Case study
A large but fragmented account standing in the way of its own return. By consolidating the structure and cleaning up Performance Max, ROAS nearly doubled the month after: the best month in three years. And no, that was not seasonality; that check is spelled out below.
nearly 2x
return (ROAS) in the month after the cleanup
best in 3 years
highest monthly ROAS in the full 3 years of data
+83%
month-over-month jump, against +17% a year earlier
Challenge
A large catalog, a sizable budget, and yet a return that kept stalling. The cause wasn't the market but the setup: too many separate campaigns, budget in structure that barely contributed anymore, and a Performance Max set so broad that nobody could see where the money went.
At an account of this size, that's expensive. Every campaign that gets too little data steers worse; every euro in a weak segment is a euro your strong segments miss out on. The account wasn't performing badly, but far below what the catalog could support.
Approach
The play was consolidation. Separate and overlapping campaigns were wound down and brought together in one clear structure, so the conversion signal no longer splintered across dozens of places but pooled where the bidding steers on it. Performance Max was cleaned up and segmented, making visible which part of the catalog carried the results and which part swallowed budget without return.
Budget then shifted to the segments that demonstrably perform. No new campaigns, no higher budget; just an account doing what it's good at, instead of spreading its strength across noise.
Result
The month after the cleanup, ROAS nearly doubled compared to the month before. It was the highest monthly ROAS in the full three years of data we have, while the budget didn't go up. The same account, the same catalog, just set up more sharply.
The fair counter-question to a jump like this. Hence the seasonality check: the same month-to-month transition, three years in a row. A year earlier, ROAS rose +17% in that same transition; the year before that it actually fell -13%. This year the jump was +83%, more than four times the best seasonal effect of the preceding years. The season explains part of it, as always; it does not explain this jump. And ROAS held its level in the months that followed, so it was no one-off spike.
Best practice
The reflex when ROAS disappoints is to create something new: an extra campaign, an extra segment, another test. But automated bidding runs on data, and every new campaign dilutes the signal. An account with too many campaigns learns a little everywhere and nowhere well.
Consolidating feels counterintuitive, because you're switching things off. Yet it's often the biggest lever: pool the signal, make visible what carries the results, and let budget follow the data instead of habit. Fewer moving parts, more control.
Self-check
The more campaigns you open, the thinner the conversion signal per campaign gets. Automated bidding needs data to steer well; fragmentation starves it.
Campaigns are running because they were once created, not because they pay off. Every euro there is a euro that doesn't go to your strong segments.
One big PMax campaign across your entire catalog hides where the money goes. Without segmentation you can't tell good from bad, let alone adjust course.
If your brand traffic sits in the same pot as everything else, your ROAS looks better than it is. You would have gotten that brand traffic anyway; it distorts what your ads really add.
Consolidation is not a goal in itself; sometimes splitting is exactly right. The point is that structure should follow the data, not your history. An account clogged with old campaigns almost always leaves return on the table.
In a free account scan I'll show you where your structure gets in the way of your return, and what's freed up when you consolidate it.
Findings within 5 working days. If there is little to gain, I'll tell you that too.
