Someone came to us with a full analysis.
By Elizabeth Greene · September 3, 2026 · Curated by George's Blog
Someone came to us with a full analysis.
Like, a real one, with charts and data.
And the conclusion was: we're losing top of search.
That's the problem. Here's the strategy to fix it.
So we looked at it.
The analysis wasn't wrong.
They were losing top of search.
But that was the symptom.
What actually happened was a new competitor had entered the space.
CPCs shifted. Conversion rate dipped on a couple of key products.
And because organic rank started slipping on those products, TACoS spiked.
But the brand had already decided it was a placement issue.
So the strategy they wanted us to execute was basically:
segment everything, crank up top of search modifiers, get back to where we were.
That would have destroyed their profitability.
When ACoS spikes or rankings drop, the question is what changed.
ACoS is a division problem.
Spend divided by sales. So either spend went up, or sales went down.
If spend went up, was it CPCs or clicks?
If sales went down, is it conversion or volume?
You keep pulling the thread until you find the actual campaign and the actual product.
In this case, it was two products.
And the root cause was competitive pressure changing the economics of a few high-volume keywords.
If you're earlier in your career managing Amazon ads, this is the thing I want you to internalize: very rarely is it every single product and every single campaign that stopped working.
It's almost always a couple of products and a few specific campaigns, with a root cause that has nothing to do with what the dashboard makes you think.
I wouldn't go first to the placement problem.
I'd go first to the date of change. Pull a comparison.
Find the product with the biggest sales drop.
Find the campaign driving that drop. Then figure out why.
Once we identified the actual competitive shift on that brand, we adjusted bids on the affected keywords, tightened up the ranking campaigns, and stopped chasing a placement metric that was never the lever.