Google Smart Bidding Has Changed: Why Your Forecasts May Now Be Wrong

The Aug. 17 Change Is Live. Here Is What Comes Next.

Google's Smart Bidding update is no longer upcoming. It is in effect. Your Target CPA and Target ROAS are now the primary levers controlling campaign efficiency, including in campaigns that are budget-limited. If you have not audited those targets against your actual business margins, you are now running on settings that the bidding system is actively using to make spending decisions.

The first week of post-rollout data is useful but needs to be read carefully. AI bidding systems go through learning periods when any significant change is introduced. Performance that looks worse than expected right now may reflect the algorithm adjusting, not a structural problem with your campaigns.

Why Your Paid Forecasts Are Now Less Reliable

A paid media forecasting analysis published by Neil Patel on Aug. 21 puts this moment in sharper context. The analysis identifies AI bidding on Google and Meta as a factor that is reducing performance predictability across the board. The core finding is direct: bid strategy adjustments have less direct impact than most teams assume. That is a significant statement to absorb right after a major Smart Bidding infrastructure change.

The same analysis identifies where paid media forecasts most commonly break. It is not at the strategy level. It is at CPC inflation and conversion rate volatility. Those are exactly the variables that become harder to control when the bidding system is recalibrating after a rollout like the one Google just pushed.

• CPC inflation and conversion rate volatility are the most common points where paid forecasts fail, according to the Neil Patel analysis
• AI bidding systems on Google and Meta are reducing predictability, meaning your expected performance ranges should be wider than they were before
• Bid strategy adjustments, including changes to Target CPA and Target ROAS, have less direct impact than most advertisers assume under AI bidding
• New campaigns and campaigns coming through a major algorithm change run negative for the first several weeks as the system learns

The Learning Period Problem

The Neil Patel analysis makes a point that applies directly to where advertisers are right now. Forecasts that skip the algorithm learning period set expectations that fail before the campaign does. This matters because the Smart Bidding rollout is a meaningful enough change that campaigns, particularly budget-limited ones with targets that were not recently calibrated, may behave like new campaigns in terms of how long they take to stabilize.

If you are looking at your Aug. 18 or Aug. 19 performance data and drawing conclusions, you are drawing them from a learning window, not a steady state. The appropriate response is to note the direction of change, not to make immediate target adjustments that force another learning cycle on top of the one already underway.

Creative Decay Is a Forecasting Variable Now

The Neil Patel analysis also flags creative decay as a predictable variable that belongs in every paid forecast from day one. This is relevant because Smart Bidding relies on conversion signals and historical performance patterns to make bid decisions. If your creative is aging and click-through rates are declining, the bidding system sees that signal and it influences how aggressively it bids, regardless of what your Target ROAS says.

After a major bidding change, the interaction between creative performance and bid behavior becomes more visible. Campaigns that were coasting on strong historical signals may surface problems faster now that the system is placing greater weight on your declared targets rather than smoothing over inefficiencies.

• Review creative performance metrics including click-through rate trends over the past 60 days for any campaign showing early post-rollout instability
• Flag campaigns where creative has not been refreshed in 90 or more days. Those are the most likely to have degraded signals feeding the new bidding behavior.
• Do not conflate creative decay with bidding system failure. The Neil Patel analysis is clear that these are separable variables and should be tracked separately in any honest forecast.
• Hold your Target CPA and Target ROAS steady for at least two full weeks post-rollout before deciding whether an adjustment is warranted

Paid media forecasts most often break at CPC inflation and conversion rate volatility, not at the strategy level. AI bidding on Google and Meta is reducing predictability, and bid strategy adjustments have less direct impact than most teams assume.

Neil Patel, August 21, 2026

What This Means for Your Conversion Tracking

Your Target ROAS is only as accurate as the conversion data feeding it. The Neil Patel analysis frames forecasting as a sequence: forecast reach first, then efficiency. That sequence only works if your conversion tracking is clean. If your pixels are misfiring or conversion values are stale, the new Smart Bidding behavior is optimizing toward a flawed signal, and no amount of target adjustment will fix that until the underlying data is corrected.

A tracking audit right now, in the first weeks after rollout, is not redundant with whatever you did before Aug. 17. The rollout itself is a good reason to verify that your conversion events are firing correctly and that your reported values match your actual transaction data. The bidding system is using that data actively and with more weight on declared targets than it was before.