Google changes Smart Bidding for campaigns limited by budget

What changed

Frederick Vallaeys’s original post describes an Optmyzr scenario model, not observed results after Google’s bidding change. The sample covered 1,943 budget-limited target CPA and target ROAS campaigns with at least 30 conversions. Its estimated median loss of about 9% assumes that performance moves fully to the stated target while the budget stays unchanged. That assumption is central to interpreting the chart.

The practical exposure is a gap between the efficiency a campaign has been achieving and the target the advertiser has configured. A target that once looked comfortably loose can become consequential if the system begins using more of that allowance. This does not mean that every campaign will experience the model’s median result. The post explicitly says that campaigns already at or worse than their targets have no loss from this specific modelled effect.

Consider an illustrative campaign spending £600 while acquiring 100 conversions at £6 each, despite a £10 target. If the same spend instead produces conversions at £10 each, the arithmetic gives 60 conversions. That is a scenario calculation, not a forecast for a real account. It shows why a target should reflect the business’s acceptable trade-off rather than a number inherited from an earlier setup.

Why it matters for advertisers

Review both sides of the bidding instruction: how much the business can invest and what efficiency it needs. After the reported change date, inspect actual CPC, conversion rate, query mix and delayed conversions before deciding what happened. A shift in those measures can help distinguish paying more for similar traffic from reaching a different traffic mix. Keep the model as a prompt for investigation, then judge the account using its own evidence.