Expanding access to ChatGPT Search has reduced traditional search engine traffic by 9.4% on average among U.S. desktop users, according to new research from Bocconi University analyzing October 2024 through July 2025. The decline has deepened further over time, reaching 17.0% after 20 weeks, and varies sharply by query type and destination website. For digital marketers, web developers, and SEO professionals, this shift raises a critical question: how much referral traffic you actually lose depends less on ChatGPT's query volume than on what the platform does with those queries.
Table of Contents
- How ChatGPT Search Diverts Traffic Differently Than Google
- Research, Reference, and How-To Queries Take the Hardest Hit
- Different Destinations See Different Traffic Flows
- Google's Dominance Remains Largely Intact
- What You Should Monitor and Adjust
- Frequently Asked Questions
How ChatGPT Search Diverts Traffic Differently Than Google
ChatGPT search's growth has fractured the search landscape in ways traditional metrics mask. According to analysis of Comscore data from 45,000+ U.S. households, the platform directs users to external websites in only 5.2% of sessions, compared to 31.1% for google searches.
Even though ChatGPT now handles roughly 12% of Google's query volume, it sends 190× less referral traffic to websites—accounting for just 0.21% of referral traffic while Google maintains around 40%. This disparity exists because ChatGPT often synthesizes answers from training data without clicking through to live web results. In fact, only 34.5% of ChatGPT queries trigger live web search, down from 46% in late 2024, meaning the majority of responses never route users to your site at all.
Research, Reference, and How-To Queries Take the Hardest Hit
ChatGPT's impact is not uniform across query types. Research, reference, and how-to queries lost the most traffic to ChatGPT, while transactional searches (buying products, checking prices) and recreational queries (entertainment, games) declined less.
This matters directly to your content strategy: if your traffic comes primarily from informational content—tutorials, documentation, reference guides, troubleshooting posts—you are in a higher-risk category than sites driving conversion-focused or entertainment traffic. The 9.4% average decline masks significant variation. Readers accustomed to typing "how do I fix X" or "explain how Y works" now have an alternative that often answers within the chat interface itself, bypassing search engines and websites alike.
Different Destinations See Different Traffic Flows
ChatGPT and Google route traffic to different types of sites. ChatGPT directs traffic to reference sites, SaaS platforms, and academic or developer resources, while Google still favors YouTube, Reddit, and Wikipedia.
If your business depends on traffic from general-audience Google search results, the aggregate 9.4% decline may not affect you as severely. However, if your site is a reference resource, technical documentation hub, or developer platform, ChatGPT's traffic diversion poses a more direct threat. This distribution reflects ChatGPT's training and design: the platform prioritizes authoritative sources and structured data, steering users toward sites with clear expertise signals rather than social platforms or video.
Google's Dominance Remains Largely Intact
Despite ChatGPT Search's expansion, Google continues to process far more queries and drive far more referral traffic. Google's global search market share remains 89.87%, and the platform processes 8.4× more daily queries than ChatGPT.
Google's referral share has plateaued around 88%, meaning traditional SEO and paid search continue to drive the bulk of external traffic to most sites. The practical takeaway: while ChatGPT creates genuine competitive pressure for certain query categories, it has not become a primary traffic driver for most websites. The 9.4% decline in traditional search volumes is real, but does not correspond to a 9.4% loss in external referral traffic for the average site.
What You Should Monitor and Adjust
For SEO and marketing teams, the immediate focus should be understanding which of your queries are vulnerable and which remain stable. Track your traffic trends by query type and source, paying special attention to how-to, troubleshooting, and reference content where ChatGPT competes most directly. If a significant share of your traffic comes from those categories, monitor referral metrics closely over the next 6–12 months to detect accelerating decline.
On the paid side, ChatGPT Search does not yet appear to disrupt Google Ads meaningfully—conversion-focused queries still route through Google in the vast majority of cases. However, brand-building and informational campaigns targeting research and reference queries may warrant budget review. Consider testing whether your reference content can be repurposed for ChatGPT (through integration partnerships or direct submission) to capture value even when users never click through to your site.
Frequently Asked Questions
Will ChatGPT Search eventually become a bigger threat to my traffic?
It depends on your content type. Reference, how-to, and research content faces the most pressure. Transactional and entertainment content remains largely unaffected. Monitor trends quarterly rather than reacting to early data.
Should I optimize my site for ChatGPT Search?
No official ChatGPT optimization exists yet. Focus on SEO fundamentals—clear structure, authoritative content, and fast performance. These serve both traditional search and any future ChatGPT integrations.
Does this change my Google Ads strategy?
Unlikely in the short term. Google Ads reaches users actively searching Google, and conversion-focused queries remain dominated by Google. Reference and educational campaigns may see slower ROI if your target audience shifts to ChatGPT for answers.
