Google Ads bidding changes: August 2026 guide

August 2026 brings one of the most consequential automatic bidding changes Google Ads has made in years. From August 17, campaigns running Target CPA or Target ROAS that carry a "Limited by budget" status will stop systematically beating their targets. For any account that quietly relied on that overperformance, the impact will be immediate and measurable.

Key takeaways

• Budget-limited tCPA and tROAS campaigns will deliver results close to the target from August 17, not below it.
• A 90-day audit of affected campaigns is essential before the deadline.
• Smart Bidding Exploration delivers roughly 10% more conversions than simply lowering a ROAS target.
• Classic Display campaigns are migrating to Demand Gen, and parallel testing should start now.
• ChatGPT Ads now offers conversion-optimized bidding and is becoming a credible performance channel.

What the August 17 change actually does

Until now, the algorithm could deliver conversions well below your CPA target or well above your ROAS target on budget-limited campaigns. Google is deliberately removing that behavior. As Search Engine Journal reports, the change covers Search, Shopping, Performance Max, Demand Gen, Travel and Display (App and Video are excluded).

The practical translation: a campaign set to a $10 CPA that has been delivering at $5 will drift toward $10 after the change. The algorithm stops overshooting and starts treating the target literally.

Your first move is a filter. Pull every campaign flagged "Limited by budget" and compare actual CPA or ROAS over the last 90 days against the set target. A wide gap between the two is your risk signal.

Three decisions, one for each affected campaign

Once you have your list, you have three options. None of them is wrong. Doing nothing is.

1. Reset the target to match reality. Google's Bid Target Adjustment Tool has been live since July 6. Use it to set a target that reflects actual performance, ideally grounded in contribution margin rather than average CPA alone. Search Engine Journal flags this as the cleanest fix for campaigns where you want to preserve efficiency.

2. Increase the budget. If a campaign is no longer budget-limited, the problem disappears. This makes sense when you know the campaign has scaling headroom and the margin supports it.

3. Accept the drift intentionally. In some cases a higher CPA or lower ROAS is commercially acceptable. The word that matters is "intentionally." Letting the algorithm decide without a conscious choice on your part is the mistake.

For Performance Max specifically, watch channel-level reports. Google has signaled that rebalancing will shift traffic between Search, Shopping and YouTube, with a risk of moving spend toward cheaper but less conversion-heavy inventory.

Smart Bidding Exploration for deliberate growth

If you want to scale after the reset rather than just defend current performance, Smart Bidding Exploration is worth testing. It lets Google probe conversions outside your safest audience segment while respecting a ROAS tolerance you define. Data cited in Search Engine Journal puts the uplift at roughly 10% more conversions versus simply lowering the ROAS target, without restructuring the campaign.

For seasonal peaks and launch windows, Promotion Mode (currently in beta) offers a short-term way to increase bidding aggressiveness, and it pairs well with the campaign total budget type for time-bound promotions.

If you are considering Google AI Max for Search, PPC Hero is direct about the prerequisite: AI Max is an amplifier. A clean account with accurate conversion tracking benefits; a messy account sees its problems multiply. Before activating, clean up mixed match types, configure text guidelines to prevent Gemini from pulling legal disclaimers or technical jargon from your landing pages, and exclude problematic URLs from Final URL expansion.

Display moves to Demand Gen, and ChatGPT Ads grows up

Running alongside the bidding change is the migration of classic Display campaigns into Demand Gen. WordStream describes Demand Gen as unifying placements across YouTube, Gmail, Discover and the display network, with a stronger emphasis on video and visually rich formats. Existing Display campaigns continue to serve, but new feature support is winding down. Starting a parallel Demand Gen test now means you will have real performance data before a forced migration, rather than scrambling when the timeline arrives.

Then there is ChatGPT Ads. Marketing Land covers the addition of conversion-optimized CPC (oCPC), a 7-day rolling average daily budget, and integrations with AppsFlyer and Adjust. ChatGPT Ads is no longer an experiment. For advertisers targeting an AI-native audience, it is a legitimate performance channel worth a pilot. Agencies and larger advertisers will find the Ads API's asynchronous bulk operations useful for scaling campaign management without manual overhead. If you run campaigns across multiple clients, the case for building this into your workflow is growing.

Five things to do before August 17

1. Filter all "Limited by budget" campaigns and compare 90-day actual CPA or ROAS against the set target.
2. Assign each affected campaign one of the three tactics: reset the target, increase the budget, or accept the drift on purpose.
3. Check Performance Max channel-level reports and prepare for potential traffic shifts.
4. Launch a Demand Gen campaign in parallel with your best-performing Display campaign.
5. Run a small ChatGPT Ads oCPC pilot if your audience actively uses AI tools.

The August change is not a disaster, but it is the moment when passive bidding management stops working. Accounts with clean structure and accurate conversion data will extract more from the new tools than those waiting to see what the algorithm does on its own.