Google has expanded Smart Bidding Exploration to Performance Max campaigns without product feeds and launched a Shopping ads beta alongside a new Promotion Mode feature, giving paid-search teams more control over how AI bidding behaves during high-demand windows.

Smart Bidding Exploration works by letting advertisers set a ROAS tolerance, giving campaigns permission to chase conversion volume beyond a strict target. The premise is that some queries convert well enough to justify a looser return threshold, and the system finds those opportunities automatically. Google reports that campaigns using the feature see an average 18% increase in unique converting search query categories and a 19% increase in conversions. Those figures come from Google’s own internal data, with no independent measurement published. Treat them as directional, not benchmarks.

The expansion matters structurally. Previously, Smart Bidding Exploration was limited to specific campaign types. Extending it to Performance Max campaigns running without product feeds, and rolling out a Shopping beta across both Performance Max and Standard Shopping, means the ROAS-tolerance mechanism now touches a much larger share of a typical e-commerce account’s budget. For advertisers running PMax without a product feed, the feature is now on the table for the first time.

Promotion Mode is the more operationally specific addition. The beta lets advertisers temporarily lift ROAS targets and increase daily budget allocations during defined high-demand periods, covering seasonal events, product launches, and flash sales. Google positions it for retailers, but any advertiser with predictable demand spikes has a reason to test it.

The practical question for a retail PPC team is whether to rely on Promotion Mode for peak periods or handle the same outcome manually through campaign-level bid adjustments and budget rules. Promotion Mode formalizes a workflow many teams already run on spreadsheets, which has value in reducing execution errors during high-stakes windows. What it does not do is give the advertiser more transparency into how the system is moving spend during those periods. If something goes wrong mid-promotion, the diagnostic path is the same as any other AI bidding issue: pull Search Terms, check Auction Insights, and compare actual ROAS against target in daily increments.

Two timeline items require immediate attention. On July 6, advertisers will receive notifications about campaign adjustments required under the expanded system. On August 17, Google will update how bidding targets are optimized specifically for budget-constrained campaigns. That second date is the one to watch for accounts where spend caps are frequently hit. Bidding behavior in constrained conditions is notoriously difficult to model, and a platform-side change to optimization logic in those scenarios can shift performance in ways that are not immediately visible in standard reporting.

The compounding effect of these changes is that Google is widening the gap between accounts that have explicitly configured their ROAS tolerance and those running on default or legacy settings. An account that has not revisited its bidding strategy since Smart Bidding Exploration first launched may be operating on assumptions that no longer match how the system behaves.

Before peak season, retail PPC teams should do three things. First, audit which campaigns are eligible for Smart Bidding Exploration and set intentional ROAS tolerance ranges rather than leaving defaults in place. Second, evaluate whether Promotion Mode fits the workflow for upcoming sale events, and if so, request beta access before those events are imminent. Third, flag the August 17 date for any budget-constrained campaigns and build a monitoring plan around it, because the system’s behavior will change and the notification may not surface clearly in the interface.

The expansion puts more surface area under AI control. That is not inherently bad, but it requires deliberate configuration to work in the advertiser’s favor.

Reported by Search Engine Land on June 15, 2026, written by Anu Adegbola.