A 4.5x tROAS target was throttling a profitable account. Lowering it added 71% revenue

A target ROAS tells Google the efficiency floor you'll accept. The higher you set it, the fewer auctions qualify, because the bidder only enters auctions it thinks will hit that return. That's fine when the target sits near break-even. But when your target is more than double your break-even, you're not being disciplined, you're refusing profitable volume. Every auction that would have returned 3x gets skipped because you demanded 4.5x.

This account had a 4.5x target and a 2.2x break-even. The gap between those two numbers was a second business it was choosing not to buy.

What happens as you lower the target

Profit is revenue x margin minus spend. Lower target, more spend, lower ROAS, more total profit. That holds right up until the marginal ROAS hits break-even.

The discipline that made it safe

1. 1

   Established the true break-even first

   One divided by gross margin gave a 2.2x floor. Everything else is measured against that number, not against a gut feeling about what ROAS looks healthy.

2. 2

   Dropped the target in 10 to 15% steps

   Never in one big move. A sudden cut throws the bid strategy back into learning and it buys low-quality volume while it recalibrates. Small steps let it re-stabilise at each level so you scale on quality.

3. 3

   Waited for each step to settle

   One to two weeks between moves so the algorithm could find its footing before the next drop. Patience here is the entire skill.

4. 4

   Watched the marginal ROAS, not the average

   At each level we checked the return on the new spend. As long as that marginal number stayed comfortably above 2.2x, we kept going.

5. 5

   Uncapped budgets on the winners

   A portfolio bid strategy with budgets uncapped on the campaigns that kept clearing the floor, so the bidder could take all the profitable volume the lower target now qualified for.

6. 6

   Held the line on noisy weeks

   Resisted yanking the target back up the first slow week. Day-to-day noise isn't signal. We only reacted to sustained trends in the marginal number.

Safe versus scaled

Before and after

ROAS dropped on purpose. Revenue and profit both grew because every extra dollar still cleared break-even.

Questions we get asked about this

Frameworks behind this case study