According to smec, cited by PPC Land, before 17 August 2026 about 55% of budget-limited campaigns beat their ROAS target, with a median overachievement of about 20%. In three weeks, that surplus vanished: about 11% in the week of 17 August, 6% in the week of 24 and about zero by 31 August. The sample wasn’t disclosed and data is preliminary, but the pattern is consistent with Google’s change.
What changed on 17 August
On 17 August 2026, Google changed bidding for budget-limited Target CPA and Target ROAS campaigns. According to smec, Google deliberately reduced bids to avoid exhausting the budget, which led campaigns to beat the target. The change eliminated that overachievement: the campaign now chases the set target instead of exceeding it.
The sources consulted indicate it applies to Search, Shopping, Performance Max, Demand Gen and Travel, and doesn’t affect App, Video, manual CPC or Target Impression Share, nor campaigns without a budget limit. Google had promised a gradual adjustment.
smec's numbers
The analysis was published by smec (Smarter Ecommerce) in an episode of 20 October 2026. Before 17 August, 55% of budget-limited campaigns beat the ROAS target, against about a third of the others. Median overachievement was about 20%, against about 5% in non-limited ones. Only about a third of limited campaigns were on target.
After: the week of 17 August had a median of about 11%, that of 24 August about 6%, and that of 31 August about zero. In preliminary data, about 50% were on target, about 25% above and about 25% below. Impression share lost to budget went from about 5% to dominating over rank.
ROAS surplus in budget-limited campaigns
20% to 0
median overachievement of target, in three weeks
Before 17 August: 55% beat the target (median ~20%). Week of 17 Aug ~11%, 24 Aug ~6%, 31 Aug ~0. smec data, preliminary, sample undisclosed.
The UK retailer case
smec presents a case from an unnamed UK kitchen-and-home retailer: average CPC went from about 17 to about 27 pence in one week. The host speaks of about 44%, but PPC Land’s calculation with rounded values gives about 59%. Target ROAS rose to about 13 and real ROAS began following that line.
It’s a single case and the video’s automatic subtitles may contain errors. It serves as an illustration that, in accounts with limited budget, cost per click may rise when the system stops “lowering the bid” to save budget.
Limitations worth acknowledging
PPC Land notes all numbers are smec’s own analysis, that Google doesn’t participate or respond, that the sample wasn’t disclosed (neither the number of campaigns, accounts or markets, nor how “budget-limited” was defined), and that post-change data is preliminary. smec describes itself as working with over 350 global retailers, but that isn’t presented as this analysis’s sample.
So read the result as a plausible pattern, not a definitive measurement. The direction (surplus vanishing) is consistent with what Google did; exact values may vary a lot from account to account.
The ROAS surplus vanished in three weeks. Does your target still make sense?
What this means for a Portuguese SME
If you have budget-limited campaigns (the platform shows “limited by budget” and impression share lost to budget), it’s likely that, since August, real ROAS or CPA converged to the target you set, and that CPC rose. This isn’t necessarily bad: the system is doing what you asked. It may, however, be different from what the account showed before.
If your target was set when the campaign always “overachieved”, the target may be too demanding or too loose for the new reality. It’s a good time to review it.
How to diagnose your account
1. List campaigns with Target CPA or Target ROAS and mark the budget-limited ones. 2. Compare, per campaign, the 4 weeks before 17 August with the 4 weeks after: CPC, conversion rate, cost per result, ROAS, impression share lost to budget. 3. See whether conversion volume fell or cost per conversion rose. 4. Record and discuss with whoever runs the account.
Google recommends waiting one to two conversion cycles before assessing target changes, and the logic applies: don’t decide with a few days of data.
Possible decisions
Raise the budget if the campaign is on target and has margin: more volume at the same cost per result. Raise the target (more demanding) if you want more efficiency and accept less volume. Lower the target (less demanding) if you want more volume and have margin. Restructure if several campaigns compete for the same budget.
The decision depends on your maximum cost per customer and your margin. Without that number, any target is arbitrary. With it, the target becomes a consequence of the business.
What to ask your manager
What changed in our campaigns from 17 August? Which campaigns are limited by budget? How has CPC evolved? Which targets have we reviewed? What did we decide for Black Friday? And how do we separate this effect from seasonality?
An attentive manager has already looked at this and has numbers. If yours hasn’t, ask. And if you want an independent analysis of your Target CPA and Target ROAS campaigns, I can do it with you.
Frequently asked questions
Did my CPC rise because of Google?
It may be, if the campaign is budget-limited and uses Target CPA or Target ROAS. Compare periods and rule out other causes, such as seasonality and competition.
Does this affect campaigns without a budget limit?
According to sources, no. The change concerns budget-limited campaigns.
Are smec's data reliable?
They’re its own analyses, with an undisclosed sample and preliminary data. Treat them as an indication and confirm in your account.