Google Ads Changed Target ROAS / CPA Mechanism – 17 August 2026: What Changed and How to Protect Your Campaigns

On 17 August 2026, Google Ads changed the way its bidding systems treat budget-limited campaigns running Target CPA and Target ROAS (and Target CPC for Demand Gen). If your account uses target-based bidding, this is one of the more consequential Smart Bidding changes of the year not because the auction changed, but because campaigns that had quietly been beating their targets no longer do so by default.

The deadline has now passed, and the change is live. But because Google rolled it out gradually across accounts worldwide, plenty of advertisers still haven’t seen the full impact land which means there is still time to review your setup and act before your numbers move on their own. Here’s a clear breakdown of what happened and what we recommend doing about it.

Before this update, a campaign that was “Limited by budget” and running a target-based strategy could comfortably outperform its stated target. A campaign with a Target CPA of $100 might have been delivering conversions at $70; a campaign with a Target ROAS of 500% might have been landing at 550%. The budget cap did the constraining, and Smart Bidding cherry-picked the most efficient auctions inside that cap.

Google reclassified that behaviour as unintended. From 17 August, budget-limited campaigns optimise more literally toward the target you actually entered, regardless of whether extra budget becomes available. In practice that means a campaign whose real performance sits well inside its stated target will tend to drift toward that target over time.

The mechanics are worth stating plainly, because a lot of the online commentary blurred them:

• The auction itself did not change. This was a bidding-behaviour change only.
• Google is not automatically adjusting anyone’s targets or budgets. The responsibility to review and reset sits with the advertiser.
• The rollout is gradual and global, so the new behaviour may not appear in every account immediately.

In scope: budget-limited campaigns using Target CPA or Target ROAS across Search, Shopping, Performance Max, Demand Gen and Travel including Search Ads 360 and Demand Gen in Display & Video 360.

Not affected: campaigns that are not limited by budget, plus App, Video Reach and Video View campaigns. Hotel and Display campaigns are already optimised this way.

If a campaign only occasionally flickered into “Limited by budget” status, it’s far less exposed than one that has been budget-constrained for weeks or months.

Google’s stated goal is predictability. Previously, raising the daily budget on a budget-limited, target-based campaign produced inconsistent results; some campaigns held their efficiency, others swung noticeably as Smart Bidding absorbed the extra spend. That made it hard to forecast what would happen when you scaled.

Under the new behaviour, a campaign with a $10 Target CPA should keep optimising toward $10 whether or not its budget changes. The trade-off Google is offering is scalability: because performance now tracks the stated target more consistently, you can increase budget to capture more volume without efficiency lurching unexpectedly.

That’s a reasonable design goal. The friction came from the fact that many advertisers had, quite deliberately, been using their targets as one lever among several inside a fixed budget, not as an exact instruction. For those accounts, 17 August introduced a decision that hadn’t existed before: does the number in the target field still represent the outcome you actually want?

The date has passed, but the audit hasn’t lost its value. The drift shows up quietly over the coming weeks, so a quick review now still pays off. Keep it simple:

1. Find your exposed campaigns: budget-limited campaigns running Target CPA or Target ROAS.
2. Reset the target close to your recent actual performance, with ~10% of breathing room: CPA $70 → ~$77, ROAS 500% → ~450%.
3. Monitor CPA, ROAS and spend over the next few weeks as the rollout reaches your account.

This wasn’t an algorithm overhaul, it was a settings problem with a real cost attached. Campaigns that were quietly overdelivering will, left untouched, deliver closer to whatever target is on file. The advertisers who came through it best were the ones whose targets already reflected intentional decisions, and the ones who audited and reset before their numbers moved for them.

If you haven’t yet reviewed your Target CPA and Target ROAS campaigns against the change, now is the time the impact builds quietly, and the fix is far cheaper before the drift shows up in your results than after.