TLDR: From 17th August 2026, budget-limited Target CPA and Target ROAS campaigns (plus Target CPC on Demand Gen) will track their bid target far more closely, even when the campaign is capped by budget. If a campaign has been outperforming its target for months, that gap is about to close. This guide walks through what’s changing, how to work out whether your targets are right, and what to do before the deadline.
What’s changing and who this affects
Right now, a budget-constrained campaign doesn’t always deliver to its stated target. Raise the daily budget on one of these campaigns and the CPA can move in ways nobody planned for, because the algorithm has been quietly borrowing headroom from the target to make the old budget work. From 17th August, that stops. Budget controls spend and the target controls efficiency. Two separate levers, instead of one leaking into the other.
Search, Shopping, Performance Max, Demand Gen and Travel are all affected, running in Google Ads or Search Ads 360. Target CPC is included too, but only inside Demand Gen, which is easy to miss in the noise around this update. Display and Hotel already behave this way, so there’s nothing to change there. App, Video reach and Video view are exempt outright. Manual CPC and Target Impression Share aren’t touched at all.
Google has put a Bid Target Adjustment Tool into accounts to help with the transition. It shows recent performance next to the current target, alongside a recommended figure, and a single Apply button that sets the target to whatever it recommends. It’s worth being clear about what that tool does and doesn’t do: it does the sums and the clicking, but it doesn’t decide anything for you. If the old target reflected a genuine cost ceiling rather than a number nobody had revisited, typing in your own figure is the better move, and the tool has no way of knowing that.
The question worth answering before you touch anything
A gap between what a campaign delivers and what its target says can mean one of two things. Either it’s deliberate headroom, built in on purpose so Smart Bidding has room to explore and find new customers, or it’s an old number nobody’s looked at since it was set. The size of the gap won’t tell you which one you’re dealing with. Only checking the target against your business numbers will.
Take the tCPA and compare it to what you can genuinely afford to pay for that conversion. Take the tROAS and check it against real gross margin, not against whatever figure has been making the client report look good. A target that survives that test was probably deliberate. One that doesn’t is the second kind, and this update just gave you a reason to fix it now rather than later.
Five ways to choose the right adjustment
Once you know which kind of gap you’re looking at, there are five sensible moves:
Keep the target
But only if you can say why that specific number is the goal, not just that it’s been producing good results. If this is your call, give the campaign more budget headroom so the target stops being the tighter constraint.
Match the target to recent actuals
If the gap was never deliberate. This is the default move for most accounts, and it’s Google’s own recommendation too. Move gradually rather than in one big jump if you’re nervous about volatility. Google says a single clean adjustment should be safe; some practitioners are choosing to nudge targets in smaller steps and watch each one before going further. Both can be right, depending on the account.
Set a custom target from real unit economics
If even recent actuals aren’t the right number, because the original figure was inherited rather than checked against margin or lifetime value.
Move to Maximise Conversions or Maximise Conversion Value
If your budget is genuinely fixed and scale matters more than a predictable unit cost right now.
Raise the budget and keep the target
If the target is right and there’s demand above your current cap. Before 17th August, this was a gamble, because increasing budget could send efficiency in unpredictable directions. After the change, that risk drops. For accounts that have been avoiding budget increases because of the volatility they used to cause, this is the one genuinely good piece of news in the whole update.
One honest caveat: resetting to today’s number is a snapshot, not a permanent fix. What’s an efficient CPA at today’s volume won’t necessarily hold once the market shifts, competition increases, or quality scores drop. A looser target used to absorb that drift automatically. Now someone has to notice it and adjust the number by hand, which is exactly why a one-off reset was never going to be the end of this.
Give it time before you judge it
Whatever you change, don’t judge it too early. Google’s own guidance is to wait one to two conversion cycles before reading the bid strategy report, and to treat Performance Planner forecasts as unreliable between 17th and 31st August. For a lead-gen account with a two to three week sales cycle and weekly offline conversion imports, “one to two conversion cycles” can mean four to six weeks, not a few days.
A rough shape for that window:
- Days 1 to 3: watch, don’t touch. Look for anything unexpected, but resist adjusting further while the system is still re-learning.
- Days 3 to 14: judge by conversion cycles rather than the calendar, using the bid strategy report rather than Performance Planner.
- Weeks 2 to 4: fix whatever you got wrong under deadline pressure, one change at a time.
- Week 4 onward: this becomes a standing check, not a finished task. Performance moves, the gap reopens, and once it does, it’s not free efficiency any more. It’s a setting that’s gone quietly stale again.
Google’s guidance is also to leave data exclusions and bid limits out of this. This is a targets problem, not a data-quality one, and stacking a second variable onto a measurement window that needs to stay clean just muddies what actually happened.
Auditing this properly, without doing it campaign by campaign
For agencies and accounts with more than a handful of affected campaigns, checking each one by hand doesn’t scale. The sequence still holds, though:
- Pull every campaign that’s carried “Limited by budget” status, looking back further than the last few weeks.
- Keep only the affected strategies: Target CPA, Target ROAS, and Target CPC on Demand Gen.
- Compare actual performance against target using a stable window, 30 days for high-volume campaigns, 60 to 90 for thinner ones.
- Rank by money at risk rather than the size of the gap. A 20% gap on £80,000 a month matters more than a 2x gap on £400.
- Check whether the target was ever real, against allowable cost or margin, not against a number that’s been making the reporting look good.
- Decide, then apply, using the options above.
- Write down what changed and why. Old target, new target, date, one line of reasoning. In four weeks’ time, that’s the difference between diagnosing a problem and guessing at one.
Portfolio bid strategies work differently here. Any adjustment has to happen at the portfolio or shared-budget level, not on the individual campaigns inside it, and Google’s own tool is the one built for that job.
The bit that matters more than the deadline
Google’s Bid Target Adjustment Tool solves the problem sitting in front of you right now. It doesn’t solve October, when a different set of campaigns starts hitting their caps because seasonality shifted or someone raised a daily budget without rechecking the target that went with it. Keeping targets honest across a lot of campaigns, without hand-picking a number every time performance moves, was never really a problem with a deadline attached. This update just made it visible.
That’s the piece worth carrying forward: not the one-time fix, but building a repeatable check into how targets get reviewed, so the accounts that drift again aren’t the ones nobody was watching.
If you’d rather have someone else keep an eye on this across your accounts, and turn it into standard reporting rather than a one-off scramble before a deadline, get in touch with Modo25 to find out what that looks like.
