An account that looked fine on the surface. We questioned the campaign structure anyway.
More Sales (Peak)
Cost Per Sale
Additional Sales
Ecommerce
Google Ads
additional sales.
11% more spend.
An ecommerce retailer’s Google Ads account was doing fine. Steady spend, steady conversions, nothing on fire. The kind of account most agencies would leave alone.
We didn’t.
The account had been running for years. We’d built most of it. The structure was sound, and the top-line numbers still looked healthy. But during a quarterly performance review, we dug into where the money was actually going inside two campaigns. What we found changed the direction of the account.
A Performance Max campaign spending $103,000 over eight months.
1,331 conversions at $78 each. Sounds fine, until you look at where the money actually went. PMax was dumping product budget into Display and YouTube impressions that generated clicks but almost never turned into sales. Google’s algorithm was optimizing for reach, not revenue.
A Dynamic Remarketing campaign: $20,000 spent. 32 sales. $630 per conversion.
Over the previous two quarters, we’d watched the conversion rate on this campaign decline steadily. The audience had fatigued. The same users seeing the same ads, converting less each month. The trend was clear: this campaign had run its course.
We ran an asset performance breakdown and the data was clear: Shopping placements inside PMax were converting. Everything else was tourism. Clicks that cost money, looked good in a report, and never turned into a sale.
We built both of these campaigns. They worked when we built them. The job is knowing when the data tells you to tear it down and build something better, even when you’re the one who built it.
Replaced Performance Max with a standard Shopping campaign. The difference: full control. We choose which products show, what we bid, and where the ads appear. Every impression goes to someone actively searching to buy, not someone scrolling through a YouTube video.
The product feed stayed the same. The targeting stayed the same. The only thing that changed was who made the decisions: us, or Google’s algorithm. The results made the case.
Didn’t replace it. Redirected every dollar into the Shopping and Search campaigns that were already the most efficient parts of the account. The logic was simple: if a campaign is converting at $45 per sale and another is converting at $630, every dollar in the second campaign is a dollar stolen from the first one.
This was a reallocation. The total budget went up 11%, but the money moved. Instead of $103,000 funding Google’s display network and $20,000 chasing users who’d already decided not to buy, that spend went into the campaigns that were already producing. More fuel into the engine that was already running.
Ecommerce
Google Ads
additional sales.
11% more spend.
Shopping placements inside PMax were converting. Everything else was tourism. Clicks that cost money, looked good in a report, and never turned into a sale.
First full peak season under the new structure:
| Metric | Before | After | Change |
|---|---|---|---|
| Conversions | 2,527 | 3,819 | +51% |
| Cost | $154,522 | $171,953 | +11% |
| Cost Per Sale | $61 | $45 | –26% |
| Clicks | 160,532 | 205,488 | +28% |
11% more spend. 51% more sales. Nearly 1,300 additional conversions during the four months that drive the business.
Then the off-season tested it from the other direction. Spend dropped 24%. Conversions held steady. Cost per sale still improved 13%. The structure performed at every volume level.
additional sales.
11% more spend.
A campaign producing 1,331 conversions doesn’t look like a problem. That’s why it’s the most dangerous kind of problem. Nobody questions it. Nobody opens it up. It just keeps running and keeps looking “fine” in a dashboard while better performance sits on the other side of a decision nobody’s willing to make.
The instinct is to leave working campaigns alone. But “working” and “working as well as it could be” aren’t the same thing. The gap between those two is where the money is. In this case, that gap was 1,300 sales over four months.
The accounts that plateau aren’t broken. They’re just managed by people who stopped asking questions.
additional sales.
11% more spend.
If the answer is “I don’t know,” that’s the answer. We’ll take a look.