The Free Lunch Is Ending: What Google’s August 17th Smart Bidding Change Really Means

On 17 August 2026, Google is making a change to how target-based Smart Bidding behaves when campaigns are limited by budget. On the surface, it sounds like a minor technical update. In reality, it exposes one of the most misunderstood parts of modern Google Ads management: the relationship between targets, budgets and commercial performance.

For years, many advertisers have benefited from what I would call “borrowed efficiency”. A campaign may have been set to a Target CPA of £85, but because the campaign was constrained by budget, it was actually delivering at £45. Or a campaign may have been set to a Target ROAS of 400%, but in reality it was returning 800%. To most teams, that looked like overperformance. The campaign was “smashing targets”. The agency could report a win. The client could feel confident that the platform was working harder than expected.

But in many cases, the campaign was not outperforming because the target was particularly smart. It was outperforming because the budget cap was doing part of the job.

When a campaign is budget constrained, the system has to ration spend. It cannot enter every auction it may otherwise want to enter, so it becomes more selective. It stretches the available budget across the day, avoids some lower-value opportunities, and naturally ends up focusing on a smaller pool of higher-quality traffic. The result is a campaign that often beats the target you set, not because the target reflected true commercial reality, but because the budget restriction created a hidden layer of efficiency.

That is the free lunch. And it is about to become far less reliable. This Smart Bidding Update changes how we plan

Google Ads free efficiency is about to end on August 17th

August 17th 2026, Google’s target-based bid strategies are expected to deliver more consistently towards the targets advertisers have actually set, even when campaigns are limited by budget. In plain English, if you tell Google that a £85 CPA is acceptable, the system will behave more like £85 is acceptable. If you tell Google that a 400% ROAS is the target, the system will behave more like 400% is the goal. That may sound obvious, but for many advertisers it will create a visible shift in performance. Campaigns that have been comfortably beating their targets may begin to move closer to the targets that were originally entered into the platform.

This is not necessarily a problem. In fact, I think it is the right direction. But it does force a much more grown-up conversation about how targets are set in the first place.

The uncomfortable truth is that many Google Ads targets are not commercial targets. They are inherited targets. They are old targets. They are testing targets. They are numbers that were set months or years ago because the platform needed an input. They are often based on media performance rather than business performance. Someone looked at a CPA or ROAS number, decided it felt reasonable, and the account continued to optimise around it. That was survivable when budget constraints were quietly adding an extra layer of efficiency. It becomes far more exposed when the system starts taking your target at face value.

This is where advertisers need to stop treating Target CPA and Target ROAS as simple platform settings. They are not just bidding inputs. They are commercial instructions to the machine. They tell Google what level of efficiency you are willing to accept. They define the trade-off between volume and profitability. They tell the system where to stop. And in an AI-led media environment, where manual control is becoming less precise, those instructions matter more than ever.

We have already moved a long way from the old world of Google Ads. Keywords are less exact. Match types are looser. Performance Max has absorbed more inventory. Demand Gen is pushing advertisers into broader, more discovery-led environments. Creative, feed quality, first-party data, conversion value and business signals are becoming more important than manual bid management. In that world, the target becomes one of the most important control layers left.

If the target is wrong, the machine can still optimise brilliantly. It will just optimise brilliantly towards the wrong outcome. That is why the August change should not be seen as a small housekeeping task. It should be treated as a commercial reset.

Every advertiser using Target CPA or Target ROAS should be asking whether their targets reflect how the business actually makes money. Is the CPA based on margin, conversion rate, close rate, stock position or lifetime value? Is the ROAS based on revenue, or does it account for profitability? Does the same target make sense across all products, markets and customer types? Are new customers being valued differently to returning customers? Are high-margin products being treated differently to low-margin products?

These are the questions that separate media optimisation from business optimisation.

For ecommerce brands in particular, ROAS is no longer enough. A 600% ROAS can still be poor if the product margin is thin, return rates are high or discounting is heavy. A 250% ROAS can be strong if the product has healthy margin, strong repeat purchase and meaningful lifetime value. This is why I believe and have preached for many years, advertisers need to move closer to POAS, or profit on ad spend, as the commercial anchor. It is not always easy to calculate perfectly, but even an imperfect margin-based target is usually better than a blended ROAS target that ignores profitability completely.

The same principle applies to lead generation. A cheap lead is not always a good lead. A high CPA is not always bad if the lead quality, close rate and contract value justify it. Target CPA only becomes meaningful when it is connected to what happens after the conversion. Otherwise, you are optimising to a form fill, not a business outcome.

Targets become the budget control layer.

This also changes how we should think about budget. Too many accounts are still managed with budget as the primary control lever. The conversation starts with “how much can we spend?”, "what is our budget for the month?", rather than “how much profitable demand exists?” That mindset needs to change. If a campaign can scale at a commercially acceptable target, budget should not be the main constraint. Budget should flex to profitable demand. The target should define the conditions under which that demand is worth buying.

This is also where the finance conversation becomes easier, not harder. Instead of asking for more budget because a campaign is “limited by budget”, marketers should be able to say: “At this target, every £1 invested is expected to return £X within this margin band. If demand exists at that level, we should allow the campaign to scale.” That is a much stronger conversation. It moves paid media away from being seen as a cost centre and towards being understood as an investment model.

So, what should advertisers do now?

Start by identifying every campaign using Target CPA or Target ROAS that has been limited by budget. Then compare the target you set with the actual performance you have been receiving. If a campaign has a £85 Target CPA but has been delivering a £45 CPA, do not simply celebrate the gap. Ask whether £85 is genuinely acceptable. If £45 reflects the real commercial target, update the target. If £65 is the right number based on margin and growth ambition, use £65. If £85 is genuinely acceptable because the business wants more volume and can afford that level of efficiency, then leave it as it is and prepare for performance to move closer to that number. Noting, that you will need more money to support that target.

The worst option is doing nothing and then being surprised when the platform starts delivering closer to the target you told it was acceptable.

What Google is doing to support advertisers

To Google’s credit, this is not a change being pushed through without any support for advertisers. From 6 July 2026, advertisers should start seeing notifications inside the Google Ads interface that guide them towards a new Bid Target Adjustment Tool. This tool is designed to help advertisers review historical campaign performance and quickly apply updated targets where needed.

That distinction is important. Google is not automatically changing your budgets or your bid targets for you. The responsibility still sits with the advertiser to decide whether the current target is commercially correct. However, the tool should make it easier to identify campaigns that may be affected, particularly those that have been limited by budget while using Target CPA or Target ROAS and have historically performed better than the target set in the account. In practical terms, this means advertisers should not wait until 17th August to react.

From 6th July, the first step should be to review the recommendations surfaced in the Google Ads UI, compare them against actual business targets, and decide whether to apply, adjust or reject the suggested changes. If the current target reflects real commercial goals, no change may be needed. But if the campaign has only been beating target because budget constraints were creating hidden efficiency, this is the moment to reset the target before the system starts optimising more consistently towards the number you originally gave it.

The key point is this: Google can surface the data and make the adjustment easier, but it cannot decide your margin, your payback period, your stock priorities or your appetite for profitable growth. That is still the job of the advertiser.

This change is undramatic by design, which is exactly why it matters. It will not look like a major platform shift. It will not require a new campaign type, a new pixel, a new feed or a new creative format. But it will quietly expose whether your bidding strategy is connected to real commercial logic.

Good advertisers will treat this as a settings review. Great advertisers will treat it as an opportunity to rebuild the relationship between media, margin and growth.

The free lunch is ending. But that is not a bad thing. It means the machine is becoming more predictable. It means scaling decisions can become cleaner. It means advertisers can stop relying on hidden efficiency created by budget caps and start building targets that reflect how their businesses actually grow.

The next era of Google Ads will not be won by the people who know which buttons to press. It will be won by the people who know what numbers, signals and creative to give the machine in the first place.

So before 17 August, the question is not simply whether your campaigns are beating target.

The better question is whether the target was ever right in the first place.