Stuck on Creative? Change the Test Axis, Not the Details

Minh watched a breakdown of how one media buying team found a new winning ad after weeks of getting nowhere, and the takeaway is worth writing down: an account stuck on creative is usually not testing too little, it is testing the same thing over and over.

1) A creative rut is a testing-axis problem, not a volume problem

The account in the video kept shipping new ads, but each one differed only in small details, so Facebook had no reason to move budget. Once the team changed four things at once — a new customer avatar, a new creative format, a curiosity-driven text hook, and an unusual filming location — the new ad took over spend and cost per purchase fell from $98 to $51, then to $45. Those are the video's numbers, not a benchmark.

2) Change big at the idea level, change one variable at the variant level

Each idea shipped as three videos that were otherwise identical: same script, same voiceover, same text hook, and a different first three to five seconds of visual. That is the only way to know why a variant won, and why a kill threshold set before launch reads cleanly. Finding a new direction and reading a clean result are two different jobs, so they do not belong in the same test.

3) One idea, one new ad set inside the same CBO campaign

Ideas get numbered and launched as new ad sets inside a single CBO campaign so they compete for budget. When one wins, the campaign budget goes up 20% a day for as long as cost per purchase stays inside target, and scaling stops the moment it does not — the same discipline that keeps scaling from breaking profitability.

How Minh reads this from experience: during a flat stretch on the Shopee Ads and Facebook Messenger accounts for Cua Di, reworking images, captions and targeting moved nothing until the customer avatar itself changed.

The point: an account stuck on creative does not need more ads, it needs a testing axis different from everything it ran last month.