Most agencies talk about creative testing like it's a production problem. Test more hooks, try a UGC cut, swap the thumbnail, run five formats and let the algorithm sort it out.
I think that's backwards. Probably because I came up as a copywriter, not a media buyer or a video editor, so before anyone touches a camera or a template, I want to test something cheaper: the angle.
The angle is the actual argument the ad is making, the specific reason someone should call you instead of scrolling past. "Board-certified surgeon" is not an angle. "You've gotten three different prices from three different clinics because none of them will tell you the real number until you're in the chair" is an angle. One's a fact about the business. The other is a reason to act, and it's in the prospect's own words, not the marketer's.
Here's how I actually find them.
Pull angles from sales calls, not brainstorms
I used to run these as team brainstorms. Whiteboard, sticky notes, the whole thing. It rarely produced anything a prospect would recognize as their own problem, because it was our language, not theirs.
What works better: go through the last 10-15 sales call recordings and write down every complaint, hesitation, and objection word for word. Not your summary of it, their actual sentence. "I don't want to sign a 12-month contract with someone I just met" is an angle. So is "every quote I've gotten so far felt like it was designed to be negotiated down."
No recordings? Whoever answers the phone knows this cold. Ask them what people push back on before they'll book.
Sales calls aren't the only source, though. Two more I trust just as much:
Your hunch counts. If you're tuned in to your audience, and you have a feeling about what's going to work, put it out there. That feeling isn't nothing, it's pattern recognition from actually paying attention to these people. Run it and see what the data says. You don't need a sales call transcript to justify every angle you test, you need enough sales calls, over enough time, that your gut has actually learned something.
If a competitor's creative has been running for a while, rip it off entirely. Not the execution, the angle underneath it. An ad that's still live after months isn't a coincidence, it's still live because it's working. Somebody already paid to find out that argument converts. Let them.
Test the angle before you test the ad
Once you've got a handful of candidates, don't go straight to production. Say the angle out loud, plainly, no visuals, no music, no editing. Just the sentence.
If it still lands when it's flat and unpolished, it's real. If it only sounds convincing once it's wrapped in a nice edit and a trending audio track, that's not an angle, that's a vibe, and vibes tend to fall apart in front of a skeptical buyer.
This is the step most people skip, which is also why so much ad budget goes into beautifully produced creative that quietly loses to a $50 talking-head video shot in someone's office. The cheap one had a real angle. The expensive one didn't.
Only then do you test execution
This is where everyone else starts, and it's honestly the least important part. Once an angle is validated, test it five different ways: founder straight to camera, a customer telling the same story, a static ad with bold text, a short explainer. A good angle survives all of those. A weak one doesn't get better no matter how you shoot it.
Order matters here:
- Pull 5-8 candidate angles from real conversations, your own gut, or what's already working for a competitor.
- Say each one plainly. Cut anything that only works with production behind it.
- Test execution variants of the 2-3 survivors, not the whole list.
Most of the money in creative testing gets spent on step 3 by people who skipped 1 and 2. That's an expensive way to find out an ad was never going to work.
If you want a second opinion on the angles you're already running, that's the first thing I'd look at on a strategy call, not the creative. The argument underneath it.