Google Ads is about to hold you to your targets

For years there has been a quiet gap in a lot of Google Ads accounts between the target an advertiser entered and the cost per conversion they actually got. Set a £40 target CPA on a campaign that is capped by budget, and Smart Bidding would often spend that capped budget on the cheapest auctions it could find and hand back £20. Free efficiency. Nobody complained.

That gap is closing on 17 August 2026. Budget-limited campaigns on target-based strategies will be pushed to deliver toward the target you actually set. Same budget, same total spend, higher cost per conversion, and for many campaigns fewer conversions from the same money.

01What is actually changing

Google's position is that this is a predictability fix. Rather than concentrating a constrained budget on the cheapest slice of available demand, Smart Bidding will spread it across the full range of auctions that still clear your stated target. That gives you a number you can plan against, and it means raising the budget produces something closer to a straight line instead of a surprise.

Three things to be clear on:

• It only bites when you are limited by budget. Google has confirmed in its FAQ that Target CPA and Target ROAS campaigns which are not budget-constrained will not change behaviour.
• It does not increase your spend. Budgets are not being raised and Google will not adjust your targets or budgets for you. What changes is what that spend buys.
• The strategy names changed too. "Maximize conversions with a target CPA" is now just Target CPA, and "Maximize conversion value with a target ROAS" is now Target ROAS. That is cosmetic. The logic underneath is the same.

Affected: Search, Shopping, Performance Max, Demand Gen and Travel. Display and Hotel campaigns already worked this way. Untouched: Manual CPC, Target Impression Share, App campaigns and Video reach and view strategies. Portfolio strategies and shared budgets are included, but any target changes have to be made at the portfolio or shared budget level, not on the individual campaigns underneath. Google's help documentation has the full scope.

02Find your exposure in ten minutes

Google has a Bid Target Adjustment Tool that has been appearing in campaign settings for flagged accounts since 6 July. It is useful, but do not let it be your audit. It only surfaces where Google decided to flag it, and its one-click option does a single thing: reset your target to match recent actuals.

Do the pass yourself instead:

• Filter every campaign on Target CPA or Target ROAS.
• Check which of those have carried a Limited by budget status, or meaningful impression share lost to budget, at any point in the last twelve months. Intermittent counts.
• For each one, put actual CPA or ROAS next to the entered target over both a 30 and a 90 day window. Two windows, because one seasonal month will lie to you.
• Rank by the size of the gap multiplied by monthly spend. That is your real exposure, and it is usually two or three campaigns rather than the whole account.

A campaign delivering £18 against a £20 target does not need your attention. One delivering £22 against a £60 target on £3,000 a month is the entire problem.

03Work out what the target should have been

This is the part that gets skipped, and it is the only part that really matters. The uncomfortable truth in most flagged campaigns is that the target was never a considered number. It was a placeholder from launch, or a figure nudged upward once to escape a learning period and never revisited.

The target should come from unit economics, not from last month's report. For lead generation, work back from the money: gross profit on a closed deal, multiplied by the rate at which leads become deals, gives you the cost per lead at which you break even. Then set your target below that by however much margin you intend to keep. For ecommerce, break-even ROAS is one divided by your effective margin after shipping, fees and returns, and your target sits above it.

Do the maths and one of two things happens. Either your actual performance turns out to be roughly where the business needs it, in which case you now have a defensible target to enter. Or you discover the campaign has been running for a year on a target nobody could justify, and this change simply made that visible.

04Pick a route, not a panic

That last option deserves a warning. Switching to Maximise Conversions sidesteps enforcement, but you are trading an enforced number for an unconstrained one. On a small budget with thin conversion volume, that can be a worse outcome than the change you were avoiding.

05Move in steps, and only move one thing

Whatever you land on, do not jump there in one edit. Large target changes restart learning and you lose a fortnight of clean data working out what you did. Move in increments of 10 to 15 per cent and let one to two conversion cycles pass before the next one. On a business with a three week sales cycle, that is not three days.

And change one variable at a time. A target adjustment, a budget increase and a structural change in the same week leaves you with no way to attribute the result. If you run seasonal peaks, handle those through Google's seasonality controls rather than yanking the target around and back again.

06Tell the client before the number moves

The rollout starts on 17 August and staggers across accounts over the following weeks, so this is not a switch flipping everywhere at once. Which means the visible effect will land somewhere in late August or September, right in the middle of a reporting period.

A cost per conversion that rises from £22 to £38 reads as a failure if the first anyone hears of it is in the monthly report. It reads as competence if you flagged it a month earlier and explained why. Annotate the account with the date you make each change, note the baseline, and put a line in this month's report now. That single paragraph is worth more to the relationship than the optimisation itself.

Then watch the right things afterwards: cost per conversion against the new target, conversion volume, impression share lost to budget, and average CPC. Give it a few days before reacting to anything. The first 72 hours of any bidding change are noise.

---

← Back to all articles