Google Ads Target CPA Enforcement (August 17, 2026): What Small Advertisers Must Do Before Monday
On August 17, 2026, Google Ads starts making budget-limited Target CPA, Target ROAS, and Demand Gen Target CPC campaigns deliver toward the number you entered in the target field instead of whatever they were previously delivering. Google’s own example in its Help Center is blunt: a campaign with a $10 Target CPA currently running at a $5 actual CPA “will deliver more closely to a $10 actual CPA starting August 17, 2026.”
Most coverage frames this as affecting “budget-constrained campaigns,” which reads like a subset. If you run a $1,500-a-month account, budget-constrained is the permanent condition of every campaign you have ever run. This is not an edge case for you. It is the whole account.
What actually changes in Google Ads on August 17, 2026?
Budget stops leaking into efficiency. From August 17, budget controls how much you spend and the target controls what each conversion costs — two independent levers instead of one bleeding into the other.
The mechanism predicts more than the headlines do. Today a campaign hitting its daily cap has two constraints fighting each other, and the tighter one wins: Smart Bidding throttles into the cheaper end of the auction to fit a day inside the cap, and the actual CPA lands below target as a by-product. That gap was never a discount. It was your budget ceiling, showing up in the CPA column.
After Monday the target binds regardless. The system bids into the auctions your target permits and stops when the budget runs out. Same spend, more expensive conversions, fewer of them.
Scope, per Google’s FAQ: Search, Shopping, Performance Max, Demand Gen, and Travel are newly affected. Target CPC is included only inside Demand Gen. Display and Hotel already work this way. App, Video reach, Video view, Manual CPC, and Target Impression Share are untouched.
Does the August 17 change make my ads more expensive?
Not your bill. Your unit price. If your daily budget is a hard cap, your spend on Monday the 24th looks identical to your spend on Monday the 10th. What falls is the number of conversions that spend buys.
Ginny Marvin, Google’s Ads Liaison, told one advertiser on LinkedIn the change “won’t result in spend changes on a campaign already budget constrained.” That is why the headlines about doubling CPAs and the reassurance about flat spend are both accurate at once. Price per conversion moves; total spend does not; volume absorbs the difference.
Real numbers. A local HVAC company runs $50 a day, or $1,500 a month. Target CPA was set at $80 when the account was built and nobody has touched it since. Actual CPA has been running $45 — about 33 booked calls a month. Once the target binds, $1,500 at $80 a call is about 19. Same invoice from Google, fourteen fewer jobs on the calendar, and no line item anywhere that explains it.
The same arithmetic applies on the value side, which almost nobody models. A Shopping campaign with a 300% Target ROAS actually returning 600% on a fixed $2,000 a month produces about $12,000 in revenue. Held to 300%, that $2,000 produces about $6,000. Nothing in your budget dashboard flags a 50% revenue cut.
Why does a direct response lens change the answer here?
Because direct response says the target was never a performance record. It was your allowable cost per acquisition, derived from margin and close rate, before you ever opened Google Ads.
That is why “match your recent actuals” — the default answer, and the one Google’s tool clicks for you — is only correct by coincidence. Recent actuals describe your budget ceiling, not your business. The right target is what you can afford to pay and still make money on the transaction. Sometimes that equals your actual. Often, it is higher, meaning the efficiency you enjoyed was driven by underbought demand rather than profit.
Freelance Google Ads manager Joey Bidner argued on LinkedIn that some of his best accounts run loose targets deliberately, to give Smart Bidding room to explore. For accounts with volume to support exploration, fair. For a three-person team at $50 a day, it almost never describes what is happening: the target is old, not deliberate.
What should you actually do before Monday?
Seven steps, an afternoon’s work for a small account. In order:
- Open the Bid Target Adjustment Tool. It went live in accounts on July 6, 2026, and notification emails went out around July 2. It shows each flagged campaign’s recent actual, its current target, and a recommendation. Read the list; do not press Apply yet.
- Widen your check to twelve months. Google’s “Limited by budget” flag looks back a full year, so a campaign that only hit its cap during a Q4 push is still on the list. Don’t trust today’s status column.
- Keep only Target CPA, Target ROAS, and Demand Gen Target CPC campaigns. Everything else comes off the list.
- Write down your allowable cost before you look at any actual. Lead gen: average job value × gross margin × lead-to-sale close rate. Ecommerce: average order value × contribution margin. Doing this second, after seeing the actuals, guarantees you anchor to the wrong figure.
- Compare, then decide. If allowable cost sits above your recent actual, match the actual for now — protect current volume, save the headroom question for a calmer week. If allowable cost sits below your actual, you have a bigger problem than August 17 and should be fixing the offer, not the bid.
- Handle thin campaigns manually. Below roughly seven conversions Google will not calculate a recommendation at all. Those still need a number typed in. If you cannot justify one, switch to Maximize Conversions and take volume over predictability.
- Log the change. Old target, new target, date, one line of reasoning. In four weeks this is the difference between diagnosing what happened and guessing.
One correction to the usual advice: raising budget while keeping the target is a better option after Monday than before. The volatility that used to punish budget increases came from exactly the borrowing this update ends. If your target is right and demand sits above your cap, more budget should now scale at the stated target instead of drifting off it — the one piece of this change working in a small advertiser’s favor.
Which dates matter between now and October?
The deadline is not the only date that shapes what you should do.
- August 14–16 — the last window to set targets deliberately rather than reactively. Google says matching actuals ahead of the deadline should not cause noticeable volatility.
- August 17 — enforcement begins, staggering across accounts over the following weeks. A quiet Monday does not mean you were skipped.
- August 17–31 — Performance Planner forecasts will be unreliable, per Google. Do not build a Q4 budget case on them.
- Late September — set a reminder now to re-audit. Targets set under deadline pressure deserve a second look with post-change data.
- October onward — Q4 auction pressure pushes more campaigns into budget-limited status, so campaigns unaffected in August become affected in November.
How long before you can judge what happened?
One to two conversion cycles, which is Google’s guidance and longer than it sounds. For a lead-gen business where enquiry to signed job runs two to three weeks, that is four to six weeks before the bid strategy report shows the change rather than the re-learning noise around it.
For the first three days, watch and change nothing. Smart Bidding reacts to target edits in real time, so stacking a fix on a system-wide change makes the cause of what happens next unknowable. Google also advises against data exclusions or new bid limits here, calling that “not recommended.” Save exclusions for genuine tracking outages: this is a targets problem, and treating it as a data problem adds a second variable to a window that needs to stay clean.
Managing several client accounts changes the sort order, not the steps: work by revenue at risk, and check any contract carrying a committed CPA or ROAS figure against the list. A client whose commitment quietly depended on a campaign over-delivering is a conversation for this week. Keeping a running target-versus-actual record is the sort of standing check CampaignPress is built to make routine.
Frequently asked questions
My campaign has never been limited by budget. Am I affected?
Probably not on August 17, but check the twelve-month lookback rather than today’s status column. Google’s flag catches campaigns that were budget-constrained at any point in the past year, including a brief seasonal spike. Campaigns that have genuinely never hit their cap already behave this way — unconstrained target campaigns have always been held to the entered target.
Should I just press Apply in the Bid Target Adjustment Tool?
It preserves current performance and beats doing nothing. But Apply sets your target to recent actuals, which describes your budget ceiling rather than a business decision. Spend ten minutes calculating allowable cost first. If the two numbers agree, apply with confidence. If they do not, type your own.
What if I miss the deadline entirely?
Nothing breaks, and Google will not change a target or budget for you at any point. Your campaigns simply start delivering toward whatever number is in the field. You can adjust afterwards, but you will be doing it while the account is re-learning, which makes it hard to separate your change from the update. Fixing it before Monday is cheaper than diagnosing it after.
Does this apply to Microsoft Ads or Meta?
No. This is a Google Ads and Search Ads 360 change, plus Demand Gen in Display & Video 360. Neither Microsoft Advertising nor Meta has announced equivalent behaviour. If you run a single-platform Google account, though, that concentration is the real exposure: one bidding policy change moves your entire acquisition channel at once.
The gap between target and actual was information all along: it told you the budget was binding, not the bid. Monday, Google stops printing it in the CPA column. The accounts that come through well will be the ones whose targets were real numbers to begin with — more breakdowns like this one daily at CampaignPress.ai.

2 Comments