I've spent about eight years on Amazon PPC and published somewhere north of 10,0
By Michael Erickson Facchin · September 7, 2026 · Curated by George's Blog
I've spent about eight years on Amazon PPC and published somewhere north of 10,000 hours of content about it.
In that time I've found roughly one original problem.
Everything else is the same four things, in the same order, in almost every account I open.
ONE. The target is borrowed, not calculated.
Someone heard 25% was good. Or they saw the platform average, which is currently 29.6%, and adopted it.
But break-even ACoS is just your profit margin. A $20 product with $3 in Amazon fees and $6 in COGS leaves $11 of pre-ad profit, which means break-even is 55%.
A seller in that position throttling at 25% is leaving thirty points of profitable headroom on the table and handing those placements to whoever did the arithmetic.
The reverse is worse and more common. A 22% margin chasing a 30% benchmark, losing money on every incremental order, while the dashboard looks completely normal.
TWO. The account structure grew instead of being designed.
Campaigns accumulate. Nobody deletes anything. Eventually you have a structure where no campaign has a single clear job, and once that's true, no metric on it is interpretable.
You cannot optimize what you cannot isolate. That's not a philosophy, it's just true.
THREE. The dead search terms were never cleared.
Typically 15 to 25 percent of spend sits in queries that barely convert. And it's never one obvious disaster keyword, because most search terms get fewer than five clicks. It's hundreds of small ones, each looking like a rounding error, collectively a five figure annual leak.
FOUR. The decision cycle is slower than the auction.
Monday morning bulk file. Adjust two hundred rows out of four thousand. Upload. That's 168 hours of live auction running on data from last week.
Winners get starved. Losers get funded. And normal weekly fluctuation gets mistaken for a trend, so people cut a bid 40% on twelve clicks of data and reverse it seven days later.
I've watched a lot of people try to fix number four first, because it's the one you can buy.
It doesn't work. Automating bids against a wrong target, inside a broken structure, full of dead search terms, gets you to the wrong answer faster and more consistently.
That's the actual reason behind most "we tried a PPC tool once and it didn't help" stories. The tool did exactly what it was told. Nobody had checked what it was being told.
One account came to us at 117.90% ACoS. A month later it was under 50%, on roughly a third of the spend, with ROAS up 133% and average CPC down $1.79.
Nothing clever happened. We just did the four things in order.
Which one is yours? Be honest, it's usually number one.