Independent tests show AI Max underperforms traditional match types

Extensive independent testing from advertising professionals has revealed significant performance gaps between Google's AI Max for Search campaigns and traditional keyword match types, contradicting the platform's promises of improved efficiency at similar costs.

Smarter Ecommerce analyzed data from over 250 retail campaigns and published findings on November 6, 2025 showing that AI Max consistently underperforms other match types within the same campaigns. The analysis found that AI Max delivers conversions at approximately 35% lower return on ad spend compared to traditional targeting methods, with cost per conversion significantly higher and average order values notably lower.

According to the Smarter Ecommerce analysis, AI Max currently represents just 0.57% of account-level ad spend across the 600+ accounts examined with active Search campaigns. Within campaigns where the feature is activated, AI Max accounts for 5% of total spend, indicating it remains secondary to other match types despite Google's aggressive promotion of the technology.

The data contradicts Google's claim that AI Max delivers 14% more conversions or conversion value at similar cost per acquisition or return on ad spend for non-retail campaigns. Google has explicitly stated it offers no performance expectations for retail campaigns using AI Max, yet the platform has pushed the feature toward all advertisers regardless of campaign maturity or vertical.

"In last week's earnings call, Google's Chief Business Officer, Philipp Schindler, stated: 'AI Max in Search is already used by hundreds of thousands of advertisers, currently making it the fastest-growing AI-powered Search Ads product,'" according to the Smarter Ecommerce post. The analysis found AI Max active in 12% of the 600+ accounts with active Search campaigns examined in their sample.

The most concerning findings relate to efficiency metrics. AI Max shows higher cost per conversion across the board compared to broad match, phrase match, and exact match within identical campaigns. Average order value for AI Max conversions runs substantially lower than other match types, creating a compounding negative effect on overall return on ad spend.

"Our (retail) data certainly contradicts Google's claims that AI Max delivers additional volume at similar efficiency (in non-retail campaigns)," according to the analysis. "However, is this result surprising? Is the data damning? In our view, no and no. It's in fact quite logical to expect that, in a well-optimized campaign, additional conversions will cost more. This is how basically every marketing channel works."

The analysis highlighted several structural concerns about how Google has implemented and marketed AI Max. The feature targets traffic that advertisers may have purposely excluded previously, such as competitor brand terms. AI Max technology overlaps significantly with Dynamic Search Ads and Performance Max campaigns, creating account management complications. Google markets AI Max as the gateway toward AI placements including AI Mode, despite the fact that these placements can be served through broad match keywords without requiring AI Max activation.

Separate testing conducted by Xavier Mantica and published on November 6, 2025 presented even more troubling results. After four months of testing, Mantica's analysis showed AI Max delivered conversions at $100.37 per conversion, representing a 90% higher cost than phrase match at $43.97 per conversion.

"I'm done with AI Max," Mantica stated in his analysis. "After 4 months of testing, the data is brutal." His test compared AI Max performance directly against exact match ($52.69), phrase match ($43.97), exact match close variants ($61.65), and phrase match close variants ($97.67). AI Max lost to every single match type, including close variants which themselves are often criticized for poor performance.

Mantica characterized AI Max as Google's attempt to rebrand broad match without calling it broad match. "They couldn't convince advertisers to use broad match because we know it burns budget on irrelevant searches. So they rebranded it as an 'AI-powered feature' with a shiny new name," according to his analysis.

When AI Max is enabled, keywords become suggestions rather than targeting parameters. The system shows advertisements to users searching for anything remotely similar to specified keywords or content found on landing pages. "Your campaign becomes full broad match. Keyword relevance goes out the window," according to Mantica's assessment.

The performance patterns align with earlier industry testing documented in August 2025 that showed 99% of impressions generated zero conversions across approximately 30,000 search terms utilizing AI Max features. Ezra Sackett, Director of Paid Search at Monks, found that less than half of search terms showed keyword matching, with the majority functioning as keywordless advertisements.

Mike Ryan, Head of Ecommerce Insights at Smarter Ecommerce, commented on Mantica's findings: "I hear ya Xavier. I looked at over 250 campaigns and found it's also the worst match type by the numbers. To be fair, I would expect 'on top' conversions to cost more (assuming the campaign is already well optimized). But on the other hand, Google claims it will match efficiency! Not pleased about that."

The testing raises questions about Google's positioning of AI Max as delivering incremental conversions without efficiency trade-offs. While additional volume naturally costs more in mature campaigns, Google's explicit promise of similar efficiency levels appears unsupported by independent testing across multiple advertisers and industry professionals.

Google announced AI Max for Search campaigns on May 6, 2025, promising automated targeting and creative optimization through three primary mechanisms: search term matching that expands keyword reach using broad match and keywordless technology, text customization that generates headlines and descriptions based on landing pages, and final URL expansion that directs users to optimized website pages.

The platform has systematically integrated AI Max across its infrastructure. Google Ads Editor version 2.10 introduced AI Max features on July 8, 2025, while Google Ads API version 21 added AI Max support on August 6, 2025. The company launched a dedicated podcast series addressing AI Max implementation and rolled out specialized reporting metrics in September 2025 to provide visibility into AI-generated traffic.

Despite these infrastructure investments, industry concerns about AI Max expansion onto Search Partner Network placements have intensified. Analysis shows AI Max generates disproportionate impression volumes across Search Partner sites compared to traditional match types. Search Partner Network placements consistently underperform Google Search proper, delivering 37% lower return on ad spend according to research from Intelligency Group.

The timing coincides with Google's broader automation push. The company unveiled its Power Pack strategy on September 16, 2025, combining AI Max for Search, Performance Max, and Demand Gen campaigns for comprehensive automated advertising. Google modified its campaign setup flow in September 2025 to prioritize Performance Max campaigns, requiring advertisers to explicitly select specific channels to avoid the automated campaign type.

The Smarter Ecommerce analysis concluded that while AI Max technology merits experimentation, advertisers should not rush into adoption. The analysis noted several specific concerns that make the feature problematic for many advertisers:

AI Max is pushed toward all advertisers regardless of campaign maturity. Mature Search campaigns with years of optimization may not benefit from aggressive expansion into untested query territory. The feature often targets traffic that advertisers excluded intentionally, including competitor brand terms and tangential queries that historically delivered poor performance.

The technology creates redundancy with Dynamic Search Ads and Performance Max campaigns. Advertisers already utilizing these automated campaign types may find AI Max adds complexity without meaningful incremental value. Account hygiene becomes more challenging as multiple automated systems compete for similar traffic across different campaign structures.

Google markets AI Max as necessary for accessing AI placements including AI Mode, despite the fact that broad match keywords can serve these placements without AI Max activation. This positioning may pressure advertisers into adopting AI Max even when traditional match types would deliver superior performance.

"That's why we find this technology interesting, and worth experimenting with, but also not a matter to rush into," according to the Smarter Ecommerce analysis.

The performance data arrives as Google celebrates its 25th anniversary of Google Ads in October 2025, marking the platform's transformation from manual campaign management to AI-powered automation. The company has systematically reduced advertiser control in favor of machine learning optimization, with AI Max representing the latest iteration of this strategic direction.

For marketing professionals managing Search campaigns, the independent testing suggests caution before activating AI Max features. Traditional exact match and phrase match keywords continue delivering superior efficiency in retail campaigns, while broad match provides expansion opportunities without requiring AI Max's additional layer of automation.

Advertisers testing AI Max should implement careful performance monitoring, comparing match type performance within the same campaigns using the specialized reporting metrics Google introduced. The platform now provides "AI Max expanded matches" and "AI Max expanded landing pages" metrics at the all keywords level, enabling granular analysis of AI-generated traffic versus advertiser-controlled targeting.

Budget requirements add another consideration. Google disclosed in an October 2025 webinar that campaigns must maintain minimum $50 daily budgets and avoid budget limitations before upgrading to AI Max. The company recommends eight-week testing periods to allow sufficient learning time for keywordless matching and automated asset generation.

The testing framework Google suggests includes a one to two week learning period during which advertisers should avoid major campaign modifications. Performance during this phase does not represent steady-state results, as algorithms actively explore targeting possibilities and asset combinations. This extended ramp-up time increases the risk for advertisers, particularly when independent testing suggests the eventual performance may not justify the learning investment.

Text guidelines functionality introduced on September 10, 2025 provides some control mechanisms for AI Max creative generation. Advertisers can specify up to 40 term exclusions and messaging restrictions using natural language instructions. However, these controls operate after AI Max activation rather than providing upfront targeting constraints, meaning budget may be wasted during the implementation and refinement period.

The independent testing results create a challenging situation for advertisers. Google continues promoting AI Max as a fast-growing product with hundreds of thousands of active users, yet the actual performance data from objective third-party analysis contradicts the platform's efficiency promises. Marketing professionals must weigh Google's internal case studies against extensive independent testing showing consistent underperformance.

The situation reflects broader tensions in the advertising technology industry around automation, transparency, and control. Platforms prioritize machine learning systems that optimize for their own revenue metrics while claiming to maximize advertiser performance. Independent testing becomes essential for marketing professionals to validate platform claims and make informed decisions about campaign strategy.

For the digital advertising community, these developments underscore the importance of rigorous testing and skepticism toward platform promises. Google's claim of 14% conversion improvements at similar efficiency has not materialized in retail campaign testing, while the 35% lower ROAS documented by Smarter Ecommerce suggests AI Max may actively harm campaign performance for many advertisers.

The path forward for AI Max remains uncertain. Google has significant infrastructure investment in the technology and continues promoting it aggressively through product integrations, dedicated reporting, and strategic positioning within its Power Pack framework. However, if independent testing consistently shows underperformance compared to traditional match types, advertiser adoption may stall regardless of Google's promotional efforts.

Marketing professionals managing Search campaigns should prioritize traditional match types until AI Max demonstrates consistent performance improvements in their specific verticals and account structures. Testing should be conducted systematically with proper controls, comparing AI Max directly against exact match, phrase match, and broad match within identical campaigns over extended periods. Budget protection measures including daily spend caps and careful search term analysis can minimize risk during testing phases.

The independent analysis from Smarter Ecommerce and Xavier Mantica provides valuable benchmarks for the marketing community. Their willingness to publish detailed performance data enables other advertisers to make informed decisions about AI Max adoption rather than relying exclusively on Google's internal case studies and optimistic projections.

Timeline

• May 6, 2025: Google announces AI Max for Search campaigns promising 14% more conversions at similar efficiency
• July 8, 2025: Google Ads Editor 2.10 adds AI Max functionality to desktop application
• August 6, 2025: Google Ads API v21 introduces AI Max support and specialized reporting views
• August 17, 2025: Industry testing reveals 99% of AI Max impressions generate zero conversions across 30,000 search terms
• August 27, 2025: Experts flag concerning AI Max expansion onto Search Partner Network placements
• August 29, 2025: Google launches podcast series addressing AI Max implementation
• September 8, 2025: Google introduces AI Max expanded matches reporting showing traffic from AI-generated keywords
• September 10, 2025: Google introduces text guidelines for AI Max campaigns with term exclusions
• September 16, 2025: Google unveils Power Pack strategy combining AI Max with Performance Max and Demand Gen
• October 2025: Google reveals AI Max budget requirements of minimum $50 daily spend in webinar
• November 6, 2025: Smarter Ecommerce publishes analysis of 250+ campaigns showing AI Max delivers 35% lower ROAS
• November 6, 2025: Xavier Mantica shares four-month test showing AI Max costs $100.37 per conversion versus $43.97 for phrase match

Summary

Who: Smarter Ecommerce and Xavier Mantica conducted independent performance testing of Google's AI Max for Search campaigns across hundreds of retail campaigns and multiple advertiser accounts.

What: The testing revealed AI Max consistently underperforms traditional match types including exact match, phrase match, and broad match, delivering conversions at approximately 35% lower return on ad spend with significantly higher cost per conversion and lower average order values.

When: Smarter Ecommerce published findings from analysis of over 250 campaigns on November 6, 2025, while Xavier Mantica shared results from four months of testing on the same date.

Where: The performance issues affect retail Search campaigns across Google Ads accounts, with AI Max representing just 0.57% of account-level spend and 5% of spend within campaigns where it is activated.

Why: The testing matters because it contradicts Google's claim that AI Max delivers 14% more conversions at similar efficiency levels, revealing the feature may actively harm campaign performance for many advertisers despite aggressive promotion by the platform as its fastest-growing AI-powered Search Ads product.