10 Comments
User's avatar
Gabriel Sandru's avatar

The denominator is where this gets me. Share against total Meta spend hides a lot when retargeting is carrying a chunk of the account: the retargeting line inflates the base, the ratio looks healthy, and prospecting is the part actually starving for new ads. I run it against prospecting spend only, since that is the budget the test is supposed to feed. Same 20% ceiling, very different answer on accounts with heavy retargeting.

Przemek Jurek's avatar

Judging tests by new-customer cost is the key point. Small brands often count repeat buyers inside test results, so a weak ad looks fine. Splitting new and returning customers is a simple fix that many owners skip.

Matt Anderson's avatar

The 26% example is a useful gut check. For a small business spending a few thousand a month on Meta, a test under $1,000 can be a big slice of the budget, so the 10 to 15% rule pushes you towards fewer, bolder tests rather than lots of tiny ones. I'd also count a new offer as a test, even when the creative is old. How do you handle accounts too small to get a test to meaningful spend inside 14 days?

Curtis Howland's avatar

That’s a great question. I do primarily work with nine-figure brands, but we’ve launched a number of new brands. In that case, we really do focus on quality over quantity, and we don’t let spend per ad launch below $1,000.

Matt Anderson's avatar

That makes sense. For a lot of Australian small businesses $1,000 per launch is most of a month’s budget, so quality over quantity is the right lesson for them too. Thanks Curtis.

Kedar Garg's avatar

Does the 15% include new landing pages, or only new creative? When a test ad sends people to a new page too, a losing CPA could be the page and not the ad. I'd guess that's part of why one producer can sit at 50% in testing.

Curtis Howland's avatar

I would include all tests, including landing pages. Yes.

Unless there’s a particular reason to exclude, all tests have a cost associated with them.

Kedar Garg's avatar

That makes the 15% bigger than it sounds. A new page usually needs design and dev time, not just a new ad. Do you count that time in it, or only the media spend?

Curtis Howland's avatar

Fair point - count it in time and just maek less big tests.

Measure tests on

Difficulty/ease

Impact (high low)

Confidence

So high difficulty tests only execute if impact and confidence are also high.

Kedar Garg's avatar

That keeps the page tests honest. A new landing page always looks high impact on paper, so confidence is the score I'd be strictest on. If you can't say why the old page lost, the new one is a guess.