makes them care.
03The weekly sprint
We ran campaigns in-house and I worked with the Creative team every week.
- Monday review.I went through the previous week's performance and picked what to iterate and what to retire.
- Brief.I worked directly with designers, wrote the briefs and defined exactly what would change in each ad.
- Ship.New ads and tests went live the same week.
- Repeat.Winners got iterated and tired ads got replaced.
The volume sounds bigger than it is. Most of the 100+ variations a quarter were a winning hook shipped in five or six versions with different opening lines. Across the account, creative tests were part of 40+ structured experiments a quarter on creative, audiences and landing pages.
04What changed by channel
- TikTok. I wrote creator briefs and tested creator-style content inside performance campaigns. The testing system I built there surfaced winning hooks faster than any channel we'd run before. More on that in do TikTok ads work for B2B and SaaS.
- Reddit. The creative that worked read like a Reddit post: conversational, no loud logo, and real merchants telling their own story.
- Meta. After iOS 14.5, broader audiences meant creative did the targeting, so creative volume had to rise. That rebuild is in my note on Meta ads account structure.
05Offer angles: two smaller examples
A hook test keeps the offer fixed. Sometimes the offer is the thing to test.
Dental Team FL: 12 offer angles
At Dental Team FL, a group of ten dental practices, I tested 12 offer angles across Google and Meta. Two winners ended up driving 70% of new patient volume and cut cost per lead in half on those segments. On Meta, plain static image ads brought the lowest cost per lead, so the gain came from the offer and not from a fancier format. Health ads have rules, and some of ours couldn't mention pricing, which limited which offers we could run at all.
SkillsWave: ad angles on LinkedIn
At SkillsWave, a workforce education platform sold to HR and learning teams, the LinkedIn ads ran three angles: proving training ROI, launching learning programs in moments instead of weeks, and knowing where learning dollars go.
The useful finding was about what to test first. Ads inside the lead form campaign ran at 0.2% to 0.4% click-through. Ads inside the website-conversion campaign ran at 1.6% to 1.9%. The gap between campaign types was far wider than the gap between angles inside either one. Sort out the big structural choice before you polish angles. The dashboard I built ranked every ad by click-through, conversion rate and cost per lead, because click-through isn't the goal.
06How to judge a winner
- Use the metric you pay for. Cost per paying customer or cost per lead. Click-through tells you the hook got attention and nothing more.
- Read it again later. When we measured long enough at Shopify, about 40% of the experiments that showed a positive short-term result had no long-term effect.
- Iterate the winner before you replace it. A winning hook in five or six versions is cheaper to make than a new concept and extends what already works.
07Mistakes
- Testing whole new ads against each other. You get a winner and no lesson.
- Starting with the call to action. It's the easiest thing to change and it mattered least.
- Vague briefs. If the brief doesn't name the one thing that changes, the designer will change three.
- Calling it on week one. Short tests flattered a lot of ideas that didn't last.
Setting up a creative testing routine for an in-house team is part of what I do as a performance marketing consultant.
FAQQuestions people ask
- What is an ad creative testing framework?
- It's a fixed routine for deciding what to change in an ad, in what order, and how to judge the result. Farhan Munshi's framework at Shopify was a weekly sprint with the Creative team that changed one variable per test: hook first, then value prop, then format, then call to action. It shipped 100+ ad variations a quarter.
- What should you test first in ad creative?
- The hook, meaning the opening line or first frame. At Shopify, Farhan Munshi found that five hooks against the same body taught more than five completely new ads, because the buyer and the value prop hold still. Value prop came second, format third and call to action last. Call to action placement gave only marginal lifts.
- How many ad creatives should you test?
- Enough to stay ahead of fatigue at your spend. At Shopify, on $20M+ a year in paid media, Farhan Munshi and the Creative team shipped 100+ ad variations a quarter. Most of them were a winning hook in five or six versions with different opening lines. A smaller account needs far fewer, built the same way.
- How do you know if an ad creative test won?
- Judge it on the outcome you pay for and read it again later. Click-through is an early signal. At Shopify, Farhan Munshi found that about 40% of experiments with a positive short-term result had no long-term effect once they were measured long enough, so winners were checked against paying merchants and not only against early clicks.
- How do you test offers in ads?
- Treat the offer as its own variable and test angles against each other on the same audience. At Dental Team FL, a group of ten dental practices, Farhan Munshi tested 12 offer angles across Google and Meta. Two winners ended up driving 70% of new patient volume and cut cost per lead in half on those segments.
