Google’s Merchant Center has launched a pilot program providing retailers with unprecedented insight into the types of shopping-related questions users are posing to its AI Mode and AI Overviews. This new feature, currently in a limited rollout, aims to equip businesses with a better understanding of consumer intent within the rapidly evolving landscape of generative AI-powered search.
Brodie Clark, an independent SEO consultant who gained early access to the pilot through a client’s sub-account, shared screenshots of the new reporting on X (formerly Twitter). He hailed it as the first Google product to offer query data specifically for these AI-driven search surfaces. While the data provided is not a direct feed of individual user queries, it offers aggregated insights into the vocabulary and categories of questions consumers are asking, which can be invaluable for optimizing product feeds.
The "AI Performance Insights" report, as it’s officially known, is designed to illuminate how brands are discovered across AI Mode and AI Overviews. Retailers with access can locate it within their Merchant Center account under the "Analytics" section, then "Products," and finally the "AI Performance" tab. Google had initially announced this reporting capability at its Google Marketing Live event in May, approximately seven weeks before the pilot’s commencement.
Understanding the Nuances of AI Performance Insights
The AI Performance Insights report categorizes shopping questions into distinct "query types," offering examples such as searching by category, researching product specifications, or seeking reviews. The "query frequency" metric indicates the popularity of these different question types. Furthermore, the report segments these queries by the "phase of the shopping journey" a consumer is in, providing insight into their progression from initial discovery to purchase intent.
"Product terms" represent the specific words and phrases shoppers use to articulate their needs. Google’s documentation provides illustrative examples like "maximum cushioning" and "arch support" for athletic footwear. Another crucial metric is "share of voice," which benchmarks a retailer’s AI impressions against those of its competitors.
Google’s accompanying documentation emphasizes the actionable nature of this data. Retailers are advised to use the insights to enrich their product listings with relevant attributes, thereby aligning their offerings with consumer inquiries. This is particularly significant for addressing attribute completeness, a persistent challenge in e-commerce. If the report reveals frequent shopper questions about a feature not adequately detailed in a product feed, it presents a clear opportunity for immediate optimization.
However, it is crucial to understand what these metrics are and are not. None of the data points in the AI Performance Insights report represent verbatim user questions. Instead, the report provides an understanding of the shape of demand rather than the demand itself. Consequently, the "product terms" are best utilized as input for feed optimization rather than as a direct keyword list for traditional search campaigns.
The "share of voice" metric, calculated as a retailer’s AI impressions divided by the total impressions across itself and its competitors, should also be interpreted with caution. The competitor set is predetermined by those already present in Merchant Center, and this cannot be manually adjusted. Furthermore, the metric can sometimes present misleading figures. A lack of sufficient impressions can result in a zero share of voice, while an absence of defined competitors may display a 100% share of voice.
Scope and Limitations of the Current Pilot
The filters applied within the AI Performance Insights report also have inherent limitations. The traffic data is restricted to organic AI traffic, excluding any paid advertising placements. Product category filters operate on a one-category-at-a-time basis, and a comprehensive report covering all categories simultaneously is not currently available. Moreover, the insights are confined to conversational queries that explicitly indicate shopping or brand intent; other types of AI interactions are not factored into this report.
The rollout of these AI-specific performance metrics follows a series of developments in Google’s approach to generative AI in search. In the preceding month, Google initiated testing of dedicated generative AI performance reports within Search Console, initially for a subset of UK websites. These reports provided impression data broken down by page, country, device, and date, but notably omitted click data and query-level metrics. The absence of click data has been a recurring point of discussion.
Simultaneously, the UK’s Competition and Markets Authority (CMA) imposed a conduct requirement on Google, mandating improved publisher controls and reporting for generative AI features. The CMA’s guidelines specifically call for the separation of impressions, click-throughs, and click-through rates for generative AI features from general search metrics. While Google has a nine-month window to implement these changes, the crucial click and click-through rate reporting for AI features has yet to materialize.
Adding to this context, in July, Google informed Chief Marketing Officers (CMOs) that third-party AI visibility tools do not possess access to its internal metrics. Google identified Search Console and Merchant Center reporting as the baseline for tracking AI-driven gains. This statement was made just three weeks prior to the launch of the Merchant Center pilot.
When viewed together, these staggered rollouts suggest a deliberate, albeit phased, approach by Google to integrating AI performance data. The initial Search Console reports addressed neither of the primary questions many sought: how users interact with AI search and how it drives traffic. The current Merchant Center pilot, while offering a partial answer to the former for US-based merchants, still excludes paid traffic and lacks crucial click data.
Strategic Placement of AI Visibility Tools
The decision to introduce AI visibility reporting into Merchant Center, rather than solely relying on Search Console, has strategic implications. SEJ contributor Slobodan Manic previously argued that placing AI visibility within Search Console was a deliberate choice, framing AI visibility as an extension of search visibility. However, the Merchant Center pilot does not necessarily contradict this. Google’s guidance to CMOs designates both dashboards as first-party reporting, suggesting that AI visibility is indeed measured where search visibility is tracked. The fact that grouped query information has landed in the merchant-focused dashboard, rather than the general search performance tool, indicates a tailored approach for specific user segments.
Implications for Search Professionals
The early access to the Merchant Center pilot provides search professionals with two significant advantages that can alter their workflows. Firstly, it offers a tangible "demand signal" for prioritization, a distinct input compared to the data available in Search Console. Secondly, "share of voice" emerges as a metric that will likely appear in monthly reports, regardless of whether its complexities are fully understood by all stakeholders.
However, as Brodie Clark noted, the current iteration of these reports shares similarities with the AI reporting in Search Console, offering limited direct actionability. While the inclusion of query data is a positive step, its aggregated nature means individual user queries remain inaccessible.
Geographic and Account Limitations
It’s important to note that the AI Performance Insights feature is not yet widely available. Most merchants have not yet received access, and simply having an eligible Merchant Center account is insufficient, as the pilot is currently limited to a select number of US accounts. The definitive check for access is the appearance of an "AI Performance" tab within the Merchant Center interface. Until this tab is visible, a retailer’s AI reporting will likely be confined to the impression data available in Search Console.
Furthermore, websites that do not operate with a product feed will not receive these grouped shopping-query insights. Affiliate sites, review platforms, and editorial teams that produce buyer guides compete for visibility within AI Mode alongside the brands they cover. Their reporting remains confined to Search Console’s impression data, which, like the Merchant Center pilot, covers AI Mode and AI Overviews but lacks any query dimension. Consequently, the gap in detailed AI insight that this pilot aims to narrow for merchants remains substantial for other types of online content creators.
Challenges for Agencies and Reporting
For digital marketing agencies, the "share of voice" metric presents a unique challenge. While it can be effectively presented in client-facing presentations or decks, its utility in direct conversations is diminished. This is primarily due to its relativity to a competitor set that the merchant did not choose and cannot modify. A "zero" share of voice might simply indicate a low volume of impressions, while a "100%" could signify an empty competitor list. Both scenarios, when presented without proper context, could be misinterpreted as poor performance during quarterly reviews.
Unmeasurable Elements and Future Developments
The most significant data point still missing from these AI performance reports is the measurement of clicks. After a year of extensive coverage on the topic of AI in search, the question of how generative AI impacts user clicks remains largely unanswered by Google’s reporting tools. While impressions indicate how often a link to a product appeared, the direct link between AI-generated answers and user engagement is yet to be quantified in these new dashboards. John Mueller, Google’s Search Advocate, has previously clarified the methodology behind counting AI search impressions, and these rules also apply to the "share of voice" metric, which is impression-based.
Individual user queries are also currently inaccessible, and details regarding the defined competitor sets have not been disclosed. Google has not yet announced whether grouped query information will eventually be integrated into Search Console, particularly for websites that do not maintain product feeds.
Expanding Horizons and Regulatory Pressures
Google has indicated that the pilot program is slated for expansion to Australia, Canada, India, and New Zealand in the coming months. This broader rollout will provide further insights into the reliability and effectiveness of these metrics across a larger and more diverse set of accounts and product categories.
The overarching question remains whether these new reporting capabilities will eventually cross over to other Google platforms. In June, Google stated its intention to add metrics to Search Console reports over time, though specific details regarding which metrics and the timeline for their implementation remain unspecified.
The CMA’s nine-month implementation window for the UK market serves as a significant fixed point. This mandate specifically covers engagement reporting, including clicks and click-through rates for UK publishers interacting with generative AI features. However, this pertains to a different dashboard, a different audience, and a different jurisdiction than the current Merchant Center pilot.
The ongoing development and phased release of these AI performance insights underscore Google’s commitment to adapting to the evolving search landscape. While the current pilot offers valuable, albeit preliminary, data for retailers, the broader industry awaits further advancements in measurement, particularly regarding user clicks and individual query data, to fully grasp the impact of generative AI on e-commerce and search visibility.
Related posts:
