The google ads bidding change that ended the cheap conversion
Padding the target and capping the budget bought cheap conversions for years. Google closed that gap on 17 August, and the cautious accounts are the ones that will feel it.
What actually changed for target CPA on 17 August#
A padded target and a capped budget quietly bought cheap conversions for years. Google ads bidding switched that subsidy off on 17 August.
Every account carries a number nobody re-reads. It was typed into the bid strategy settings during onboarding, it looked prudent at the time, and it has been left to sit in the undergrowth ever since, because the actual cost kept landing underneath it. That number is your target CPA, and on 17 August 2026 google ads bidding stopped treating it as a ceiling you happily undershot and started treating it as a price Google will genuinely try to charge you.
Google's own paperwork is unusually plain about this. The help page covering the google ads bidding switch, published by Google Ads Help, 2026, says the new behaviour begins on 17 August 2026 and offers a worked example that leaves nothing to interpretation: "If your campaign's Target CPA is $10, but your recent actual CPA performance is $5, your campaign will deliver more closely to a $10 actual CPA starting August 17, 2026." Same budget, dearer conversions, fewer of them.
The trigger is not the strategy on its own, it is the strategy plus a cap, and that pairing is the whole of what google ads bidding now treats differently. Not all google ads bid strategies are caught by it: only the target-based ones, and only while they are capped. Google's accompanying Google Ads Help FAQ, 2026 states that campaigns which are not budget-constrained see no behavioural change at all, and that daily and monthly budget limits will always be respected. So the affected population is precise rather than sweeping: campaigns wearing the Limited by budget status while running Target CPA, Target ROAS, or Demand Gen Target CPC.
Why google ads bidding never gave you a discount#
Read the strategy's own definition and the logic falls out. Google's page on About Target CPA bidding, 2026 describes an average, not a limit: "some conversions may cost more than your target and some may cost less, but altogether, Google Ads will try to keep your cost per conversion equal to the target CPA you set." A campaign persistently delivering at half its target was therefore not obeying a cheaper instruction. It was hitting a wall. The budget ran out before the bidding could climb, and the target CPA sat above it like a scent the campaign never got close enough to follow.
The auction literature has described that constraint for years, in plainer language than a help page manages. Deng, Mao, Mirrokni and Zuo, 2021 model auto-bidding as an agent operating under a return-on-ad-spend constraint and a budget constraint at once, and Duan and colleagues, 2025 define the auto-bidder's task as maximising cumulative value within a specified budget, allocating spend across stages so it does not exhaust early. Persistent underdelivery against a target CPA is simply what a binding budget looks like from outside the system, and google ads bidding has stopped presenting it as a result.
That doubling is not a folkfox extrapolation. Greg Finn of Cypress North gave PPC Land, 2026 the same arithmetic in July, in spoken English: "that 5 dollar acquisition cost you have is going to be moved up to 10." He noted the change arriving dressed in the language of predictability, and predicted rising click costs as the visible symptom.
Why the padded target CPA punished the cautious#
The padding is rational, and it is documented in Google's own rulebook rather than invented by nervous marketers. The policy on Google Ads gambling and games, 2026 requires certification before an operator may run gambling ads at all, with country-specific licences to hold and maintain, while the policy on Google Ads financial products and services, 2026 states that advertisers must complete a verification process to advertise financial services in some locations. Every one of those gates is a cost that lands somewhere, and it usually lands as headroom in the target that google ads bidding has now been told to reach.
So picture the pendulum, because that is the shape of it. A cautious team set the target high to survive the worst month, the cap held the campaign well below that height, and the account banked the difference every single day without anyone recording it as a benefit. The swing back up is not google ads bidding punishing you for doing anything wrong. It is the pendulum being released, and the further you had pulled it from where you actually operate, the further it now travels. That distance is the gap.
So the accounts about to feel this hardest are the careful ones, not the reckless ones. A team that set a tight, honest target and let the budget breathe has almost nothing to do this week. A team that padded generously and capped hard has the largest gap between the price it pays and the price it authorised, and that gap is now a scheduled rise. It is a quiet reversal of who gets rewarded for prudence.
Your $35 CPA on a $50 Target CPA wasn't free efficiency. As of Aug 17, Google steers budget-limited Target CPA/ROAS campaigns closer to the target you actually set not the number Smart Bidding found. Audit that gap before your next budget bump.
The practitioner's phrase is the one worth stealing: the gap was never free. It was a structural artefact of a budget cap, sitting in reports as evidence of skill, and it survived only because nobody had cause to interrogate it. Audit that gap now, before the next budget increase, because a budget rise and a google ads bidding change landing in the same fortnight will make the resulting cost movement impossible to attribute to either. That is the diagnostic discipline folkfox brings to every paid search engagement, and it matters twice over in iGaming and fintech accounts where the padding was largest.
Which campaigns this google ads bidding change actually touches#
Before you touch a target CPA setting, establish whether the google ads bidding change reaches your campaign type at all, because the published record is not perfectly consistent and it is worth knowing which side of the disagreement you stand on. Three documents describe the scope, two of them written by Google, and they do not name the same list.
Here is the disagreement stated without spin. Google's change documentation, 2026 names Search, Shopping, Performance Max, Demand Gen, Display, Hotel and Travel as impacted. Google's own FAQ page, 2026 lists Search, Shopping, Performance Max, Demand Gen and Travel, and does not name Display or Hotel. PPC Land, 2026 goes further and reports Hotel and Display as excluded outright. Where two Google pages differ, treat the change page as the stricter reading and check your Display and Hotel campaigns anyway.
The exclusions, by contrast, are unanimous. App campaigns, video reach campaigns and video view campaigns keep the previous google ads bidding behaviour, and nothing about their targets needs revisiting. If your account is mostly app install work or upper-funnel video, the correct action here is genuinely nothing, and that is a better answer than a busy one.
Target ROAS moves the other way, and that catches people out#
One inversion is worth flagging because it reverses the direction of the worry. Google's page on About Target ROAS bidding, 2026 describes the same averaging logic in reverse: the system tries to keep conversion value per cost equal to the target you set. So a budget-capped campaign that has been quietly returning above its target ROAS will now be pulled down toward it, as PPC Land, 2026 reports. Overachievers on value get trimmed. Overachievers on cost get dearer. Both are the same mechanism.
Neither movement is a fault in the machinery. Google's page on Smart Bidding, 2026 describes auction-time bidding as weighing signals about a person and their context in every auction, and smart bidding strategies were never promised to beat your target, only to hit it. What changed inside google ads bidding is which constraint wins when a cap and a target disagree. Until 17 August the cap won. Now the target wins, within the cap.
The arithmetic that fixes a target CPA in one sitting#
| Reading | Where it lives | What it tells you |
|---|---|---|
| Actual cost per conversion, last 30 days | Cost per conversion column, campaigns table | This is the price you have really been paying. It is the number that matters. |
| Target CPA, as configured | Bid strategy settings on the campaign | This is the price Google will now steer toward. Often nobody has revisited it in a year. |
| The gap between the two | Subtract the first from the second | Every unit of gap is a scheduled cost rise, not a hypothetical one. |
| Campaign status | Status column, look for Limited by budget | No cap means no change. A cap plus a gap means act this week. |
| Conversions, last 30 days | Campaigns table | Under seven and Google's adjustment tool will not advise you at all. |
Five lines, one sitting, no drama. Fixing a target CPA is arithmetic rather than strategy, and treating a google ads bidding change as a crisis is precisely what costs money, because a panicked account makes four changes at once and then cannot say which one moved the numbers. Read the worksheet, act on the fourth row, and leave everything else where it stands.
The single decision is whether to cut the target or accept the rise, and for most accounts it is a cut. If your recent actual cost per conversion is the price at which the account is profitable, then that price, not the padded one, is your honest target. Setting the target to roughly where you have been operating preserves what you were getting. Leaving it high hands Google permission to spend up to a number you never actually intended to pay.
From 17 August 2026 the google ads bidding change begins reaching accounts gradually rather than all at once, so an untouched campaign today may still be running old behaviour tomorrow.
The shift arrives silently. There is no dated notice per campaign, which is exactly why the worksheet beats waiting for a signal.
Google's own guidance asks advertisers to wait one to two conversion cycles before judging performance, which on typical cycles lands in mid-September to mid-October.
Compare cost per conversion and conversion volume against your pre-17-August baseline, and only then decide whether a further target adjustment is warranted.
There is one more trap in the guidance, and it is the sort that looks like diligence. Google's FAQ, 2026 advises against applying data exclusions or bid limits purely in response to this change. Both feel like sensible defences and both quietly corrupt the record: an exclusion removes the very period you need for comparison, and a hard bid limit throttles the system in ways that surface later as a mystery. Adjust the target instead.
Small accounts sit in the awkward middle, and it is worth saying so rather than pretending the tool covers everyone. A campaign turning six conversions a month has the least reliable actual cost and no recommendation waiting for it. For those dens the honest move is a longer window, ninety days rather than thirty, and a target CPA set on the median rather than the best month you can find.
What else is moving in your bid strategy settings this month#
Attribution is the real casualty of a busy month, and August has been busy. PPC Land, 2026 notes that three separate changes are landing close together, this google ads bidding shift among them, alongside campaign-level language targeting removal and an AI Max conversion launch in September. Its assessment is blunt and correct: that is a poor setup for attributing any observed movement to any single cause. Which is an argument for changing one thing this week, not four.
The second item worth knowing about touches creative rather than cost. Search Engine Land, 2026 reported on 17 August that Google is extending Performance Max asset handling so that generative AI produces video aspect ratios an advertiser has not uploaded, filling the vertical and square gaps automatically. PPC News Feed, 2026 carried the same news on 19 August. It is on by default, and both reports give the same deadline for declining it: 4 September 2026.
You supply the ratios you approved
A brand uploads horizontal, vertical and square cuts it has signed off, and any placement it has no asset for simply goes unserved. Reach is narrower and every frame that runs has been through review.
Generative AI fills the gaps for you
Missing aspect ratios are created automatically from the assets you did upload. Reach widens, and frames a compliance reviewer has never seen can appear in placements a regulated brand is answerable for.
We could not verify the opt-out mechanism itself, and we would rather say so than tidy it away. Both trade reports describe an opt-out form and an account team route, but we located no Google blog post announcing the change and no public URL for that form, so treat the mechanism as reported rather than confirmed. If you run creative in a licensed category, the sensible move is to raise it with your Google account team directly this week rather than hunting for a form that may not be publicly listed.
For brand-led advertisers and anyone whose creative carries mandated wording, that default matters more than the target CPA change does. A generated crop can move a legal disclaimer out of frame without anybody deciding it should, and the same applies wherever paid social assets are reformatted by a platform rather than a person.
The fox that survives the lean months does not bolt at every rustle in the thicket, it knows the price of its own quarry. Read your actual cost, cut the padded target CPA to match it, watch the volume for a fortnight, and leave the rest of your google ads bidding alone. The accounts that hold a steady trail through September will be the ones that did the arithmetic in August, not the ones that prowled the settings hardest.
Frequently asked questions#
What is a good target cpa for google ads?
A good target CPA is the cost per conversion at which your account is genuinely profitable, not a padded number set for comfort. Since 17 August 2026 Google steers budget-limited campaigns toward the figure you configured, so a target well above your real economics is now an instruction to spend more per conversion. Take your actual cost over thirty to ninety days and set the target there.
Will my costs actually go up because of this change?
Only if the campaign is both budget-limited and delivering below its target, because google ads bidding only reaches for the target when a cap was holding it back. Google's worked example shows a campaign at a five dollar actual cost against a ten dollar target moving toward ten. Your budget is still respected, so total spend does not rise, but the same money buys fewer, dearer conversions.
Does this affect campaigns that are not limited by budget?
No. Google's FAQ is explicit that campaigns without a budget constraint see no change in google ads bidding behaviour. The mechanism depends on a cap preventing the bidding from reaching the target you set. Check the status column first: if Limited by budget never appears, this update does not apply.
Which campaign types are excluded from the change?
App campaigns, video reach campaigns and video view campaigns keep the previous google ads bidding behaviour, and every document we read agrees. Search, Shopping, Performance Max, Demand Gen and Travel are in scope across all sources. Display and Hotel are named by Google's change page but not its own FAQ, so check those two yourself.
What is the Bid Target Adjustment Tool and should I use it?
It is a Google tool, live since 6 July 2026, reachable from an account notification banner or the campaigns page. It reads recent performance and recommends a revised target. Use it as a second opinion: it only advises campaigns with seven or more conversions, and Google has confirmed it will not change targets or budgets on its own.
Should I lower my budget instead of lowering my target?
Lowering the budget tightens the constraint, which is the opposite of what you want, because the cap is what created the gap. Cutting the target CPA is the change that resets the price Google steers toward. Google also advises against applying data exclusions or bid limits purely in response to this update, because both damage the data you need to judge the outcome.
How long should I wait before judging the result?
Google asks advertisers to allow one to two conversion cycles before evaluating performance, which for most accounts puts a reliable reading between mid-September and mid-October 2026. The rollout is gradual over several weeks, so an early comparison may be measuring a campaign that has not yet been reached. Set a baseline now.
Read more on this topic#
Customer Match lists in Google Ads
The audience side of the same account: what Google will and will not let you build a list from.
Read the pieceInvalid traffic and the AI Max migration
One of the other September changes that will muddy your attribution if you measure this one too late.
Read the pieceGoogle Ads language targeting removed
The third change landing this quarter, and the reason to alter one setting at a time.
Read the piecePPC reporting and the Merchant Center restatement
What to do when the numbers in your report move for reasons that have nothing to do with your bidding.
Read the piece
Want the arithmetic done before the rollout reaches you?
We audit target-based bid strategies in regulated accounts, find the gap between what you configured and what you were really paying, and reset it before the cost rise arrives.