ChatGPT Access Reduces Traditional Search Queries by 9.4%
A new study from Bocconi University reveals that broader access to ChatGPT has reduced weekly traditional search queries by 9.4%. After 20 weeks, the decline reaches 17%. For businesses relying on organic search traffic, this is a structural shift worth understanding.
The research, based on U.S. desktop Comscore clickstream data, shows that ChatGPT directs users to external websites in only 5.2% of sessions, compared to 31.1% for Google. And when ChatGPT does send traffic, it favors non-profit, subscription, and freemium platforms—not ad-supported sites.
How ChatGPT's Referral Traffic Differs from Google
ChatGPT's referral traffic is 27.6 percentage points less likely to go to ad-supported sites compared to Google's. Instead, it sends users to reference and knowledge sites, tools, SaaS platforms, and academic or developer sites. This means publishers who depend on ad revenue from search referrals face a double hit: fewer visits and lower monetization potential per visit.
The study also notes that ChatGPT's monthly referral rate grew from 2.5% to 6.5% during the measurement period, indicating increasing engagement. However, the absolute referral volume remains far below Google's.
Which Search Categories Are Most Affected?
The biggest declines are in informational searches. Academic research referrals dropped 32.8%, and reference searches fell 26.5%. Transactional and recreational searches barely changed. This suggests ChatGPT is primarily substituting for queries where users want quick answers, not for shopping or entertainment.
For businesses that create content targeting informational keywords—like how-to guides, definitions, or research summaries—the risk is highest. If your traffic comes from people asking 'what is' or 'how does,' ChatGPT is likely eating into your audience.
What This Means for Your Business
If your business relies on ad-supported content or informational search traffic, you need to adapt. Consider diversifying revenue streams—subscriptions, freemium models, or direct partnerships. Also, optimize your content for AI-generated answers: structure it clearly, use schema markup, and aim to be the source ChatGPT cites.
For e-commerce and transactional sites, the impact is minimal so far. But as AI search evolves, monitoring referral patterns will be critical.
Bottom Line: A Structural Shift, Not a Fad
The Bocconi study provides hard data that AI search is not just a novelty—it's changing how people find information. The 9.4% drop in search queries is significant, and the trend is accelerating. Businesses that treat this as a temporary blip risk losing traffic and revenue. The smart move is to start experimenting with AI-optimized content and alternative monetization now.
FAQ
By 9.4% on average, with a 17% reduction after 20 weeks of broader access.
Ad-supported sites lose 27.6 percentage points of referral traffic compared to Google. Academic and reference sites see drops of 32.8% and 26.5%.
Not yet—transactional searches barely changed. But monitor trends as AI search evolves.
