The old budget constraint could create accidental efficiency
Google completed a global Smart Bidding change on August 27, 2026 for campaigns using target-based bidding while limited by budget.
The change was announced earlier, but its operational consequence is becoming clearer now that advertisers have several weeks of post-rollout data.
Google’s official position is straightforward: budget-limited campaigns using Target CPA, Target ROAS and some Target CPC setups should now optimize more consistently toward the target advertisers explicitly set.
Before the change, some campaigns constrained by budget could materially outperform that target.
A campaign with a $100 Target CPA might deliver at $70 because the system had more profitable demand available than the budget allowed it to capture.
That looked excellent in the dashboard.
It also meant the target was not functioning as the effective performance boundary operators assumed it was.
The configured target now matters more
Google gives a blunt example in its documentation.
If a campaign has a $10 Target CPA but historically achieves a $5 actual CPA while limited by budget, the new system may move actual performance closer to $10 unless the advertiser changes the target.
That changes the interpretation of “historically overperforming.”
Before August, a media buyer could leave a loose Target CPA in place while budget scarcity effectively suppressed bidding.
After the update, the target itself becomes more economically active.
The account now has fewer hidden guardrails.
This means target settings need to represent a real business tolerance, not a forgotten number inherited from an old campaign setup.
Early post-update analysis suggests CPC pressure
Search Engine Land reported September 17 on analysis from Smarter Ecommerce’s Mike Ryan examining post-update behavior.
The analysis suggested that the previous bid suppression effect may have weakened or disappeared in some budget-limited campaigns, with higher CPCs and lower impression share appearing after the update.
That is not a Google-published aggregate result.
It is early practitioner analysis.
The distinction matters because the magnitude can vary by account, campaign type, competition and target.
But the mechanism is consistent with Google’s stated product change.
If the system is now willing to bid closer to a loose target even when budget is constrained, it can pay more for eligible traffic than it did under the previous behavior.
This is not a simple “Google raised CPCs” story
CPC is an auction outcome.
The product change is about how the bidding system behaves when two controls conflict:
Budget says: there is limited money available.
Target says: conversions are worth paying up to a certain economic threshold.
Historically, budget scarcity could cause the system to become more conservative than the target required.
Google wants target-based campaigns to operate more predictably around the stated target.
That can increase CPC without necessarily making the system worse.
If higher bids buy more valuable conversions while the campaign stays inside a rational Target CPA or ROAS, the economics can still improve.
The problem occurs when the target itself is too loose.
Loose targets are now more dangerous
A Target CPA should represent the acquisition cost the business is actually willing to pay under the modeled conversion definition.
A Target ROAS should reflect the value signal being sent.
If those numbers are disconnected from contribution margin, lead quality or customer value, the bidding system can now exploit that disconnect more aggressively.
This creates a direct link between bidding configuration and finance.
The operator needs to know:
- what the conversion is worth;
- whether value reflects margin;
- whether target CPA includes lead quality;
- whether the campaign can afford scale at the configured threshold.
The update reduces the usefulness of target settings as soft suggestions.
Google recommends reviewing campaigns that historically beat target
Google specifically advises advertisers to review budget-limited campaigns that have historically performed better than their targets.
The company also provides a Target Adjustment Tool and says it does not automatically change targets or budgets.
That means responsibility sits with the advertiser.
If recent average CPA is $60 and the configured target is $100, the account should not assume historical $60 performance will persist simply because it has for months.
The operator now needs to choose deliberately:
- lower the target;
- increase budget;
- remove the target and maximize volume/value within budget;
- or accept performance closer to the stated target.
Budget and target are different control surfaces
This update clarifies a conceptual mistake common in PPC.
Budget controls how much capital can be deployed.
Target controls what level of efficiency the bidding system should pursue.
They are not interchangeable.
A team that wants to preserve a strict efficiency level should encode that requirement in the target.
A team that simply cuts budget to force the account to become more efficient may no longer receive the same side effect.
That is a healthier model conceptually.
It is also less forgiving of sloppy settings.
Failure mode: reacting immediately to a few days of volatility
Google says the rollout completed August 27 and advises advertisers with long conversion delays to wait one to two conversion cycles before evaluating performance.
That is important.
A change in CPC can appear immediately.
A change in actual conversion value can take longer to mature.
Lead-gen accounts can be especially vulnerable to premature conclusions.
If the conversion takes 30 days to qualify, a two-week CPC spike is not enough evidence to determine whether downstream economics worsened.
Failure mode: lowering targets without understanding volume loss
A campaign that delivered $60 CPA under a $100 target may tempt the operator to immediately set the target to $60.
That may preserve efficiency.
It can also reduce reach.
The correct target depends on the marginal economics of additional conversions.
If the business can profitably acquire more customers at $80, forcing the system back to $60 may sacrifice profitable scale.
The update creates a reason to revisit economics, not simply restore the old dashboard number.
Failure mode: comparing pre- and post-update periods without controlling for market changes
September auction conditions can differ from July.
Holiday demand, competitor budgets, product launches and creative changes can all move CPC and conversion rates.
The strongest analysis should compare:
- similar campaign types;
- similar demand periods;
- target changes;
- budget changes;
- conversion lag;
- impression share;
- CPC;
- conversion volume;
- contribution or downstream quality.
A platform change can explain a pattern without explaining every fluctuation.
What operators should audit
Start with campaigns that meet three conditions:
- Target CPA or Target ROAS.
- Limited by budget.
- Historically significant overperformance versus target.
For each, record:
- configured target;
- actual pre-update average;
- actual post-update average;
- CPC;
- impression share;
- conversion volume;
- value or downstream quality;
- budget utilization.
Then classify the gap.
If the target is intentionally loose because the business can afford it, higher CPC may be acceptable.
If the target is stale, fix it.
If the business cannot afford performance near the configured target, the setting is now exposing a risk that already existed.
The broader lesson
Automation is making campaign settings more literal.
Media buyers can no longer rely as heavily on accidental behavior created by platform constraints.
A budget is a budget.
A target is a target.
A conversion value is a value signal.
As platforms automate more of the auction, the remaining human inputs become more important because the system increasingly acts on them.
The August Smart Bidding change is a clear example.
The platform is not taking control away from the operator.
It is making the operator’s stated control more consequential.
