When the Smart Bidding change went live in mid-August, we published a piece on the mechanism and a checklist, with the advice to inspect first and avoid big moves. More than a month on, Google confirmed the update completed on August 27, the first outside data is out, and the direction is clear, CPCs on budget-limited campaigns have climbed. That leaves one question, how should accounts be adjusted.

For this piece I went through the official help documentation, coverage from Search Engine Land and Search Engine Roundtable, and analyses from two third parties, Smarter Ecommerce and Relevant Audience. The mechanism does not need repeating; this article covers the judgment calls and moves for after the update wrapped up.

One Paragraph on What Actually Changed

Google began an update to the bidding behaviour of budget-limited campaigns on August 17 and completed it on August 27, covering Search, Shopping, Performance Max, and Demand Gen campaigns that use a target-based bid strategy. One line in the help documentation is easy to skim past, Google will not adjust your targets or budgets automatically. Your settings stay exactly as they were, and changing them is your call.

The mechanism that was removed is called bid suppression in industry parlance, a third-party term; the official description is that the system automatically decreases bids to avoid hitting the budget cap. Its effect was to buy cheaper traffic for budget-limited campaigns and let them outperform their stated targets. That suppression is now gone. Search Engine Land's headline framing is that the update appears to have ended bid suppression, and one third-party summary is worth keeping, the direction is established by Google's own documentation, while the magnitude is unmeasured for your account until you look at your own numbers.

How to Tell If You Are Affected

There is exactly one filter for the affected set, campaigns labelled Limited by budget that use Target CPA or Target ROAS, across the four campaign types above. Campaigns outside that set are not directly affected, but they face second-order effects, covered below.

The comparison method comes from the third-party write-ups, compare August 1 to 16 against August 28 onward, and skip the August 17 to 27 rollout window, when the change was phasing in and the data blends old and new behaviour. Under that comparison, look at four signals.

First, CPC. Rising CPCs on budget-limited campaigns are the documented effect of this update, not an anomaly. If you see CPC climbing and clicks falling on a limited campaign, check the target you set before suspecting the auction environment.

Second, where impression share is being lost. This is the most counterintuitive one. Before August, budget-limited campaigns lost impression share mainly to "rank", and only a small share to "budget", an irony baked into the label itself. After the update the two flip, budget losses expand clearly while rank losses narrow. If a campaign is going through that flip, it is direct evidence of the change.

Third, the volume-and-price combination. Target attainment has tightened, but the price may be fewer conversions at a higher cost each. That is a real trade, not an error. If volume and price both deteriorate, go to the fourth point.

Fourth, do not watch only the limited campaigns. In the third-party sample, non-limited campaigns saw cheap clicks flow in from the opposite direction, meaning low-cost inventory that suppressed campaigns used to absorb entered other campaigns' auctions. If your non-limited campaigns got more efficient recently, this may be why, rather than anything you did.

One more check is easy to miss. Before comparing windows, confirm conversion tracking was consistent across both, with no changes in GA4 or in the conversion actions themselves. Limited campaigns may have been subsidising blended efficiency numbers all along, which makes this a reporting problem first and a bidding problem second. E-commerce accounts feel it first, because Shopping and PMax campaigns most often run against a hard budget ceiling.

Finally, put the data itself in context. The magnitudes come from one vendor's client sample, with neither sample size nor industry distribution disclosed, and readings can differ completely across markets and verticals. Trust the direction; hold your account responsible for the magnitude.

Four Response Paths

First, widen the budget and the target together. If the campaign is worth keeping, move the target close to what it actually delivers and raise the budget to where it can spend fully, changing both variables together. Google's line is that scaling gets smoother once a campaign transitions from limited to non-limited, but only if the target is right. Moving the budget while leaving an outdated target in place is locking a second door behind yourself.

Second, move constrained budget toward incremental scenarios. Much of the spend that was being stretched by the old mechanism is now free, and the first question is whether it buys incremental growth, not whether it restores the old efficiency. There are two natural destinations, new scenarios such as new regions, new audiences, and new structural tests; and the non-limited campaigns that are benefiting from the cheap-click flow. This is not a promise, just a more honest starting point than forcing the old efficiency back.

Third, set targets by profit rather than by ROAS. A ROAS inflated by suppression is a false baseline, and setting targets against it only steers the system the wrong way. Rebuild the target from your profit structure, and if needed accept buying volume within your budget via Maximize Conversions, a campaign type that structurally cannot be budget-limited. Whether to switch is a decision to run through the margin math first.

Fourth, hold a 2 to 4 week observation window before changing structure. Third-party conversion attribution runs until late September, so every ROAS-level conclusion before then is an interim reading. Inside the window, limit yourself to recording and small tweaks, changing one variable at a time, and do not touch budgets and targets simultaneously. Once the two windows are clean, set the final numbers. This is the most counterintuitive discipline of the wrap-up period, and the most important one.

Pre-Q4 Monitoring List

For the daily view, watch CPC and impression share movements on budget-limited campaigns, plus budget spend-through and its rhythm. The rhythm deserves its own line, if spend-through turns erratic, the system is usually settling between the old and new behaviour, so watch for several days before concluding.

For the monthly view, split target attainment by limited and non-limited sides, calculate blended CPA and blended ROAS separately on a profit basis, recheck conversion tracking health, and rerun a plan in Performance Planner, since Google notes the forecasting models have stabilised with the update and the earlier transition-period bias can be considered over.

Q4 adds one of its own. Before peak season, re-project budgets and targets under the new behaviour instead of carrying over elasticity assumptions from before August. Account optimisation is worth staging earlier than usual this Q4.