Google Ads confirms the August 17 bidding update finished on August 27, and CPCs are rising

Google has confirmed that its 17 August target-based bidding change finished rolling out globally on 27 August 2026, and the first outside data on what it did to cost per click is now public. Search Engine Roundtable reported both on 15 September 2026. The short version for advertisers: bid suppression on budget-limited campaigns has ended, CPCs on those campaigns are rising, and the cheap clicks they used to absorb have moved to campaigns that are not budget-limited.

We covered the announcement of this change on 28 July and Google's follow-up video Q&A on 7 August. What is new here is a completion date, which was missing before, and the first measured effect from outside Google. Neither of those existed when the change was announced.

What Google confirmed

Google updated a help document with the rollout window. The text states that on 17 August 2026 changes to target-based bid strategies rolled out globally across campaigns that are limited by budget, that the global rollout was completed on 27 August 2026, and that the new bidding behaviour is now fully live across all affected accounts.

The same help box carries an instruction that is easy to skim past. Google recommends reviewing your settings to check they align with your business goals, specifically if your budget-limited campaigns historically performed better than your targets, and states plainly that Google does not automatically adjust your bidding targets or budgets. That second sentence is the whole operational point. Google changed the bidding behaviour and left the targets exactly where the advertiser set them.

Hana Kobzová at PPC News Feed, credited by Search Engine Roundtable for spotting the update, noted that the help page also indicates stabilisation of the Performance Planner's forecasting models, and that campaigns which previously overperformed are now expected to align more closely with entered targets.

What the change actually does

Google's original description, quoted by Search Engine Roundtable, is worth reading literally. From 17 August 2026 Google updated its bidding systems to deliver more predictable campaign performance, applying to Search, Shopping, Performance Max and Demand Gen campaigns that use a target-based bid strategy such as Target CPA or Target ROAS. Campaigns that are limited by budget and use a target-based strategy will more consistently perform toward the target, including when budgets are adjusted.

The behaviour it replaced is described just as clearly in Google's own text: when a campaign is limited by budget, it may perform better than the target you have set, and when you increase the budget on those campaigns it is hard to know exactly what performance to expect. That was the problem Google set out to fix. Budget-limited campaigns used to beat their targets, and then broke that pattern unpredictably the moment you funded them properly.

The mechanism underneath is bid suppression. To keep a budget-limited campaign inside its daily budget, Google was holding bids down. A campaign with a Target CPA of 500 baht and a budget too small to spend at that target would buy the cheapest available clicks and report a CPA below target, which looked like outperformance and was really rationing.

The first data, and what it shows

Mike Ryan of Smarter Ecommerce published median CPC data on X on 14 September 2026 and drew two conclusions from it, both quoted by Search Engine Roundtable. First, CPCs are rising for budget-limited campaigns. Second, a sizable pool of cheap clicks has been liquidated and is now available to campaigns that are not budget-limited. His chart shows the median CPC for the two groups diverging after 17 August.

His explanation of the mechanism is the part worth keeping. Google had been lowering bids on budget-limited campaigns to stop them exceeding budget, which meant most of those campaigns were effectively rank-limited by CPC rather than by budget, and that situation is now reversing.

Kirk Williams asked in the same thread whether this supports the idea that Google improved the overall system for everyone except the budget-limited accounts themselves, particularly for their non-limited competitors. Search Engine Roundtable's answer was that it does.

That framing is the uncomfortable one. If you run campaigns at a budget that lets them spend freely, a pool of cheap inventory that was previously locked up in throttled competitors' campaigns has been released into the auction you are bidding in. If you run campaigns against a budget cap, the discount you were quietly receiving has been withdrawn.

What to check in your own account

The important part of Google's own guidance is that it changes nothing for you automatically. Targets and budgets stay exactly as they were.

• List every campaign carrying the Limited by budget label that uses Target CPA or Target ROAS, across Search, Shopping, Performance Max and Demand Gen. Those are the only campaigns in scope.
• Compare CPA or ROAS for 1 to 16 August against 28 August onward, skipping 17 to 27 August because the rollout was in progress. A campaign that used to beat its target and now sits on it is behaving exactly as designed.
• Look at CPC over the same two windows for those campaigns specifically. Rising CPC on a budget-limited campaign is the documented effect, not an auction anomaly.
• Check whether volume fell while target attainment improved. Hitting the target more consistently at a fixed budget can mean fewer conversions at a higher cost each, which is a real trade and not an error.
• Look at your non-limited campaigns too. If cheap clicks were released into the auction, the effect should show up there as well, in the opposite direction.
• Only then decide whether to move a target or a budget. Doing it before you have compared the two windows means changing two variables at once.

For accounts where budget-limited campaigns had been quietly subsidising the blended CPA, this is a reporting problem before it is a bidding problem. The comparison only works if conversion tracking was stable across both windows, so confirm that in GA4 and in the conversion actions themselves before drawing conclusions. Shopping and Performance Max accounts will usually see this first, because e-commerce campaigns are the ones most often run against a hard budget ceiling.

What is not established

The data published so far is one practitioner's median CPC series, shared on X, covering the accounts that practitioner has visibility into. It is directionally useful and it matches Google's own description of the change, but it is not a controlled study and it carries no stated sample size, vertical breakdown or geography.

Google did not publish any measurement of the change's effect. It did not say how much CPCs would move, in which verticals, or in which markets. Nothing in the source material addresses whether the effect differs in smaller auctions such as Thailand, where the pool of competing advertisers on a given term is thinner than in the United States. And there is no statement about whether the released cheap inventory persists or is bid back up over time as non-limited campaigns absorb it.

Treat the direction as established by Google's own documentation and the magnitude as unmeasured for your own account until you have looked.

What this means for Thai advertisers

Budget-limited campaigns are the normal state of affairs in Thailand, not an edge case. A large share of Thai accounts run on a fixed monthly budget set in baht at the start of the year, which means their campaigns sit against the cap for most of the month. Those are precisely the accounts that were receiving the suppressed-bid discount, and precisely the accounts now seeing CPCs rise.

The practical consequence is that a Thai account reporting a CPA improvement last quarter may have been reading a rationing artefact rather than an optimisation win. If the reported CPA was consistently beating the Target CPA on campaigns flagged Limited by budget, the target was probably set too generously and the constraint was doing the work. Now that the constraint has been removed from the bidding logic, the target itself has to be right.

There is a second-order effect for smaller Thai auctions. In a thin auction, a handful of throttled competitors releasing cheap clicks can move the median CPC noticeably for everyone else, in both directions. That is not something to plan around, but it is worth knowing before attributing a CPC change in September to your own campaign work. Accounts that want the before-and-after read done properly can start with a structured review of their Google Ads setup.

Frequently asked questions

When did the August 17 bidding change finish rolling out?

Google confirmed that the global rollout completed on 27 August 2026, having started on 17 August 2026, and that the new bidding behaviour is fully live across all affected accounts. Search Engine Roundtable reported the confirmation on 15 September 2026.

Which campaigns does this affect?

Campaigns that are limited by budget and use a target-based bid strategy such as Target CPA or Target ROAS, across Search, Shopping, Performance Max and Demand Gen. Campaigns that are not budget-limited are not directly in scope, although the released inventory can reach their auctions.

Will Google adjust my targets or budgets for me?

No. Google's help text states that it does not automatically adjust your bidding targets or budgets, and recommends reviewing your settings yourself, particularly if your budget-limited campaigns historically performed better than your targets.

Are CPCs going up because of this?

For budget-limited campaigns, the early data says yes. Mike Ryan of Smarter Ecommerce published median CPC data on 14 September 2026 showing CPCs rising for budget-limited campaigns while cheap clicks became available to non-limited campaigns. Google published no figures of its own.

My CPA got worse after August. Is that this change or something else?

It could be either, and the way to tell is to compare 1 to 16 August against 28 August onward on the specific campaigns labelled Limited by budget, leaving out the rollout window of 17 to 27 August. If target attainment tightened while CPA rose toward the target, that pattern matches the documented change.

The short version

Google removed a subsidy nobody had asked for and few advertisers knew they were receiving. Budget-limited campaigns used to beat their targets because their bids were being held down, and that is over as of 27 August. The reported numbers on those campaigns will now sit closer to the targets you actually entered, which is more honest and, for accounts whose targets were set loosely against a suppressed baseline, more expensive. Nothing adjusts itself, so the work is to compare the two windows and then decide whether the target or the budget was the thing that was wrong. Advertisers who want that comparison run across an account can talk to our Google Ads team.