Here’s the question worth asking before anything else: if a campaign has been quietly beating its own target for months, is that a problem you’re about to lose, or a mistake you never caught? Starting August 17, 2026, you won’t get to avoid answering it.

Google is changing how campaigns running Target CPA and Target ROAS strategies behave once they’re limited by budget. Per Google’s Help Center, campaigns that are “Limited by budget” on Target CPA, Target ROAS, or Target CPC for Demand Gen “will more consistently perform toward your bid target, including when you make budget adjustments.”

Google’s own example: “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.”

Here what’s actually changing, and who it hits

Google frames this as a predictability fix, and on its own terms, that holds up. The same target should mean the same efficiency whether the daily budget is $50 or $500. Right now it doesn’t: raise the budget on a constrained campaign and the CPA can swing in ways nobody asked for. After August 17, budget controls spend and the target controls efficiency, full stop. Those become two separate levers instead of one leaking into the other.

Here’s the scope of the change: Search, Shopping, Performance Max, Demand Gen, and Travel, running in Google Ads or Search Ads 360, plus Demand Gen in Display & Video 360. Target CPC is included too, but only inside Demand Gen—most coverage of this update skips that detail. Already behaving this way: Display and Hotel, so there’s nothing to audit there. Exempt entirely: App, Video reach, and Video view campaigns. Untouched regardless of budget status: Manual CPC and Target Impression Share. (Google’s FAQ is the source for all of that, if you want to check your own account type against it directly.)

A Bid Target Adjustment Tool went live in accounts on July 6, six weeks ahead of the deadline. It’s worth being precise about what it does, because the name oversells it a little. For every flagged campaign, it shows three things: recent actual performance, the currently entered target, and a recommended target calculated from that recent performance. Then one button, Apply, which sets the target to whatever the tool recommends.

So the tool does the arithmetic and the clicking. It doesn’t do the deciding. That’s an important distinction, and not just a technicality. Google will not touch a single target or a single budget on your behalf, before or after the 17th. Ignoring the tool is a real choice with a predictable outcome, not a way of deferring the decision. And “Apply” locks in one specific answer — match recent performance — which is often right but isn’t the only option. If the old target reflected a genuine cost constraint rather than a number nobody revisited, typing in your own figure is the better move, and the tool won’t tell you that; it just does the math on what’s already happened.

One more thing worth knowing: for campaigns under roughly seven conversions, Google won’t calculate a recommendation at all, because the data’s too thin to trust. Those campaigns still need a decision. They just won’t get a suggested one.

The rollout itself isn’t instant. In practice, we’re seeing it stagger across accounts over the following weeks rather than flip everywhere at once on the 17th. A quiet first few days doesn’t mean you dodged it.

Why PPC advertisers are concerned about Google’s bidding change

Joey Bidner, a freelance Google Ads manager, posted on LinkedIn:

“I have never been more frustrated by a Google Ads update than this… To me, this is one of the most self serving Google-centric changes we’ve seen in years’.”

His specific claim:

“Some of my best-performing accounts INTENTIONALLY run with low tROAS or high tCPA targets because they give smart bidding the freedom to explore, discover new customers, and find efficiencies over time.”

That’s not a complaint about losing money. It’s a claim that the “problem” Google is fixing was, for some accounts, a working strategy.

The post drew dozens of replies. Maggie Humphrey, Director of Ecommerce at Cypress North, pushed Ginny Marvin, Google Ads Liaison on the framing directly: was this really a broader Smart Bidding change rather than a budget-specific one? Marvin, who was fielding a lot of these questions in the same thread in real time, replied:

“Yes this only impacts budget-constrained campaigns using a target because this is already the bidding behavior when campaigns using a target aren’t budget constrained… In short we’re making the controls clearer; the target will more precisely control your ROI.”

On the Marketing O’Clock podcast, co-host Greg Finn worked through that same exchange live on air and came away with a different objection than Bidner’s. Marvin had told one commenter that nothing changes in how unconstrained campaigns behave, then told the room that budget-constrained campaigns will now behave the way unconstrained ones already do. Finn’s read: “There has to be two different systems… or this is all a lie.” His point wasn’t that Google is trying to squeeze more spend out of advertisers — it was that the explanation, read closely, was hard to reconcile with itself. Here’s the full clip of the discussion.

The gap between your bid target and actual performance does not tell you what to do

Only you know that.

A gap between what a campaign is actually delivering and what its target says could mean one of two things: deliberate headroom, something built on purpose, or an outdated number nobody’s revisited since it was set. The size of the gap alone doesn’t tell you which one applies. That’s not something Google or anyone in the LinkedIn thread said explicitly. It’s the actual question worth asking before you touch anything.

If you’re running the first kind, like the accounts Joey Bidner described in the previous section, this update is a real problem. A target set loose on purpose so Smart Bidding could explore was a working strategy, and August 17 removes the lever that made it work.

If you’re running the second kind, the diagnosis is different: not a loss, but a conversation about your targets that was overdue anyway, whether or not this update had forced it.

Here’s how to actually tell them apart, rather than guessing. Take the tCPA and check it against what you can genuinely afford to pay for that conversion. Take the tROAS and check it against your real gross margin, not against whatever number has been producing good headlines in the client report. A target that survives that check was probably deliberate, while a target that doesn’t, that turns out to be an old number nobody stress-tested, is the second kind, and the update just gave you a reason to fix it now instead of later.

Ginny Marvin’s own advice fits that second case, even though she wasn’t drawing this same two-way distinction herself. Per Brooke Osmundson’s Search Engine Journal’s coverage of the thread, she told a commenter named Jack Carr that the change “won’t result in spend changes on a campaign already budget constrained,” and that her broader guidance for “budget-constrained campaigns currently over-performing on their target is to ensure the targets are in line with your goals.”

That’s the actual test: were you using that gap on purpose, or had nobody looked at it in a while? Answer that honestly before you touch a single target.

How to choose the right target adjustment before August 17

Once you have made that call, here are five options to help you out.

Keep the target, only if you can say out loud why that specific number is the goal, not just that it has been producing good results. “It’s been working” is not the same as “this is my actual target.” If you land here, Google’s FAQ has a specific tactic for it, not just “do nothing”: “you should try to provide as much budget buffer as possible during periods when your campaigns are limited by budget and scale your campaign budgets when you can,” and if several campaigns share a target, “you can consolidate your campaigns under portfolio bidding or shared budgets to make better use of your limited budgets.”

Keeping the number means giving the campaign enough budget room that the target stops being the tighter constraint, not leaving it alone and hoping.

Lower or raise the target to match recent actuals if you were not running it loose on purpose. This is the default move for most accounts, and it is also Google’s own stated recommendation: “To maintain the current performance, you should update targets to match the recent performance before 17 August.” There is a real, public example of exactly how it plays out.

Chris Ridley asked Ginny Marvin directly on LinkedIn: his client wanted a target CPA between $30 and $50, was currently averaging $35, and the target had been left at $50 as an upper limit. Would leaving it there push CPA up toward $50? Marvin’s answer: “the campaign will more consistently perform towards your bid target, which in this case is $50… If you want the campaign to continue to perform at the current average target of 35, we recommend changing the target to that.” The number in the box is the number you will get. If $50 was never the real goal, $35 needs to be in the field.

On how aggressively to make that change, Google and a practitioner watching real accounts do not fully agree. Google’s guidance is that matching actuals ahead of the deadline “should not experience noticeable volatility or changes to your existing performance,” and separately that Smart Bidding “reacts to target changes in real time… and can perform well with rapid adjustments,” whether the change is large or small.

Aaron Levy, Optmyzr’s evangelist, is less reassured.

His recommendation: “start nudging your CPA targets down/ROAS targets up (slowly of course) and seeing how the new change reacts. methinks this could yield some headaches for those who aren’t ready for it.”

Google says a clean, single adjustment before August 17 should be fine. A practitioner running this in production is choosing to move in smaller steps anyway and watch each one before going further. Both can be right for different accounts; you will not know which camp yours is in until you try it.

Note: To help advertisers spot this before the deadline, Ginny Marvin told Aaron Levy directly on LinkedIn:

“We’re rolling out notifications in Google Ads accounts starting this week, where they can review historical campaign performance and apply updates to targets.”

If that notification has shown up in your account, you already know where to look. If it hasn’t, check anyway rather than assuming you’re clear: the lookback window can flag campaigns that were only briefly budget-constrained sometime in the past year, easy to miss if you’re only looking at today’s status. Pull the list yourself instead of waiting on Google to surface it.

Set a custom target from real unit economics when even recent actuals are not the right number, because the target was inherited or never checked against margin or lifetime value in the first place. Google’s FAQ confirms this is expected: “You can set a more efficient target than your current performance at any time to reflect your business goals,” with the caveat that “your daily spend will likely be impacted,” and that you can model the trade-off first using Google’s bid simulator tool rather than guessing. Matching to actuals fixes drift; it does not fix a target that was wrong to begin with.

Move to Maximize Conversions or Maximize Conversion Value if scale matters more than efficiency right now. Google frames this as the fallback for a genuinely fixed budget, instead of the default: “If your budget is strictly inflexible and you are comfortable with ROI fluctuations when budgets are adjusted, you can switch your campaign to a Maximize conversions or Maximize conversion value bid strategy.” It trades a predictable unit cost for volume, and that should be a deliberate trade for an inflexible budget, not a shortcut because picking a target feels harder than picking a strategy.

Raise the budget and keep the target, and this one is worth pausing on because the update quietly makes it a better option than it used to be. Before August 17, increasing the budget on a budget-constrained target campaign was a gamble: efficiency could swing in ways nobody predicted, because the algorithm was already borrowing from the target to make the old budget work. After the 17th, that borrowing stops. If the target is genuinely right and there’s demand sitting above your current cap, giving the campaign a daily budget comfortably higher than average spend should let it scale at the stated target instead of drifting away from it. This isn’t a hedge or a consolation prize. For advertisers who’ve been avoiding budget increases specifically because of the volatility they used to cause, this is the one piece of good news in the whole update.

One complication is worth naming here because it is the honest limitation of “just match recent actuals.” Rob Johnston, a digital marketing consultant, pointed out on Joey Bidner’s post that resetting to today’s number is a snapshot and not a fix:

“What was an efficient CPA at a particular volume today, won’t necessarily be tomorrow… as soon as the SERP gets more competitive, if quality scores reduce, etc., you then need to manually increase the tCPA again when volumes fall. Currently the system figures all of that out for you when setting more conservative targets.”

He is right that a looser target used to absorb that drift automatically. Now a person has to notice it and adjust the number by hand, which is exactly why a one-time reset is the start of this problem. More on that later.

Wait one to two conversion cycles before judging the change

Whatever you change, give it time before you judge it, including if you’re following Aaron Levy’s phased approach from above. Each nudge still needs its own full evaluation window. Google’s own guidance is to wait one to two conversion cycles before evaluating performance in the bid strategy report, and to expect Performance Planner forecasts to be briefly unreliable between August 17 and 31.

“One to two conversion cycles” sounds like a few days until you do the math for a lead-gen account with offline conversion imports. If a CRM batches conversions weekly and the average sales cycle from lead to closed deal runs two to three weeks, that is not a few days of patience. It is potentially four to six weeks before the bid strategy report reflects what actually happened, rather than the noisy re-learning period right after the change. Judge it on day three, and the number being measured is not the change. It is Smart Bidding still figuring out what was asked of it.

Here’s roughly how that window breaks down in practice.

Days 1 to 3: watch and don’t touch.

Look for anything you didn’t predict, a campaign you assumed was unaffected suddenly moving, or spend reallocating between channels inside a Performance Max or Demand Gen campaign. Resist the urge to adjust anything yet. Smart Bidding reacts to target changes in real time, and stacking a fix on top of a system-wide change you’re still watching just makes the cause of whatever happens next unknowable.

Days 3 to 14: judge by conversion cycles instead by the calendar.

This is the stretch where the temptation to react early is strongest, and it’s also where Google’s own forecasting tools are least trustworthy. Read the bid strategy report, where actual performance sits next to target directly, and hold off on trusting Performance Planner between the 17th and the 31st.

Weeks 2 to 4: fix what you got wrong in August.

Some of the targets you set under deadline pressure will turn out to be slightly off, and that’s normal. Adjust them now, with real post-change data in front of you, one change at a time rather than several at once, for the same reason you didn’t touch anything in the first three days.

Week 4 onward: this becomes a standing check instead a finished task.

Target versus actual on budget-limited campaigns doesn’t stay settled just because you fixed it in August. Performance moves, the gap reopens, and once it does, it’s not free efficiency anymore. It’s a setting that’s quietly gone stale again.

That last point is worth sitting with, because it’s the whole reason a one-time reset was never going to be enough. More on that below.

This is also why Google’s FAQ specifically advises against reaching for data exclusions or new bid limits in response to this update, calling that “not recommended, as it can cause performance fluctuations.” Marvin gave the same advice directly in the thread. Save exclusions for genuine tracking outages. This is a “targets” problem, and not a data-quality problem, and treating it like one just adds a second variable to a measurement window that already needs to stay clean.

How to audit budget-constrained campaigns without doing it one at a time

This matters most for agencies, large ecommerce accounts, and any budget-constrained advertiser with enough affected campaigns that reviewing them one at a time is a real burden. It matters less if you’re running a handful of enterprise campaigns that were never budget-constrained to begin with, or a lead-gen account where most of the exposure sits in one or two brand campaigns rather than spread across dozens.

If that’s you, auditing by hand is genuinely realistic. Here’s the sequence:

1. Pull every campaign that’s carried “Limited by budget” status, but widen the window. Google’s own flag looks back 12 months, and a campaign that hit its cap during Q4 peak and hasn’t since is still sitting on a target set for that constraint.
2. Keep only the affected strategies. Target CPA, Target ROAS, and Target CPC on Demand Gen. Drop everything else from the list.
3. Put actual next to target for each one, using a window long enough to be stable, 30 days for anything with real volume, 60 to 90 for anything thinner.
4. Rank by money at risk, but not by the size of the gap. A 2x gap on $400 a month matters less than a 20% gap on $80,000 a month. Multiply the gap by the spend and work the top of that list first.
5. Ask whether the target was ever real, the way we covered above: check it against allowable cost or actual margin, but not against what’s been producing good numbers.
6. Decide and then apply. Use the options from the previous section.
7. Write down what you changed and why. Old target, new target, date, one line of reasoning. In four weeks, this is the difference between diagnosing a problem and guessing at one, and it’s also what you’ll actually paste into a client conversation.

Across dozens of accounts, or hundreds of campaigns spread over a handful of them, that sequence stops scaling. The steps are still right. But it isn’t easy to run them against fresh data every time performance shifts, which is where a rule-based approach earns its keep.

Here is what that looks like built in Optmyzr’s Rule Engine, worth describing precisely rather than conveniently.

We built a strategy that keeps budget-limited Target CPA and Target ROAS campaigns performing at the efficiency they’ve actually been hitting, instead of letting performance drift toward a looser target once the change lands.

How it works: it finds campaigns that have consistently beaten their target across the last 7, 30 and 90 days, then eases the target toward actual performance, gradually, and only where budget is the constraint. It can run on a schedule, and either apply the changes for you or just notify you, whichever the account prefers.

Access the strategy here: https://tools.optmyzr.com/rule_engine/loadInstant/240683

How does this compare to Google’s own Bid Target Adjustment Tool?

Google’s tool is a one-time, manual review you run before August 17. You find the affected campaigns, look at recent performance, and apply new targets by hand. It does the job once, but it doesn’t keep watching after that.

Ours is built to run as an ongoing strategy instead of a one-off:

• It keeps working. Runs on a schedule and re-checks targets over time, not just once before the deadline.
• It doesn’t react to a lucky week. Only acts when a campaign has beaten its target consistently across 7, 30 and 90 days.
• Every threshold is adjustable per account, and you can run it in safe mode where it only notifies and changes nothing.

There is no single, native “Limited by budget” flag to filter on, because Google Ads treats budgets and campaigns as separate reporting objects. What you do have is Google’s own “Recommended Budget Amount”, a suggestion generated specifically when Google detects a campaign “consistently hitting its daily budget cap before the day ends,” and a metric called Search Lost IS (rank/budget), which blends rank-based and budget-based lost impression share into one number rather than isolating budget alone. Together, those two are a workable proxy for “this campaign is budget-constrained,” even without one clean checkbox for it.

For the reset itself, Rule Engine supports an Expression value type, so instead of hand-picking a number for every flagged campaign, you can set an action that adjusts a target relative to the campaign’s own gap. Optmyzr’s documentation gives the formula (ROAS-Target ROAS)/10 as a working example (at the ad group level, though the same mechanic applies at campaign scope), specifically so the adjustment is proportional to the gap rather than a blunt jump, because a drastic target change can send a campaign back into the learning phase and cost you the stability you were trying to protect in the first place.

The part that matters most is what the strategy does with that proposed change, and it does not apply it automatically. Schedule the strategy with “Add to Alerts” or “Send Notifications” checked, and leave “Apply Changes” unchecked. The suggestions land on the My Alerts page, or in Slack, and sit there until someone reviews them and applies them by hand. That is the actual mechanic, not a vague gesture toward “review and approve”: the strategy proposes, a person applies.

One honest limitation: campaign-level target edits like this only work on standard bid strategies. Portfolio bid strategies are fully automated by Google, and no third-party tool, including Optmyzr, can edit their targets at the campaign level. Change the target directly and the campaign exits the portfolio.

That limitation matters more than it might sound like, because portfolio strategies and shared budgets have their own mechanics under this update, and they’re easy to get wrong. Google’s guidance is very specific: any target adjustment has to happen at the portfolio or shared-budget level, but not on the individual campaigns inside it. If the budget isn’t shared, only the campaigns actually hitting their cap are affected. If the budget is shared and constrained, the impact spreads across every campaign in that pool, whether or not each one individually looks budget-limited. Advertisers running portfolio strategies need Google’s own Bid Target Adjustment Tool for this, since Rule Engine can’t reach in at the campaign level without breaking the portfolio.

For an agency managing this across a book of accounts, the shape of the problem is a little different from managing it for one account. The audit itself is the same seven steps, but it has to run against every client at once, sorted by revenue at risk rather than whichever account happens to come up first on the call schedule.

Before the 17th, that also means checking contracted CPA or ROAS commitments against the campaigns on this list. Any client whose contract quietly depends on a campaign over-delivering is a conversation to have this week. After the 17th, it means adding a target-versus-actual line to reporting for at least a quarter, since it’s now the metric that explains most of what moves, and setting a re-audit for late September, because targets set in a hurry under deadline pressure deserve a second look once there’s real post-change data to look at.

Optmyzr’s CEO, Frederick Vallaeys, described a version of this same approach in a Reddit thread about the update, and it is worth quoting directly because it is more specific than “reset to actuals” alone:

“If you don’t want to audit every campaign manually, we actually have a pre-built automation in Optmyzr that automatically decreases your tCPA whenever your actual CPA drops below 70% of your target. You can, ofc, customize it with your own thresholds.”

He also raised a second lever worth considering alongside a target reset:

“If you’re worried about volume drop, try using a shared budget, so campaigns with better performance get a better share of it.”

Neither of these replaces your judgment. But both replace doing this one campaign at a time.

A one-time target reset will not prevent future drift

Google’s Bid Target Adjustment Tool solves the problem directly in front of you right now. It surfaces the budget-constrained campaigns in your account, shows their historical performance, and lets you keep the current target, match it to recent actuals, or set a custom one. That is genuinely useful, and if it has not been opened yet, it is worth ten minutes before August 17.

But it is a one-time nudge tied to one deadline. It does not solve what happens in October, when different campaigns start hitting their budget caps because seasonality shifted or a client raised daily spend without anyone rechecking the target that came with it. Staying efficient across a lot of campaigns, without hand-picking a target every time performance moves, was never a problem with a deadline attached to it. It was ongoing before August 17 too. This update just made it visible, and gave everyone a reason to finally look.

That’s the piece worth carrying forward from everything above: the audit, the four-week window, the phase where you stop watching and start correcting. None of it was really about the 17th. The deadline forced a one-time version of a check that needed to exist permanently.

The method holds regardless of what builds it: flag the campaigns showing the signal, propose a target relative to their own gap, have a person review and apply it, and check back once the campaign has actually had time to settle. Optmyzr is one way to run that method without doing it by hand every time performance shifts, but the method itself does not require it. Build the rule, review the proposal, apply it, and go check again next quarter.

The campaigns that drift again are the ones nobody built a repeatable check for, not the ones that got a one-time reset in August.