I'll be honest. I've seen scale tests that proved absolutely nothing.
By Elizabeth Greene · September 12, 2026 · Curated by George's Blog
I'll be honest. I've seen scale tests that proved absolutely nothing.
A brand came to us.
Their agency said they needed to spend more.
They were sitting at 12% total ACoS, didn't want to go higher, but couldn't grow.
The agency's answer was just more budget.
So we audited the account.
What we found was that the spend was overwhelmingly concentrated on branded keywords.
The "scale test" had basically just pumped more money into people who were already searching for them by name.
Amazon happily took the extra budget. Revenue barely moved.
The campaign labels said "non-branded."
But when we actually looked at what was being targeted inside those campaigns?
Branded ASINs and branded keywords, defense spend wearing a growth costume.
I can't tell you how many times I've gotten into an account and it's just a mess.
Campaigns labeled one thing, doing something completely different.
And then someone runs a "scale test" on top of that and goes, "See? We're maxed out. There's no more room to grow."
No. You tested whether Amazon would let you pay more for demand you already owned. It said yes.
That's a toll booth on a road people were already driving.
If your test doesn't separate branded from non-branded at the keyword and ASIN level, not the campaign label level, you're just testing how much branded traffic costs when you bid higher on it.
And the answer is always more.