Bidding (from 17 August 2026): Budget-limited Target ROAS/CPA campaigns that have been quietly beating their goals will now be steered back toward the actual target set — expect costs to drift upward on overperforming campaigns unless targets are reviewed.
Display → Demand Gen: Standalone Display campaigns are being retired and folded into Demand Gen, which also runs across YouTube, Discover, Gmail and Maps. Migration is voluntary for now, with automatic migration to follow.
Performance Max: After years of being a “black box,” PMax is testing the ability to switch off Search Partners and Display within campaigns, alongside new account-level placement exclusions and channel reporting.
Search AI Max: Becoming the default layer for Search campaigns, with older features like Dynamic Search Ads, automatically created assets, and broad match being auto-upgraded into it on a set timetable.
Bottom line for leaders: Google Ads now needs ongoing oversight, not a one-time setup — budgets, targets, and creative all need regular review as automation takes on more of the decision-making.
If you run paid media, you’ve probably noticed something odd this summer: Google Ads doesn’t look like the platform you learned two years ago. Target ROAS is behaving differently. Display campaigns have quietly started disappearing into something called Demand Gen. Performance Max, long criticised as a “black box”, is suddenly letting advertisers switch channels off. And Search campaigns are being nudged, sometimes without asking, into a new AI-powered layer called AI Max.
None of these changes happened in isolation. Together, they tell a clear story: Google is trading manual control for automated scale, and asking advertisers to trust the machine a little more each quarter. For marketing leads, that’s not just a technical shift — it’s a question of predictability, budget discipline and boardroom credibility. Below is a plain-English walkthrough of the four biggest updates, and what each one means for the expectations you’re setting with your leadership team.
1. Target ROAS and Target CPA: the “free lunch” is ending
From 17 August 2026, Google changed how budget-limited Search, Shopping, Performance Max, Demand Gen and Travel campaigns behave when they’re running Target CPA or Target ROAS bidding. Previously, if your budget capped a campaign, Smart Bidding would often let it quietly outperform its stated target — a £10 Target CPA campaign might have been delivering £5 conversions for months without anyone tightening the dial. That “overperformance” is being phased out. Campaigns will now be steered more consistently toward the actual target you’ve set, even after you adjust budgets.
For a business leader, this sounds technical, but the consequence is direct: cost per acquisition or ROAS on campaigns that have quietly been beating expectations may start drifting upward, purely because Google is now enforcing the number you originally typed in, rather than the number you’ve actually been getting. Google has been explicit that it won’t change budgets or targets automatically — a Bid Target Adjustment Tool has been rolled out so teams can review and reset targets before performance shifts. The practical takeaway is simple: any campaign you’ve been quietly celebrating for beating its goals needs a review before those gains evaporate.
2. Display is folding into Demand Gen
Standalone Display campaigns, one of the oldest formats in Google Ads, are being retired. Since mid-2026, eligible advertisers have been able to migrate Display campaigns voluntarily into Demand Gen using a purpose-built tool that carries across up to 42 days of performance history. New Display campaigns can now only be created inside Demand Gen, and any campaigns left un-migrated will eventually move automatically.
This isn’t the end of the Google Display Network itself — GDN inventory across roughly two million sites and apps is still very much alive. What’s changing is the wrapper around it. Demand Gen bundles GDN together with YouTube, Discover, Gmail and Maps in one AI-optimised campaign, and Google reports an average 9.5% increase in ROI for advertisers who add GDN into Demand Gen rather than running it standalone. The trade-off is a loss of some granular controls: manual CPC, certain bid adjustments and portfolio bidding don’t carry across, and content suitability exclusions need to be set at account level rather than campaign level before migration will complete.
For business leaders, this matters because Display was often the “always-on brand awareness” line item in the media plan — cheap, broad, low-drama. Folding it into a demand-generation format means richer creative requirements (video, multiple aspect ratios, stronger imagery) and a fresh learning period each time a campaign migrates. Budget owners should expect a short dip in reported efficiency immediately after migration, followed by a recovery, rather than a like-for-like continuation of historic numbers.
3. Performance Max is finally admitting it has a “platform” problem
Performance Max has spent three years being criticised for being a black box: advertisers hand over budget and creative, and Google decides where it spends across Search, Display, YouTube, Gmail, Maps and Discover, with limited visibility into which channel actually drove the result. That’s now changing, in stages.
Google has rolled out account-level placement exclusions that apply across PMax, Demand Gen, YouTube and Display simultaneously, added channel-level reporting through the Ads API, and permanently removed low-quality “parked domain” inventory from the Search Partner Network. More significantly, Google is alpha-testing the ability to switch off Search Partners and the Display Network entirely within individual PMax campaigns — something that hasn’t been possible since Smart Shopping was retired. Industry data cited in trade coverage suggests Search Partner Network traffic converts at a return roughly 37% below Google Search proper, which explains why this control is in such demand.
The business implication is a genuine shift in negotiating power. For years, leadership teams were told PMax’s opacity was the cost of its performance. That argument is getting harder to make as Google itself adds the dials advertisers have been requesting. If your agency or in-house team is still running PMax exactly as it was configured in 2023, this is the moment to revisit it — new exclusion and reporting tools mean “we can’t see or control that” is no longer a fully accurate answer to a leadership question about spend quality.
4. AI Max is becoming the default for Search, whether you opt in or not
AI Max for Search campaigns — Google’s optional layer of keywordless targeting, AI-generated ad copy and landing-page expansion — has moved from beta to a mainstream, increasingly mandatory part of the Search product. Dynamic Search Ads, automatically created assets and campaign-level broad match settings began auto-upgrading to AI Max from September 2026, with DSA campaigns following in February 2027 after Google extended the original timeline in response to advertiser pushback.
Google’s own data claims AI Max delivers an average of 7–14% more conversions or conversion value at a similar CPA or ROAS when the full feature set is switched on, and adoption has already passed half a million advertisers. New controls have followed the criticism: branded search settings now let advertisers keep AI Max focused on “unbranded only” traffic, addressing the common concern that it was cannibalising dedicated brand campaigns. AI Max ad placements are also increasingly the only way to appear inside AI Overviews and Google’s AI Mode — surfaces that a growing share of high-intent commercial searches now pass through.
For business leaders, the message is that “we haven’t touched Search automation” is no longer a neutral position — it’s an active decision with a fixed expiry date, since legacy features are being sunset around it. The sensible response isn’t blind adoption, though. As several industry commentators have pointed out, AI Max performs best where conversion signals are already strong and first-party data is clean; lead-generation advertisers without robust offline conversion tracking are the ones most likely to repeat the early mistakes made with Performance Max — trusting the automation before feeding it the business context it needs.
What this means at the leadership level
Pulled together, these four changes point in the same direction: Google is consolidating campaign types, tightening the gap between stated targets and actual delivery, and pushing automation deeper into every layer of the account. For marketing leaders reporting into the business, that has three practical consequences worth flagging now, before performance shifts and someone asks why:
- Set expectations on volatility, not just targets. Budget-limited campaigns beating their goals are likely to see costs rise toward the stated target over the coming weeks. Build that into forecasts rather than reacting to it as underperformance.
- Reframe “loss of control” as a temporary, not permanent, state. PMax’s new exclusion controls and AI Max’s branded search settings show Google responding to advertiser pressure. Teams that keep pushing for granularity, rather than assuming automation is fixed and final, are getting more of it.
- Treat creative and data quality as the new levers. With bidding, channel selection and even copywriting increasingly automated, the remaining competitive advantage sits in the signals you feed the system: clean conversion data, strong first-party audiences and genuinely differentiated creative assets.
None of this means panic. It means treating Google Ads less like a campaign you configure once and more like a platform you actively govern — reviewing targets, auditing new controls and asking what’s changing before it changes for you.
No — Google has confirmed it won’t automatically alter your budgets. What changes is how closely Smart Bidding sticks to the target you’ve set within that existing budget. If a campaign has been quietly beating its target, its reported cost per acquisition or ROAS may drift back toward that stated figure rather than your spend increasing outright.
Not immediately, but it’s worth planning for. Google is running a phased, voluntary migration first, with automatic migration following later for campaigns left untouched. Migrating proactively with Google’s own tool preserves up to 42 days of performance history and shortens the learning period compared with waiting for an automatic switch.
This control is currently in limited alpha testing, so most advertisers don’t yet have it. Account-level placement exclusions, which apply across PMax, Demand Gen, YouTube and Display together, are already broadly available and are the best current tool for cutting out low-quality placements while the fuller channel toggle rolls out.
It’s officially still optional, but Google is auto-upgrading many legacy features into it. Campaigns using Dynamic Search Ads, automatically created assets or campaign-level broad match are being migrated on a set timetable, so in practice most active Search advertisers will end up running some form of AI Max whether they actively enable it or not.
Look past headline CPA or ROAS in isolation and track it against your pre-change baseline for at least two to four weeks after any migration or bidding change, since automated systems typically need a short relearning period. Pair that with incrementality checks — branded search lift, assisted conversions, and blended customer acquisition cost — rather than relying on last-click platform reporting alone.
