How Google's Smart Bidding Changes Should Shift Your Budget Strategy

If you advertise with Google Ads and run Target CPA or Target ROAS campaigns, something significant is about to break your account, and most advertisers won't see it coming until it's too late.

On August 17, 2026, Google is removing bid suppression for budget-limited campaigns. It sounds like a routine platform update. It isn't. For the past two years, budget-constrained campaigns running Smart Bidding have been quietly outperforming their stated targets, delivering efficiency that looked great in reports but was largely an artifact of how suppression worked behind the scenes. That hidden efficiency is disappearing, and the budget floors and bid guardrails you've built your account around were calibrated to a reality that no longer exists.

This piece breaks down exactly what changed, why it invalidates the assumptions most e-commerce and SaaS advertisers have been running with, and what you should actually do about it before the deadline hits. We'll cover how to diagnose your exposure, recalibrate your targets the right way, rethink your minimum budgets, and manage the learning phase through the transition, including what this means for Q4 planning.

What Google Quietly Changed and Why It Matters

If you've been running Target CPA or Target ROAS campaigns on a limited budget, Google's algorithm has been quietly doing you a favour. When a campaign was budget-constrained, the system suppressed bids to stretch that budget further, delivering results meaningfully better than your stated target. Your campaign said "acquire at a certain CPA." Google was actually acquiring well below that figure. That gap is bid suppression, and most advertisers never saw it documented anywhere.

That changes on August 17, 2026. From that date, campaigns using Target CPA or Target ROAS bidding will deliver toward your stated target rather than the suppressed actual. If you haven't adjusted your targets to reflect what was really happening, your performance numbers will shift.

Affected campaign types are Search, Shopping, Performance Max, Demand Gen, and Travel. App campaigns, Video Reach, and Video View are untouched, so accounts weighted toward those formats have less immediate exposure.

Google's stated reason is improving predictability when you scale budget. That framing is an implicit admission the gap was real enough to cause genuine performance fluctuation whenever advertisers increased spend on a suppressed campaign.

Notifications started rolling out on 6 July 2026, and affected accounts received alerts inside Google Ads along with access to the Bid Target Adjustment Tool, covered in detail later. If you haven't checked your account for those alerts, do that before reading further.

This matters well beyond August 17. If you're heading into Q4 planning or thinking about how to run a January sale campaign that actually converts, the performance baselines you've been using as reference points are built on suppressed data.

Two Years of Budget Assumptions Just Broke

The problem with the efficiency you've been reporting to stakeholders is that a meaningful chunk of it was never yours to keep.

That hidden efficiency gap, two years in the making, didn't just distort your reports; it distorted every budget floor built from them.

That created a false floor. Advertisers looked at delivered CPA, said "right, that's what this channel costs," and set minimum budgets accordingly. The number felt earned. It wasn't. It was partly structural, a byproduct of the suppression mechanism itself, and always fragile.

The failure mode is obvious in retrospect: adding budget relieved the constraint, suppression lifted, and the campaign started spending toward its stated target rather than the suppressed actual. The benchmark was never achievable at scale because scale was precisely what killed it.

This is why the August 17 change doesn't just affect live campaigns. It retroactively invalidates the guardrails built around two years of suppression-inflated data, particularly for e-commerce accounts that locked in Q4 budget plans based on those benchmarks. (If you're also seeing platform-level efficiency shift on paid social, the performance breakdown happening across Meta Ads in 2026 is worth reading alongside this.)

SaaS advertisers are the most exposed. Thin margin structures leave almost no room between a real CPA and an unsustainable one, and if your Target CPA was calibrated against suppressed delivery, that gap may have already closed without you noticing.

Diagnosing Your Exposure Before August 17

So now you know the assumptions were wrong. The next question is which of your campaigns actually has exposure, and how much.

Start in the Campaigns view and add the Budget status column. Filter for any campaign flagged as "Limited by budget" that is also running Target CPA or Target ROAS. Those are your candidates. Use the Bid Target Adjustment Tool (surfaced with July 6 notifications) to review queued recommendations, then run the core diagnostic: a 90-day comparison.

Pull your stated Target CPA or Target ROAS alongside actual delivered CPA or ROAS for the same period on each affected campaign. The gap between those two numbers is your suppression gap. A campaign with a materially lower delivered CPA than its stated target is not evidence of a well-tuned algorithm; it is evidence of suppression doing the heavy lifting.

The wider that gap, the higher your exposure after August 17. Large-gap campaigns have been leaning on suppression rather than genuine efficiency, which means their real cost structure has been hidden.

When the Bid Target Adjustment Tool surfaces a recommendation to lower your Target CPA or raise your Target ROAS, accept it, even though it feels like you are conceding ground. You are not. You are aligning the stated target with where the algorithm has actually been operating.

Finally, segment this analysis by campaign type before drawing conclusions. In my broader look at what's actually working in Google Ads right now, suppression behaviour varies noticeably across formats. A Shopping campaign and a Search campaign in the same account can carry very different gap sizes, and treating them as equivalent will skew your prioritisation.

How to Recalibrate Your Target CPA and ROAS the Right Way

With actuals as your baseline (not stated targets, that gap was covered above), the recalibration logic differs by account type.

For SaaS accounts, if a campaign was set to a higher Target CPA but consistently delivered well below that figure, the real efficiency ceiling sits just above the delivered figure. Setting the new target at the original stated number post-change gives the algorithm far too much room to spend inefficiently once suppression is removed. The stated target needs to be close enough to the actual delivered figure that the algorithm's mandate is accurate, not generous.

For e-commerce, the logic is inverted. A campaign delivering materially higher ROAS than its stated target means the algorithm has been holding back spend because the stated target is too conservative. The fix is tightening the stated ROAS target toward the delivered figure, the same principle covered when examining how ROAS discipline affects margin-aware campaign performance.

One critical constraint: adjusting targets triggers re-entry into the learning period, which now requires around 50 conversion events to complete. Stagger adjustments across campaigns rather than updating everything simultaneously.

The objective is alignment, not aggression. You are not pushing targets lower or higher indiscriminately; you are giving the algorithm an accurate mandate. That framing also matters for stakeholders, since conversion rate assumptions built into your targets may need revisiting alongside the CPA and ROAS numbers themselves.

Rethinking Your Minimum Budget Floors for Affected Campaigns

If a campaign was converting at a suppressed CPA but your stated target was materially higher, your daily budget was calculated against the wrong number. Post-August 17, that same budget now needs to cover conversions at something closer to your stated target. The same spend buys fewer conversions, which directly threatens Smart Bidding stability.

Here's a concrete example. Say you're running a Shopping campaign where actual delivered CPA has been well below the stated Target CPA. Your floor was set assuming more conversions per day than the real post-suppression CPA will support. At the true post-suppression CPA, that same budget may yield significantly fewer daily conversions, likely below the threshold needed to keep Smart Bidding out of learning.

That threshold interaction is the compounding risk. Campaigns that drop below the minimum conversion volume re-enter the learning phase, and if that happens across multiple campaigns simultaneously during the transition window, you're managing instability at scale. I'd flag this in paid ads fundamentals after the cost inflation era as one of the more predictable failure modes when platform mechanics shift.

To be clear, this change does not increase what you pay per click. The google ads cost structure at the platform level is unchanged; what changes is that the real cost per result for suppression-dependent campaigns surfaces correctly for the first time.

For new budget floor calculations on campaigns with a meaningful suppression gap, build in a meaningful CPA buffer on top of your actual delivered CPA (the right size depends on your margin structure and account history), then recalculate the daily budget needed to hit your conversion threshold. Treat the initial weeks post-August 17 as a re-baselining period, not a performance optimisation period.

Learning Phase Management After the Change

The harder sequencing problem is account-wide learning phase exposure.

August 17 creates pressure to act on multiple campaigns at once. If you update targets across five or six campaigns in the same week, you can push the entire account into learning simultaneously, compounding instability rather than containing it. The fix is straightforward: adjust your highest-exposure campaigns first, meaning those with the largest gap between stated and actual targets. Work through medium and low-exposure campaigns on a staggered schedule across two to three weeks.

A campaign that was marginal before August 17 is a candidate for consolidation, not remediation. Trying to manage suppression exposure on something already borderline invites a cycle of learning phase re-entry that drags performance through Q3. If the campaign was not generating enough conversions to hold Smart Bidding stability before the change, it will not get easier after it.

Performance Max requires a separate read. PMax's automated asset and audience signals mean suppression and budget constraints interact with the underlying model differently than in Search or Shopping. Adjusting a PMax target triggers signal recalibration across assets, audiences, and placements simultaneously, so the learning phase effect is broader and less predictable. I cover the 2026 benchmarks in more detail in my Google Ads in 2026 data breakdown, but the short version is: treat PMax adjustments as their own sequencing tier and do not bundle them with Search or Shopping changes in the same window.

Q4 Implications for E-Commerce and SaaS Advertisers

All of that learning phase sequencing work only matters if it's done in time. August 17 sits less than two months before the critical Q4 ramp-up window, which means any delay in recalibrating targets translates directly into learning phase instability during the highest-value traffic window of the year. Black Friday and Cyber Monday do not wait for algorithms to settle.

For SaaS advertisers running lead generation on sponsored Google Ads, the exposure is different but equally sharp. Pipeline forecasts built on suppressed CPL benchmarks are now structurally inaccurate. The actual delivered CPL, not the stated target, is the realistic floor for any Q4 planning, and revenue projections built on the old figures will need to be revised upward before board or sales leadership sets quotas off them.

The opportunity angle is real though. Advertisers who recalibrate accurately before mid-August will have stabilised campaigns entering Q4, while competitors who ignore the change will be firefighting performance degradation during peak season. That is a genuine competitive advantage, and it compounds because disrupted accounts tend to over-correct with budget increases that just accelerate the problem.

On budget reallocation: if CPAs rise materially on specific campaign types after August 17, adding more budget is not the answer. The better move is evaluating whether channel mix needs to shift, reallocating spend toward campaigns with smaller suppression gaps where efficiency is more structurally sound.

More broadly, if your growth model relied on the hidden efficiency gap, Q3 2026 is a structural reset. Rebuild budget plans from post-change actuals, and treat any pre-August benchmarks as legacy data. For the kind of forward-looking framework this requires, marketing fundamentals that actually drive growth in 2026 is worth reading alongside this recalibration process.

What I'd Actually Do This Week

So given everything above, here is what I would actually action this week, before anything else.

First, run the 90-day suppression gap diagnostic covered earlier, your stated target versus actual delivered CPA or ROAS on every budget-limited Smart Bidding campaign. That gap is your exposure number and your prioritisation order.

Second, open the Bid Target Adjustment Tool inside Google Ads and review what Google is recommending. Treat those suggestions as a starting reference, not gospel. Cross-check each one against your own historical data before accepting anything, because Google's recommendation is based on its model, and your account history may tell a different story.

Third, stagger your target changes as covered in the learning phase section, highest-exposure campaigns first, to avoid pushing the entire account into learning simultaneously ahead of Q4.

Fourth, recalculate your budget floors using actual delivered CPAs and build in a meaningful CPA buffer (sized to your margin structure and account history). Flag any campaign where rising post-change costs could push monthly conversion volume below the threshold Smart Bidding needs to stay stable. Those are your highest-risk campaigns and they need a plan before 17 August.

Finally, and this matters more than any tactical step: set stakeholder expectations now. The window between now and when your campaigns restabilise is a re-baselining period, not an optimisation period. Performance may look softer than usual. That is the correct outcome of surfacing real costs, not a signal that something is broken.

Conclusion

Google's Smart Bidding changes are not a minor update you can absorb passively. They expose the gap between the targets you set and the costs your campaigns actually carry, and that gap has real budget consequences heading into one of the most competitive quarters of the year.

The core takeaways are straightforward: audit your suppression gaps now, recalibrate targets against historical data rather than Google's recommendations alone, stagger your changes to protect learning stability, and reset stakeholder expectations before performance looks different on the dashboard.

The advertisers who will come out of this strongest are the ones who treat the next six weeks as a structured re-baselining exercise, not a crisis to manage reactively.

Start with your exposure audit today. The data you need is already in your account. Use it before August 17 makes that choice for you.