Google Ads bidding is about to change in a way every advertiser running Target CPA or Target ROAS needs to understand before August 17, 2026. On June 15, Google announced its mid-year bidding and budgeting overhaul — and the headline item, Bidding Target Optimization, will pull budget-limited campaigns that have been over-delivering their stated targets back toward those targets.
For accounts that quietly beat their goals, this is not a feature you opt into — it is a behavioral change that arrives automatically on August 17. The good news: Google is shipping a Bid Target Adjustment Tool on July 6 that surfaces your historical performance and gives you three ways to respond. Between those two dates sits a six-week window to decide whether your over-performance was a deliberate strategy or simply a target you never updated.
This playbook covers all three June 15 announcements — Bidding Target Optimization, the expansion of Smart Bidding Exploration, and the Promotion Mode beta — plus the separate June 1 pacing change that now stacks on top of them. We will map who is affected, recompute the math behind Google's own examples, and lay out a week-by-week audit so you walk into August 17 with intent rather than surprise.
- 01Bidding Target Optimization lands August 17, 2026.Budget-limited campaigns that have historically beaten their stated Target CPA or Target ROAS will be steered back toward the target you set. Google states it will not automatically adjust your targets or budgets — advertiser action is required.
- 02The Bid Target Adjustment Tool arrives July 6.Triggered by account notifications for advertisers with budget-limited target-based campaigns in the past 12 months. It surfaces historical performance and three options: keep the target, match it to recent performance, or set a custom target.
- 03The real question is the intention gap.If your conservative targets were a deliberate lever to keep campaigns scaling, you should lower them before Aug 17. If the targets simply drifted out of date as performance improved, you can accept the change and monitor.
- 04Smart Bidding Exploration expanded — mind the GA vs beta line.It is now globally available for Performance Max campaigns without product feeds. For Shopping and PMax-with-feeds it is launching in beta. Google reports internal lifts of +18% and +19% from its own test window — vendor-stated, not independently verified.
- 05Promotion Mode is a new Search + PMax beta.It schedules temporary ROAS-tolerance changes and extra daily budget for peak events. It is distinct from seasonality adjustments and, at beta launch, confirmed only for Search and Performance Max — not Shopping or Display.
01 — What Google AnnouncedThree changes, one announcement.
Google bundled three distinct moves into its June 15, 2026 announcement, and conflating them is the fastest way to make a wrong call. The first is Bidding Target Optimization, a backend change effective August 17 that affects how budget-limited campaigns deliver against their stated targets. The second is the expansion of Smart Bidding Exploration, which moves from Search-only to all Performance Max campaigns without product feeds, with a separate beta for Shopping. The third is Promotion Mode, a new beta for Search and Performance Max that schedules temporary ROAS-tolerance and budget boosts around peak events.
The package rolls out across Google Ads, Search Ads 360, Display & Video 360, Google Ads Editor, and the Google Ads API. Google also renamed two strategies in June 2026 with no behavioral change: "Maximize conversions with a Target CPA" is now simply Target CPA, and "Maximize conversion value with a Target ROAS" is now Target ROAS. No account adjustment is required for the rename — it is purely cosmetic.
Bidding Target Optimization
Budget-limited campaigns that historically over-deliver their stated tCPA or tROAS get steered toward the set target. Google says it will not auto-adjust your targets or budgets — you act, or you accept the pullback.
Smart Bidding Exploration
Globally available for Performance Max without product feeds across all languages. Shopping and PMax-with-feeds remain in beta. Lets the algorithm bid on queries with unproven conversion histories within a ROAS tolerance range.
Promotion Mode
Schedule temporary ROAS-tolerance changes and extra daily budget across a defined peak window — flash sales, seasonal events, launches. Works with both daily and campaign total budgets. Not Shopping or Display at beta launch.
02 — The August 17 ChangeWhen your campaign beats its target, that ends.
Here is the mechanism in plain terms. Today, a budget-limited campaign with a Target CPA can deliver conversions well below that target — Google's own Help Center uses an illustrative example of a campaign set to a $10 Target CPA that has been delivering conversions at around $5. After August 17, that campaign will aim to deliver closer to the $10 target it was actually set to, rather than continuing to over-perform at $5. The $10-versus-$5 figures are Google's illustration, not a measured advertiser result — treat them as a teaching example.
The same logic applies to Target ROAS. A campaign set to a 300% tROAS that has been delivering closer to 400% will be steered toward the stated 300%. The change is limited to budget-limited campaigns — those constrained by their daily or total budget rather than by their target. Campaigns that are not budget-limited are not the focus of this change.
"We're making backend bidding target optimization updates to help campaigns limited by budget see more predictable performance in line with CPA and ROAS targets."— Ginny Marvin, Google Ads Product Liaison, via Search Engine Journal, June 15, 2026
Two campaign types are explicitly out of scope. The change applies to Search, Shopping, Performance Max, Demand Gen, Travel, and Display campaigns. It does not apply to App campaigns, Video reach campaigns, or Video view campaigns (VVC), which continue their existing behavior. Note that Hotel and Display campaigns already operate under the new behavior, so for those there is nothing new to do.
03 — Decision MatrixThe Aug 17 risk matrix: who is affected, and how much.
No single piece of Google's coverage combines campaign type, performance gap, and recommended action into one decision table — so we built one. The gap figures below are computed from each row's stated scenario (a $10 target delivering at $5 is a 2x over-performance; a 300% tROAS delivering at 400% is a 33% over-performance, since 400 is one-third above 300). The dollar and percentage scenarios are illustrative, mirroring Google's own teaching example — your real numbers come from the Bid Target Adjustment Tool.
| Campaign scenario | Budget-limited? | Target gap | Risk level | Recommended action |
|---|---|---|---|---|
| Search — tCPA $10, delivering $5 | Yes | 2x over-performance | High | Lower the target toward $5 before Aug 17 if the efficiency is the goal |
| Search — tCPA $10, delivering $9 | Yes | ~11% over (minor) | Low | Accept the change; monitor over one to two conversion cycles |
| Shopping — tROAS 300%, delivering 400% | Yes | 33% over-performance | Medium | Raise the target toward ~380% to lock in recent efficiency |
| PMax — no product feed, SBE active | Not budget-limited | Not applicable | None | Opportunity — expand reach via Smart Bidding Exploration |
| App campaigns | Not applicable | Not applicable | Not affected | No action needed |
| Video reach / VVC | Not applicable | Not applicable | Not affected | No action needed |
| Display (already updated) | Varies | Not applicable | Already live | Already operating under the new behavior — no action |
The interpretation that matters: a wide gap is not automatically a problem. Some accounts deliberately set conservative targets to keep budget-limited campaigns spending and scaling — for them, the gap is a feature, and August 17 is a threat to be neutralized by lowering the target. Other accounts simply never revisited a target after performance improved — for them, the gap is drift, and accepting the change can be perfectly reasonable. The matrix tells you the magnitude; only you know the intent.
04 — The Six-Week WindowJuly 6 to August 17 is your window.
The Bid Target Adjustment Tool rolls out July 6, 2026, triggered by account notifications for advertisers who have run budget-limited target-based campaigns in the past 12 months. It surfaces your historical performance data and presents three action options — with a fourth available outside the tool. From July 6 to August 17 is almost exactly six weeks, which is enough time to audit, decide, apply, and watch the algorithm re-stabilize before the change is mandatory.
Keep the current target unchanged
Campaigns that have over-performed will move toward the stated target after Aug 17. Choose this only if your stated target genuinely reflects the cost per conversion you want. For most over-delivering accounts this means accepting higher CPA.
Match the target to recent performance
Lower the target to where the campaign has actually been delivering — for example, drop a Target CPA from $10 to $5. This locks in your current efficiency rather than letting it drift back toward the old, looser target.
Set a custom target for the goal
Pick a target that reflects your actual business economics rather than either the old number or recent delivery — useful when margins, seasonality, or a new product mix mean neither the historic target nor recent CPA is the right anchor.
Switch to a Maximize strategy
Outside the tool, you can move to Maximize Conversions or Maximize Conversion Value, removing the target constraint entirely. It frees the campaign to chase volume but reintroduces CPA and ROAS variability — only sensible when efficiency is genuinely secondary to scale.
We recommend treating the six weeks as a phased sprint. Week 0 (now, before the tool): pull a list of budget-limited campaigns running Target CPA or Target ROAS and rank them by the size of the gap between stated target and actual delivery. Weeks 1 to 2 (from July 6): use the tool's historical data to classify each campaign as intentional under-target or simple drift. Weeks 3 to 4: apply the target changes for the intentional group. Weeks 5 to 6: monitor. Google recommends waiting one to two conversion cycles after any target or budget change before evaluating performance — a conversion cycle being the time from click to reported conversion, including any CRM-import lag.
05 — Smart Bidding ExplorationMore reach without loosening your ROAS target.
Smart Bidding Exploration lets you set a ROAS tolerance range so the algorithm can bid on queries with unproven conversion histories — expanding traffic beyond established performance patterns without requiring you to significantly loosen your global ROAS target. It is the latest step in the broader arc of Google Ads bidding automation. As of June 15, 2026 it is globally available for all Performance Max campaigns without product feeds, across all languages. Previously it was limited to Search campaigns since its 2025 launch.
The critical distinction for planning: only the PMax-without-feeds rollout is generally available. For Shopping ads — both standard Shopping and Performance Max with product feeds — Smart Bidding Exploration is launching in beta, a separate track from the global GA rollout. Do not assume Shopping has the feature until you see it live in the relevant account.
Smart Bidding Exploration — Google-reported lifts
Source: Google internal data via Search Engine Land and the Google Marketing Live 2026 blog post · vendor-stated, not independently verified · results vary by advertiserThe structural argument for the Shopping beta is worth a separate note. Shopping queries behave differently from text search — they are price-sensitive and category-driven, so an algorithm exploring unproven queries is making a different kind of bet than it does on Search. That is exactly why Google is staging Shopping behind a beta rather than shipping it to GA alongside PMax-without-feeds: the risk-reward profile needs its own validation. If you run Shopping, test the beta on a controlled subset before trusting it across the account — the same discipline we apply when running Performance Max asset experiments.
06 — Promotion ModeA scheduled boost — not a seasonality adjustment.
Promotion Mode is a new beta for Search and Performance Max campaigns that lets you schedule temporary changes to your ROAS tolerance and add extra daily budget across a defined peak date range — a flash sale, a seasonal event, a product launch. It works with both daily budgets and campaign total budgets (the latter launched earlier in 2026 for Search, Shopping, and PMax). At beta launch it is confirmed for Search and Performance Max only — not Shopping, not Display.
Practitioners routinely conflate Promotion Mode with seasonality adjustments, and the difference is the whole point. The two tools answer different questions, and both can run in parallel.
Tells the model what to expect
You signal an anticipated change in conversion rate over a window — Smart Bidding then adjusts its bids in expectation. It does not directly change budget or ROAS tolerance; it changes the model's conversion-rate assumption.
Changes tolerance and budget directly
You directly adjust ROAS tolerance and inject extra daily budget across the peak window. It is an action on the levers, not a forecast — built for events where you actively want to spend more and accept a looser efficiency bar.
They are not mutually exclusive
Because one is a forecast and the other is a direct lever change, they compose. For a major sale you might both signal the conversion-rate lift and schedule a Promotion Mode budget boost over the same dates.
For agencies, Promotion Mode is the cleaner mechanism for time-boxed pushes than ad-hoc manual budget edits, because it builds the start and end dates into the campaign rather than relying on someone remembering to revert a budget after the sale ends. As a beta, expect availability and behavior to shift — validate it on one campaign before standardizing it across a client roster.
07 — The Pacing StackThree overlapping budget changes, mapped as one stack.
By August 2026, three separate Google budget and pacing changes overlap, and most coverage treats each in isolation. Stacking them in one view is where the real planning risk surfaces — particularly for ad-scheduled campaigns running target-based bidding, which now sit at the intersection of all three. The table below maps each change to its effective date, mechanism, affected population, and whether action is required.
| Change | Effective | What it does | Who it hits | Action required? |
|---|---|---|---|---|
| Campaign Total Budgets | Earlier 2026 | Sets a fixed-duration budget (days to weeks) instead of a daily rolling budget | All Search / Shopping / PMax advertisers | Optional feature — Google reports a 66% average reduction in manual budget adjustments where adopted (vendor-stated) |
| Ad-schedule pacing change | June 1, 2026 | Paces toward the full monthly cap (30.4x daily budget) regardless of a restricted schedule | Campaigns with day or hour ad scheduling | Yes — review daily budgets; restricted-schedule campaigns can see materially higher effective monthly spend |
| Bidding Target Optimization | August 17, 2026 | Steers stated tCPA/tROAS and actual delivery into alignment for budget-limited campaigns | Budget-limited tCPA/tROAS campaigns (Search, Shopping, PMax, Demand Gen, Travel, Display) | Yes — audit targets and act before Aug 17 |
08 — The Agency PlaybookWhat to do, in the order to do it.
Pulling the pieces together, here is the operating sequence we are running across managed accounts. It assumes you manage a mix of campaign types and want to walk into August 17 with deliberate targets rather than inherited ones. This sits alongside our broader paid media services, where target governance is a standing discipline rather than a once-a-year scramble.
Inventory the at-risk campaigns
List every budget-limited campaign on Target CPA or Target ROAS. Sort by the gap between the stated target and actual delivery — the widest gaps carry the most August 17 exposure. This is your Section 03 risk matrix, applied to real accounts.
Classify intent vs drift
For each wide-gap campaign, decide: was the conservative target a deliberate lever to keep it scaling, or a stale number nobody revisited? Intentional under-targeting means lower the target; drift can mean accept the change.
Apply, then wait out a cycle or two
Make target changes for the intentional group via the Bid Target Adjustment Tool. Then hold — Google advises one to two conversion cycles before judging the result. Resist the urge to re-tune mid-learning.
Layer the budget side on top. Separately check restricted-schedule campaigns against the June 1 pacing change so a campaign you thought was capped at a few hundred a month is not quietly pacing toward its full monthly cap. And where Smart Bidding Exploration is available for your Performance Max campaigns without feeds, treat August as a chance to expand qualified reach — the same window in which you are tightening targets elsewhere can be the window in which you open up exploration where the economics support it.
The strategic read is that Google is steadily removing the gap between what advertisers say they want and what the system delivers — renaming strategies to Target CPA and Target ROAS, enforcing target consistency, and adding exploration and promotion levers that assume you actually mean the targets you set. For teams that have used loose targets as a crude scaling hack, that era is closing. Looking past August, expect this consolidation to continue: features like Demand-Led Pacing, announced at Google Marketing Live on May 20 and described as coming in the months after, point to a future where pacing and targeting are increasingly automated and the advertiser's job is to set the economics correctly, not to manually nudge daily budgets. It is the same trajectory we trace in our look at agentic PPC campaign management. The accounts that win will be the ones whose targets reflect real business math — which is precisely the discipline this change rewards.
09 — ConclusionWalk into August 17 with intent.
The targets you set are about to be the targets you get.
Google's June 15 announcement is, at its core, a single consistent move: close the gap between stated targets and actual delivery. Bidding Target Optimization enforces it on August 17, the renamed Target CPA and Target ROAS strategies signal it, and Smart Bidding Exploration and Promotion Mode give you deliberate levers to expand reach or push through peaks without abandoning your targets.
The honest framing is the one practitioners need: the August 17 change is not opt-out, and inaction is itself a decision to accept your over-performing campaigns drifting back toward their stated targets. The Bid Target Adjustment Tool on July 6 gives you a six-week window to decide whether each gap is strategy or drift, and to lock in the efficiency you want before the system does it for you. Just remember Google's own internal lift figures are vendor-stated and the modeled spend increases are illustrative — plan from your own account data, not from headline percentages.
The teams that benefit are the ones that treat this as a target-governance exercise rather than a fire drill: inventory the at-risk campaigns, separate intentional under-targeting from stale drift, apply changes early enough to let the algorithm re-stabilize, and keep the June 1 pacing change in view for restricted-schedule campaigns. Do that, and August 17 is a non-event you planned for — not a surprise you inherited.