On 17 August 2026, Google changed how Target CPA and Target ROAS behave in campaigns that are limited by budget. The change is already live. If your cost per acquisition has drifted upward over the past two weeks and nothing in your account changed, this is almost certainly why - and the fix most advertisers reach for first is the wrong one.
What Google Actually Changed
Before 17 August, a campaign constrained by its budget behaved conservatively. Facing a daily budget it could exhaust quickly, Smart Bidding concentrated spend on the cheapest auctions available to it. The practical result was familiar to anyone managing a small account: you set a Target CPA of €20, and the campaign quietly delivered at €11, month after month.
Google's framing of the change is direct. Budget-limited campaigns using a target-based strategy will now perform more consistently toward the bid target you set. The documentation gives the example plainly: a campaign with a Target CPA of $10 that has been delivering an actual CPA of $5 will begin delivering closer to $10.
The change applies to Search, Shopping, Performance Max, Demand Gen, Display, Hotel and Travel campaigns, and to Search Ads 360 as well. App campaigns, Video reach campaigns and Video view campaigns continue with the previous behaviour. There is no regional carve-out. The rollout is global, which means Greek accounts moved on exactly the same date as accounts in the United States or Germany.
What It Does Not Do
The most common misreading circulating right now is that Google has found a way to make you spend more. It has not, and the documentation is unambiguous: the change will not directly result in increased spend, and daily and monthly budget limits continue to be respected in full.
What changes is not how much is spent but where inside the budget it goes. Instead of harvesting only the cheapest inventory it can reach, the bidding system now distributes the same budget across the full range of auctions available up to your stated target. You are not paying more. You are paying differently, and reaching a broader set of opportunities in the process.
Whether that is good or bad for your business is a real question. It is also a different question from the one most people are asking.
The Uncomfortable Thing This Exposes
Here is what the change actually surfaces. If your Target CPA was €20 and the campaign delivered at €11 for eighteen months, then the €20 was never doing any work. It was a number typed into a field once and never revisited, sitting far above anything the campaign was allowed to attempt. The budget constraint, not the target, was governing the account.
The target was never tested. It was a ceiling so high the campaign never approached it. Google has now made that number mean something, and a great many accounts are about to discover that the figure in the target field does not correspond to any business decision anyone consciously made.
So the useful question is not what Google is doing to your CPA. It is this: what is a conversion actually worth to this business, and does the number currently sitting in the target field resemble that at all?
Why This Lands Harder in the Greek Market
In markets dominated by large budgets, being limited by budget is an occasional condition - something you notice, correct, and move past. In Greece it is closer to a permanent state, and that makes this change disproportionately significant here.
The State of Balkan eCommerce 2025 survey of Greek online merchants puts monthly advertising budgets in stark terms. Thirty-four percent of Greek eCommerce businesses spend up to €1,000 per month on advertising, and a further 24.5 percent spend between €1,000 and €5,000. Close to six in ten Greek eShops operate below €5,000 a month. The same survey ranks Google Ads as the clear number one channel at 41.5 percent of respondents.
A campaign running on €1,000 a month is working with roughly €33 a day. At that level, limited by budget is not a status you occasionally encounter. It is the condition the account lives in permanently.
Which leads to an uncomfortable implication. The comfortable CPA many Greek advertisers have been reporting was, in part, a product of the constraint rather than of the setup. It was not earned by better structure or better creative. It was subsidised by the fact that the campaign could never reach far enough to test its own target. That subsidy has now been withdrawn, and it has been withdrawn across a far larger share of Greek accounts than of accounts in higher-budget markets.
The Trap: Lowering the Target on Reflex
The instinctive response is to drop the target until the old CPA returns. Target was €20, delivery was €11, so set the target to €11 and the problem disappears.
Sometimes that is correct. Often it is not, and it fails in a specific and predictable way.
A target set at or below what the account can realistically deliver does not produce cheap conversions. It produces a campaign that cannot find enough qualifying auctions to spend its budget at all. Impression share falls, volume drops, and the campaign moves from limited by budget to a bid strategy that is throttling itself. You have locked in an efficiency number and paid for it with the volume that made the account matter in the first place.
The decision underneath is not a Google Ads decision. It is whether €11 per acquisition at low volume is worth more to the business than €20 per acquisition at meaningfully higher volume. That depends on gross margin, on repeat purchase rate, on lifetime value, and on whether the business can service the additional volume at all. Until 17 August, the platform was answering that question silently on your behalf. It has stopped.
What to Do Now
- Filter, do not guess. In the campaigns view, isolate campaigns whose status is limited by budget and whose bid strategy is Target CPA or Target ROAS. Those are the only campaigns in scope. Everything else can be left alone.
- Compare target against reality over 30 days, not 7. For each campaign in scope, put the target you set next to the actual CPA or ROAS for the previous 30 days. The gap between them is your exposure. A campaign already delivering close to its target has nothing to do here.
- Make the business decision before touching the field. Where the gap is large, choose deliberately: reset the target to reflect the performance you actually want to keep, or hold the target and raise the budget to buy the additional volume it now unlocks. Both are defensible. Doing neither is not.
- Use the Bid Target Adjustment Tool. Google shipped a dedicated tool for exactly this review and it is live in the interface. It will not make the decision for you, but it removes the spreadsheet work of identifying which campaigns are exposed.
- Then wait. Google's guidance is to allow one to two conversion cycles before evaluating the bid strategy report. Judging a target change on three days of data is how accounts enter a spiral of adjustments that never settles.
One Caveat Before You Reset Anything
If you are about to anchor a new target to the CPA in your dashboard, be certain that number is real. Platform-reported CPA carries the attribution model's assumptions inside it, and in most accounts those assumptions flatter the channel. Resetting a target to a CPA that was inflated by over-attribution relocates the error rather than removing it.
Where you have incrementality data, blended figures, or a margin-based target of your own, use those. Where you do not, at minimum hold the reporting window constant so you are comparing like with like across the change.
The Pattern Underneath
This change belongs to a longer sequence. Over several years Google has steadily removed the accidental efficiencies that came from constraint - the cheap conversions that arrived not because an account was well built but because the system was boxed in. Each removal converts a setting that used to be decorative into a setting that governs outcomes.
Accounts where targets reflect real unit economics will absorb 17 August with a modest adjustment, and some will find volume they did not know was available to them. Accounts where the targets were placeholders will watch performance move toward numbers nobody chose on purpose.
The change is not a threat. It is an audit, and it does not particularly care whether you were ready for it.