Note: This analysis is based on “ChatGPT Referrals to E-Commerce Websites: Do LLMs Outperform Traditional Channels?” by Maximilian Kaiser and Christian Schulze (October 2025), which represents the first large-scale empirical study comparing organic LLM traffic performance to traditional digital marketing channels.
When ChatGPT began offering direct shopping links in August 2024, industry observers predicted a seismic shift that would fundamentally reshape e-commerce. The narrative was compelling: AI-powered recommendations would outperform traditional digital marketing, potentially making ChatGPT a “Google killer.” But here’s the reality check that’s sending shockwaves through the industry: new research reveals ChatGPT shopping links are dramatically underperforming expectations.
A comprehensive academic study analyzing 12 months of performance data from 973 e-commerce websites with $20 billion in combined revenue has shattered the AI shopping hype. The findings are stark: organic LLM traffic outperforms only paid social media while getting crushed by every other traditional channel, including the very Google search it was supposed to displace.
The Numbers Don’t Lie: ChatGPT’s Conversion Crisis
The data reveals a brutal truth about AI shopping that no one in the industry wants to admit. While early studies reported oLLM conversion rates of 6.7% versus 3.9% for organic search, the largest-scale analysis to date tells a drastically different story.
Organic search: the channel everyone expected ChatGPT to destroy: outperforms ChatGPT by approximately 13%. But that’s just the beginning of AI’s performance problem. Affiliate marketing demolishes ChatGPT with an 86% greater likelihood of conversion. Even paid search, email marketing, and direct traffic leave ChatGPT in the dust.
Only paid social media performs worse than ChatGPT shopping links, and that gap is narrowing as platforms optimize their commerce features. This means ChatGPT currently ranks as the second-worst performing major traffic channel for e-commerce conversions.
The study examined over 50,000 oLLM transactions alongside 164 million transactions from traditional channels: a sample size that dwarfs previous research and makes these findings impossible to dismiss as statistical noise.
Why the AI Shopping Revolution Stalled
The disconnect between ChatGPT’s theoretical advantages and real-world performance stems from three fundamental problems that are reshaping how we must approach AI-optimized commerce strategies.
Platform Immaturity Is Killing Conversions
ChatGPT remains optimized for conversation, not commerce. Only 2.1% of ChatGPT conversations involve purchasable products, revealing a platform that hasn’t fully committed to shopping optimization. While Google has spent decades refining its understanding of commercial intent, ChatGPT is essentially starting from scratch in the commerce game.
The Optimization Gap Creates a Vicious Cycle
Here’s the catch-22 that’s trapping ChatGPT traffic: weak performance depresses investment incentives, which leads to poor website optimization for AI traffic, which further weakens performance. Most retailers haven’t optimized their sites for the unique behavioral patterns of LLM-driven visitors, creating a coordination problem that’s difficult to solve.
Consumer Behavior Reveals a Trust Problem
Users are treating ChatGPT as a research tool, not a purchase platform. They’re getting product recommendations from AI, then completing purchases through traditional channels they trust. Under last-click attribution models, this behavior assigns conversions to other channels, making ChatGPT’s actual influence invisible in standard analytics.
This behavioral pattern fundamentally changes how businesses must approach organic SEO strategies and attribution modeling to accurately measure AI’s impact on their sales funnel.
The Hidden Engagement Story That Changes Everything
Despite poor conversion performance, ChatGPT traffic shows surprising engagement signals that reveal untapped potential. Users referred by ChatGPT have favorable bounce rates compared to most traditional channels, suggesting high relevance when clicks occur.
However, these engaged users spend less time per session and view fewer pages, indicating a more targeted browsing behavior. They arrive knowing what they want but leave without purchasing: a pattern that suggests optimization opportunities are being completely missed.
This engagement data reveals that ChatGPT users aren’t low-quality traffic; they’re high-intent visitors that current e-commerce sites are failing to convert. The problem isn’t the traffic source: it’s how businesses are handling that traffic once it arrives.
The Trajectory That Should Worry Google
While current performance disappoints, the improvement trends tell a different story. Both conversion rate and revenue per session have consistently improved throughout the first 12 months of ChatGPT shopping links.
The trajectory suggests conversion rates will continue narrowing the gap with traditional channels, though projections indicate ChatGPT won’t reach parity with organic search within the next year. However, the consistent month-over-month improvements reveal a platform rapidly learning and optimizing.
Average order value declined during the study period, suggesting users are purchasing lower-value items or that early optimization focused on conversion probability rather than basket size. This presents a clear opportunity for businesses that can crack the code on maximizing AI-driven order values.
What This Means for Your E-Commerce Strategy Right Now
The research findings have immediate implications that demand strategic shifts in how businesses approach digital marketing and traffic optimization.
Traditional Channels Aren’t Going Anywhere
The data provides zero support for narratives positioning LLMs as immediate “Google killers.” Organic search, paid search, email marketing, and affiliate programs continue delivering superior conversion rates and revenue per session. Businesses abandoning these proven channels for unproven AI strategies are making a costly mistake.
AI Requires Different Optimization Approaches
ChatGPT traffic behaves fundamentally differently than traditional search traffic. These users arrive with specific intent but need different conversion paths, trust signals, and user experiences. Standard e-commerce optimization playbooks don’t apply to AI-driven traffic.
The Attribution Model Is Breaking
Current attribution models fail to capture AI’s influence on the customer journey. Businesses relying on last-click attribution are missing ChatGPT’s actual impact on their sales funnel. This invisible influence means many companies are underinvesting in AI optimization because they can’t measure its true contribution.
The Action Plan for Forward-Thinking Businesses
Smart e-commerce managers are already adapting their strategies to capture AI’s growing influence while maintaining strong performance in traditional channels.
Start Optimizing for AI Traffic Now
Begin testing different landing page experiences for AI-referred traffic. These visitors need streamlined purchase paths, enhanced trust signals, and mobile-optimized experiences that accommodate their browsing patterns.
Implement Advanced Attribution Modeling
Move beyond last-click attribution to understand AI’s influence on your customer journey. Consider implementing first-touch, time-decay, or data-driven attribution models that capture ChatGPT’s research-phase influence.
Maintain Traditional Channel Excellence
The biggest mistake businesses can make is neglecting proven channels while chasing AI trends. Double down on organic SEO excellence while gradually building AI optimization capabilities.
The Long-Term View: Patience Required, Preparation Essential
One year after launch, ChatGPT shopping represents an interesting but not yet transformative e-commerce channel. The platform shows promise through improving conversion trends and high engagement signals, but it’s nowhere near displacing traditional channels.
The message for e-commerce leaders is clear: pay attention to AI traffic and begin optimizing for it, but don’t abandon traditional strategies that continue delivering superior results. The AI shopping revolution is happening, but it’s taking significantly longer than the hype suggested.
As LLM platforms enhance their commerce capabilities, retailers optimize their sites for AI traffic patterns, and consumers develop greater trust in AI recommendations, the performance gap will likely narrow. But for now, the data shows that rumors of Google’s death have been greatly exaggerated.
The businesses that will win in this evolving landscape are those that optimize for both worlds: maintaining excellence in traditional channels while building capabilities for the AI-driven future that’s slowly taking shape.
Ready to develop a comprehensive digital marketing strategy that balances traditional excellence with AI preparation? Book a consultation to discuss how your business can thrive in this changing landscape.
