Do you run uncapped budgets? Campaigns that spend to the end every single day and never hit a ceiling? Then you can close this article. The change in question doesn't affect you.
Do you and your clients keep hitting the budget ceiling? Then read on. In a month your account economics may shift, and Google will send you a notification that's easy to click away.
What's changing
On August 17, 2026 Google adjusts how target-based bid strategies behave. Specifically:
- It affects target cost per acquisition (tCPA) and target return on ad spend (tROAS). For Demand Gen, also target cost per click.
- Campaign types hit: Search, Shopping, Performance Max, Demand Gen, Display, Hotel and Travel. Including accounts managed through Search Ads 360.
- It does not affect App campaigns, Video Reach, Video View, target impression share, target CPM, or manual CPC.
- And crucially: it does not affect campaigns that aren't limited by budget.
The core is simple. Today a budget-limited campaign often beats its target. You set a target cost per acquisition of $20 and conversions actually come in at $10. From August 17, such a campaign will run closer to the $20 you entered.
The bulk target-adjustment tool (Bid Target Adjustment Tool) has been in accounts since July 6. Google stresses it won't change your targets or budgets on its own — and that your budget caps still hold.
Why your campaign beats its target today
Nobody really explains this, and yet it's the whole point.
When a campaign is limited by budget, it means auction demand exceeds the money you have. The system has to choose which auctions to even enter. And so far it has chosen conservatively — going mainly into auctions where it was most confident of a conversion, so the budget would last.
The result? Real performance came out better than what you entered. Not because your target was set well, but because the budget quietly overrode it.
Which leads to an uncomfortable conclusion: on budget-limited campaigns, your target has been mostly decoration for years. Real performance was decided by the budget. Anyone running the account by budget and leaving the target alone was pulling a lever that wasn't connected. From August 17, it will be.
Google's argument for the change is actually fair. Today you raise the budget on a campaign that ran at $10 per conversion, and it jumps to $20. You don't understand why, because all you did was add money. Now you'll know upfront what you're running.
Before you panic: three things have to line up
There's plenty of noise around this change. Peers abroad call it a self-serving move that pushes advertisers toward higher budgets. The opposite view also shows up — that the system today leans too hard on safe auctions and ignores a wider set of opportunities that still fit inside your target.
Before you pick a side, notice that an affected campaign has to meet three conditions at once:
- It has a target set. Without tCPA or tROAS there's nothing to enforce.
- It's genuinely limited by budget. Not "losing 0.3% of impressions," but actually throttled.
- It beats the target. A campaign running below its target has nothing to lose on August 17.
Drop any one of them and the change doesn't touch you. And that's a rarer combination than the headlines suggest.
I ran 42 accounts and did the math
Instead of guessing, I went through my entire portfolio. A 28-day window (June 30 – July 27, 2026), the last three days dropped for conversion lag. To keep dead campaigns from skewing the numbers, I counted only those spending at least 1,000 CZK (~$45) over those 28 days.
The funnel came out like this:
| Step | Campaigns |
|---|---|
| Campaigns spending ≥ 1,000 CZK / 28 days | 42 |
| … has a target set | 16 |
| … and is genuinely budget-limited | 9 |
| … and is beating that target | 7 |
Seven campaigns out of forty-two. In money: 118,648 CZK out of 476,370 CZK spent, or 24.9%. Five accounts out of sixteen active ones.
Let me dwell on the second row from the bottom, because everything rests on it. "Budget-limited" cannot be read off a single number. Budget-lost impression share is a Search Network metric — for Performance Max it covers only the search slice, and for Demand Gen it isn't returned at all. Rely on it alone and your PMax campaigns drop out of the result, even though they hit the budget cap every single day.
So I use three signals, and any one of them counts: budget-lost impression share, the campaign status Google itself reports (BUDGET_CONSTRAINED), and budget utilisation — spend against daily budget over the period. Three of the seven affected campaigns are PMax with budget-lost impression share below five percent or at zero — and Google still flags them as budget-constrained while they burn 99 to 100% of their daily budget.
One more note on the last row, because somebody will ask: at what level of overshoot did I count a campaign in? Google gives no threshold, so it's my call — and the good news is that it doesn't matter here. The smallest overshoot among those seven campaigns is 37%. Set the bar anywhere between ten and thirty-five percent, you always get seven. You only inflate that count if you label as "budget-limited" a campaign that loses a tenth of a percent of impressions to budget. That's statistics, not reality.
An honest caveat: this is one portfolio, mostly Czech small and mid-sized e-commerce stores and services. Forty-two campaigns is not a representative sample of the market, and large accounts may look different. But nobody has published anything better — every article I read on the topic just retells Google's announcement without a single number of their own.
One of them hurts the most
If you're tempted to file this under "storm in a teacup," look at the worst of the seven.
It's a category Search campaign for one e-shop:
- Target return on ad spend set to 1.0, i.e. 100%
- Real return 3.46 — beating the target by 246%
- Losing 8.8% of impressions to budget, daily budget ~$45
- Over 28 days: spend 11,555 CZK, conversion value 39,957 CZK
This is a textbook case of a common tactic: set the target return low as a safety net, so it runs almost like value maximization, just with a brake. The budget throttles it anyway and the system picks the best auctions for me.
This exact tactic stops working on August 17. Leave the target at 1.0 and the system starts honestly aiming at a return of 1.0. Upper-bound impact: conversion value drops from 39,957 CZK to 11,555 CZK over 28 days. That's 28,402 CZK over four weeks, roughly 370,000 CZK (~$16,000) a year.
That estimate is deliberately the worst case — it assumes a full slide to target at unchanged spend. In reality it'll be milder. But the direction is certain, and the fix takes five minutes.
So no, it's not a storm in a teacup. It's a narrow but deep problem. Which is exactly the kind that gets overlooked.
What to do about it
First, take a snapshot of the current state. Today, not August 16. If you don't freeze the numbers before the change, you'll have nothing to compare against afterward and you'll just be guessing whether a swing was Google, the season, or you. Target versus reality, cost per click, volume, conversion value and impression share for each campaign is enough.
Then find your affected campaigns. Filter: "Limited by budget" status combined with tCPA or tROAS. Ignore campaigns where budget-lost impression share is in the low single-digit percents.
Then decide for each one separately:
| Situation | What to do |
|---|---|
| Gap between target and reality under 10% | Nothing |
| Large gap, budget is a hard ceiling | Align target to reality — preserves today's economics |
| Large gap, economics support volume | Keep the target and add budget. This is the opportunity |
| Economics have changed since | Recompute the target from margin, not history |
| Volume matters more than exact cost | Consider Maximize Conversions or Conversion Value |
Don't rely on Google's tool blindly. It recommends a target derived from your history, not your margin. Clicking "Apply" means letting Google bless a number you never derived from economics. For a campaign that's supposed to be profitable, that's the wrong move.
And do it with a buffer. The system needs one or two conversion cycles to settle. Changing targets on the evening of August 16 is the worst option. And between August 17 and 31, don't plan from planner forecasts — Google itself warns they'll be temporarily unreliable.
The takeaway
Most accounts won't be directly affected. That's exactly why so many people will click the notification away — and the minority that is affected will take it full force.
Anyone with targets derived from margin is looking at twenty minutes of work. Anyone with eyeballed targets, or none at all, finds out in September from the P&L.
And if, while reading this, you realized half your campaigns have no target at all, then your problem isn't August 17. You've had that problem already — the budget was just kindly hiding it from you.
