Sign in Subscribe
Data

Google to fix AI Mode tracking bug affecting referrer data

Google confirms AI Mode referrer tracking issue was unintentional and will fix noreferrer elements.

Luis Rijo

• 5 min read
HTML code showing Google AI Mode links with noreferrer attributes that block referrer tracking.
HTML code showing Google AI Mode links with noreferrer attributes that block referrer tracking.

Google confirmed on May 23, 2025, that tracking issues affecting its recently launched AI Mode feature were the result of an unintentional bug. The issue prevented website owners from accurately tracking referrer data from AI Mode clicks, which Google executives acknowledged as unexpected behavior that required immediate attention.

According to the Search Engine Roundtable report published May 23, 2025 - one day before today's date of May 24, 2025 - industry experts Tom Critchlow and Patrick Stox first identified the problem after Google released AI Mode in the United States several days earlier. The feature, which provides enhanced search capabilities using advanced AI technology, was inadvertently stripping referrer information from outbound links.

John Mueller, Senior Search Analyst from Google's Search Relations team, addressed the issue directly in LinkedIn comments and Reddit discussions. Mueller characterized the situation as "unexpected" and confirmed that Google's internal teams were "considering this a bug" that would be resolved through upcoming fixes.

The technical root of the problem involved noreferrer elements embedded within AI Mode's link code structure. According to Patrick Stox's analysis, these elements caused referrer values to be stripped from outbound clicks, resulting in traffic appearing as "unknown" in most analytics platforms. This classification meant that clicks from AI Mode would register as direct traffic rather than being properly attributed to Google's AI-powered search feature.

The referrer policy implementation in AI Mode differed significantly from standard Google search results. While traditional Google search links typically preserve referrer information to help website owners understand traffic sources, the AI Mode implementation included HTML attributes that completely eliminated this tracking capability. Tom Critchlow noted the comprehensive nature of the referrer stripping, observing that clicks from AI Mode were "completely strip[ping] the referrer" rather than simply redirecting through Google's standard tracking mechanisms.

The technical specifications surrounding referrer policies involve HTTP headers that control how much referrer information gets included with requests. According to Mozilla's documentation, the noreferrer directive specifically omits the Referer header entirely, meaning that requests include no referrer information whatsoever. This approach contrasts with other referrer policy options that might preserve origin information while protecting more detailed path and query parameters.

Google's AI Mode represents a significant expansion of the company's search capabilities, introducing multimodal search features powered by advanced language models. The feature allows users to conduct more complex queries using natural language processing and provides enhanced result presentation formats. However, the referrer tracking issue threatened to create significant gaps in website analytics data, particularly affecting businesses that rely on detailed traffic attribution for marketing optimization.

The impact extended beyond simple traffic counting concerns. Website owners use referrer data to understand user behavior patterns, optimize content strategies, and measure the effectiveness of search engine optimization efforts. When referrer information gets stripped, analytics platforms like Google Analytics typically categorize the traffic as "direct," which obscures the actual traffic source and can skew marketing attribution models.

John Mueller's response timeline provided insight into Google's internal bug tracking processes. His initial LinkedIn comment, posted deep within discussion threads, acknowledged the issue as "being tracked as a bug." Mueller subsequently elaborated on Reddit, stating that Google was "currently considering this a bug" and expressing expectation that it would be resolved, while apologizing for the unexpected behavior.

The Search Engine Roundtable documentation included visual evidence of the problematic code implementation. Screenshots revealed HTML link elements containing specific attributes that triggered the referrer stripping behavior. The technical implementation showed how noreferrer attributes were being applied systematically across AI Mode links, suggesting the issue stemmed from a broader configuration problem rather than isolated incidents.

The timing of the bug discovery coincided with Google's broader rollout of AI-powered search features across multiple platforms. The company has been integrating artificial intelligence capabilities throughout its search ecosystem, including the introduction of AI Overviews and enhanced question-answering features. This context suggests that the referrer policy issue might have emerged from rapid feature deployment rather than intentional design decisions.

Industry commentary on the incident revealed skepticism about whether the referrer stripping was truly accidental. A comment from a user identified as DataDonkey on the Search Engine Roundtable suggested possible intentional implementation during rollout phases. The commenter speculated that Google might have temporarily disabled referrer tracking to manage traffic expectations while adjusting other AI Mode features.

The broader implications of the bug extend to Google's relationship with website publishers and content creators. Accurate referrer tracking plays a crucial role in the digital advertising ecosystem, helping publishers understand traffic sources and optimize monetization strategies. When referrer data gets compromised, it affects revenue attribution and can complicate relationships between search engines and content providers.

Google's approach to fixing the bug involved careful rollout timing considerations. Mueller noted that the resolution would be delayed due to a "long weekend coming up," with the company preferring to avoid deploying changes immediately before extended periods. This cautious approach reflects standard software development practices that minimize risks associated with weekend deployments when support staff availability might be limited.

The incident also highlighted the interconnected nature of Google's search and analytics ecosystems. When referrer policies change unexpectedly, the effects cascade through multiple Google products, including Google Analytics, Search Console, and various third-party analytics platforms that rely on referrer data for attribution modeling.

Search Console integration emerged as another aspect of the AI Mode rollout. Mueller confirmed that "AI Mode reporting is not yet live in Search Console" but indicated that documentation updates would reflect the upcoming availability of AI Mode metrics within the platform. This suggests that Google had planned comprehensive analytics integration for AI Mode but encountered technical obstacles during implementation.

The referrer policy bug serves as a case study in the complexity of modern search engine feature rollouts. As search engines incorporate increasingly sophisticated AI capabilities, maintaining compatibility with existing web standards and analytics infrastructure becomes more challenging. The incident demonstrates how seemingly minor technical configurations can have significant downstream effects on the broader web ecosystem.

For the marketing community, this incident represents both a cautionary tale and an opportunity for deeper understanding of referrer policy implications. Digital marketers who depend on accurate traffic attribution must now consider how emerging AI-powered search features might affect their analytics data. The bug also underscores the importance of diversified traffic tracking methods that don't rely exclusively on referrer headers.

The resolution timeline remains somewhat uncertain, though Mueller's communications suggest that Google treated the issue as a priority. The company's acknowledgment of the bug and commitment to fixing it indicates recognition of the importance of maintaining accurate analytics data for website owners and marketing professionals.

This situation illuminates the delicate balance between innovation and infrastructure compatibility in search engine development. As Google continues expanding AI capabilities across its search platform, maintaining backward compatibility with existing web standards becomes increasingly complex. The referrer tracking bug demonstrates how technical implementation details can have far-reaching consequences for the digital marketing ecosystem.

The marketing community's response to this incident likely involves increased scrutiny of AI-powered search features and their potential impact on analytics accuracy. Marketing professionals may need to develop new methodologies for tracking and attributing traffic from AI-enhanced search results, particularly as these features become more prevalent across search platforms.

Google's handling of the incident also provides insight into the company's communication strategies around technical issues. Mueller's direct engagement in community forums demonstrated a commitment to transparency, though the discovery of the bug through external observers rather than internal quality assurance processes raises questions about testing procedures for new feature rollouts.

Timeline

The Google AI Mode referrer tracking bug timeline demonstrates the rapid identification and acknowledgment of an unexpected technical issue:

May 20-21, 2025: Google releases AI Mode feature in the United States 

May 22-23, 2025: Industry experts Tom Critchlow and Patrick Stox identify referrer tracking problems 

May 23, 2025: John Mueller confirms the issue as a bug in LinkedIn and Reddit discussions 

May 23, 2025: Search Engine Roundtable publishes detailed analysis of the technical problem 

May 24, 2025: Google commits to fixing the noreferrer elements while delaying deployment due to weekend timing

Stories like this, in your inbox

Enter your email
Subscribe