Why your Google Ads CPA rose: the August 2026 Target CPA change
On 17 August, Google changed how target-based bidding works, and on many accounts actual CPA climbed without anyone touching a campaign. What changed, who it hit, and how to tell that rise from a real one.
Adgent · Notes from building the analyst
Most Google Ads changes are quiet: a label moves, a report gets a new column, nobody notices. This one was not. On 17 August 2026, Google changed the mechanics of target-based bidding in a way that lifted actual CPA on many accounts, whether or not anyone touched a setting. If your CPA jumped in late August, this is very likely why. Your campaign did not break. The rules changed underneath it.
The frustrating part is that the accounts most affected are often the well-run ones: the ones that were quietly beating their targets. So before you spend a morning hunting for a broken pixel, understand what actually moved.
What changed on 17 August
Google's own framing is that target-based bid strategies make budget-limited campaigns perform more consistently toward the target you set, including after budget adjustments. Before the 17th, a campaign constrained by budget could drift well under its target, because Google was not pushing to fill every dollar toward that goal. Now it does. The strategy steers delivery back up to the target you have told it to hit.
The clearest way to see it is Google's own example:
“If your Target CPA is $10 but recent actual CPA is $5, after August 17 the campaign will deliver closer to a $10 actual CPA if you make no changes.”
Read that again, because the direction matters. A campaign acquiring conversions at $5 against a $10 target was not underperforming. It was outperforming, comfortably. The change told Google to stop leaving that room on the table and spend up toward the $10 you set. Your acquisition volume may rise, but so does the price you pay per acquisition, and so does total spend. The number you set as a ceiling became, in effect, the number Google aims for.
Alongside the mechanics, Google renamed the strategies. “Maximize conversions with a Target CPA” becomes simply “Target CPA”, and “Maximize conversion value with a Target ROAS” becomes “Target ROAS”. Same underlying strategy, cleaner label. The rename is still a good reminder to go and look at what each target is actually set to.
| Change date | 17 August 2026 |
|---|---|
| Bid strategies | Target CPA, Target ROAS |
| Campaign types | Search, Shopping, Performance Max, Demand Gen, Display, Hotel, Travel |
| Hit hardest | Budget-limited campaigns beating their target |
| Target review tool | Bid Target Adjustment Tool, since 6 July 2026 |
Who this hits hardest
The single biggest tell is the gap between your target and your real delivery. If your actual CPA has been sitting comfortably below your Target CPA, or your actual ROAS comfortably above your Target ROAS, you had been leaving delivery on the table, and the change is designed to fill exactly that gap. The bigger the gap, the bigger the move you saw.
Budget-limited campaigns feel it most sharply. They are the ones where Google now has explicit room to spend up toward the target after adjustments, so the effect concentrates there. If you run tight daily budgets on Search, Shopping, Performance Max or Demand Gen with a target-based strategy, assume you are in scope. The change applies across Search, Shopping, Performance Max and Demand Gen, and also Display, Hotel and Travel campaigns using Target CPA or Target ROAS.
Accounts already delivering right at their target will not see much: there is no gap for Google to close. It is the outperformers, the accounts that set a target with headroom and then beat it, who got the surprise.
Why your CPA looked like it jumped, though nothing broke
Here is the diagnostic trap. Someone opens the account in late August, sees CPA up thirty or forty percent, and starts pulling levers: pausing campaigns, cutting budgets, blaming the creative. None of that is the right response, because none of that is the cause. This is a settings and system change on Google's side, not a campaign problem. The funnel did not break. The auction did not turn. Google changed what your target means.
That distinction is the whole game. A CPA rise that comes from the auction, a tracking break or creative fatigue demands one kind of response, and a rise that comes from a bidding-rule change demands a completely different one. If you cannot tell them apart, you will react to the wrong thing. For the general version of that diagnosis, the usual causes of a CPA spike and how to rank them, see why did my CPA suddenly jump. This change is one more cause worth adding to that list for the second half of 2026: the target itself moved.
What to do now
The change went live on 17 August 2026, so the checklist below is no longer preparation. It is a catch-up audit. The window to prepare has closed; the tool has not. Google's Bid Target Adjustment Tool, available since 6 July 2026, still lets you review and reset targets by hand.
- Audit target against true break-even. Compare each Target CPA or Target ROAS against the most you can actually afford to pay, not your comfortable recent actual. If your target was set with slack, that slack has already become spend. The break-even ROAS calculator gives you the floor.
- Use the Bid Target Adjustment Tool. If you do not want the higher CPA, lower your Target CPA (or raise your Target ROAS) to your real ceiling now, rather than leaving Google to spend up to a loose target.
- Flag your budget-limited campaigns. These move most. Know which campaigns are capped: that is where the CPA rise concentrated.
- Do not panic-edit. If CPA rose after the 17th, resist reworking bids and budgets in a hurry. Big edits reset learning and cost you more. If you do need to scale, do it deliberately: scale a winner without breaking it.
The one decision that matters is whether your target is a real ceiling or a loose placeholder. If it is your genuine break-even, leave it: the change is bringing delivery up to a number you can live with. If it was set once and forgotten, fix it, because Google is already acting on it.
Reading the rise before reacting to it
A number moving because Google changed the rules is exactly the kind of thing that should be explained before anyone touches the account. A CPA up in late August is not automatically a fire. If it lines up with the 17th, your campaigns are target-based, and actual CPA moved toward your target rather than past it, the cause is this change.
The same discipline runs the other way: a target only means something if the account can reach it. On one live account, Adgent was asked to set a target CPA the numbers could not support:
| Target CPA requested | ₺250 ($5.16) |
|---|---|
| Actual cost per booking | about ₺4,028 ($83) |
| What it did | declined the target and explained why |
A target set far from what the account delivers, in either direction, is an instruction the bidding system will follow literally. That is the work the Trust Gate exists for: checking a request against the account's own numbers before the budget moves.
Adgent reads every connected Google Ads and Meta account each day, carries the brand and confirmed targets into the analysis, and puts an estimated monthly amount on each decision. The target gap, evidence and proposed move arrive together; after the account changes, the result returns to the same history. If you want that read on your own targets now the change has landed, start with the audit.
Google's own announcement of the change is in the Google Ads Help Center.
Bring the same question to your own account. Adgent returns its position, the evidence underneath and what the finding is worth a month.
Pricing