Some Google Ads accounts have a target CPA that means “we would tolerate this.” Starting August 17, Google plans to treat it more like “please aim here.”
That difference can get expensive.
Google is changing target-based bidding for campaigns marked Limited by budget. According to its official change notice, affected campaigns that have been beating their stated Target CPA or Target ROAS may begin delivering closer to the target after August 17, 2026.
Google gives a blunt example. A campaign with a $10 Target CPA that has recently delivered a $5 CPA may start moving closer to $10. The budget cap still applies, but the efficiency can change inside that cap.
A bid target is no longer a motivational poster. It is an operating instruction.
Find the campaigns that are actually affected
Do not turn this into a full-account panic. The change is narrower than the headline.
Start with campaigns that meet both conditions:
- They use Target CPA, Target ROAS, or Target CPC for Demand Gen.
- They are currently, or have recently been, limited by budget.
The change applies across Search, Shopping, Performance Max, Demand Gen, and Travel campaigns. Google says Display and Hotel campaigns already use the new behavior. Manual CPC and Target Impression Share are not part of this update.
If you received a service announcement but see no recommendation, keep looking. Google says the notice can reflect any affected campaign that was budget-limited during the previous 12 months, while recommendations use more recent data.
Replace the target with a business number
A platform target should come from unit economics, not from the number somebody typed during setup.
For lead generation, work backward:
- What percentage of tracked leads become qualified?
- What percentage of qualified leads become customers?
- What is a new customer worth after delivery costs?
- How much of that value can acquisition consume?
Suppose a professional-services firm closes 20 percent of qualified leads, only half of its tracked conversions are qualified, and it can spend $1,000 to acquire a client. The rough ceiling is $100 per tracked conversion: $1,000 multiplied by 20 percent, then multiplied by 50 percent.
If the account target says $180 because that once kept spend moving, August 17 may expose the fiction. Fix the conversion signal and the business math before you congratulate the campaign for spending the budget.
This is where our offline conversion import checklist matters. A precise Target CPA attached to junk form fills is still junk optimization with better posture.
Use a three-column decision sheet
For each affected campaign, record three numbers:
- Stated target: the Target CPA or Target ROAS in the account.
- Recent actual: performance across a window long enough to cover at least two conversion cycles.
- Business target: the efficiency the economics can support now.
Then make one of three decisions.
- Keep it. The stated target matches the business target. Document that choice and watch the rollout.
- Change it. Recent performance is better and the old target is too loose. Use Google’s Bid Target Adjustment Tool or enter a target supported by the business math.
- Escalate it. The numbers disagree because conversion quality, lag, attribution, or margin is unclear. Do not let the platform settle the argument.
Google says its adjustment tool may not recommend a target for campaigns with fewer than seven conversions. That is not permission to guess. Low-volume campaigns need a wider evidence window, pooled business data, or a different bidding decision.
Do not change five things at once
The temptation is to update targets, raise budgets, swap bid strategies, restructure campaigns, and clean up conversions before lunch. Then performance moves and nobody knows why.
Google’s target-bidding FAQ says Smart Bidding responds to target changes in real time, but teams should wait one to two conversion cycles before evaluating results. It also advises against adding data exclusions or bid limits solely because of this update.
Log the old setting, the new setting, the reason, the owner, and the evaluation date. Screenshot the bid-strategy report. Add a note to the reporting calendar. This is the same basic discipline behind a useful marketing causality check: preserve enough evidence to know what changed.
Watch the awkward two-week window
Google warns that Performance Planner and related forecasts may be inaccurate from August 17 through August 31 while systems transition.
That lands in the middle of back-to-school campaigns and close enough to fall planning that a tidy forecast can look more certain than it is. Put a note directly in any budget recommendation built from that window. Compare the forecast with actual spend, conversions, conversion value, CPA or ROAS, and channel allocation.
Performance Max and Demand Gen deserve extra attention because Google says spend can shift across channels even when the overall effect is consistent. If channel mix matters to the business, save the before view. Otherwise, the change may disappear inside an aggregate number.
The 30-minute preflight
Before August 17:
- Filter for budget-limited campaigns using an affected bid strategy.
- Export stated targets, recent actual performance, budgets, and conversion volume.
- Calculate the current business target from qualified outcomes and margin.
- Choose keep, change, or escalate for each campaign.
- Record every setting change and its reason.
- Set the review date one to two conversion cycles later.
- Flag August 17–31 forecasts as transition-period estimates.
Then leave the account alone long enough to learn something.
Automation works best when the instruction is honest. This update is a good excuse to make sure yours is.
Is the ad account optimizing toward the right business outcome?
We can audit the targets, conversion signals, and reporting logic before a platform setting quietly becomes the strategy.