Retailers are receiving more visitors from ChatGPT when users ask about products and follow its links to ecommerce pages. However, the share is still very small. The study analyzed[1] first-party analytics from 973 online stores with 20 billion dollars in combined revenue and found that ChatGPT accounted for about 0.2 percent of all sessions, roughly two hundred times smaller than organic Google Search volumes according to the working paper. Traffic from other large language models such as Gemini, Perplexity, and Copilot appeared negligible in comparison.
Conversion Rate Behind Search and Affiliate Channels
The researchers used regression models to compare more than 50,000 ChatGPT-referred transactions with 164 million orders from traditional acquisition channels. The results showed organic search and affiliate marketing produced higher conversion rates than ChatGPT referrals. Paid social performed worse. Organic and paid search also generated stronger revenue per session than ChatGPT, while ChatGPT still exceeded paid social on that metric. Average order values from ChatGPT referrals declined as adoption grew, based on the same dataset.
Bounce Rate Low, Deeper Browsing Limited
Engagement measured through bounce rate and session depth showed mixed outcomes. Bounce rates were lower than most channels except search. The paper states that “session depth was generally lower than most channels” which indicates fewer pages viewed after arrival. Users following AI assistant recommendations appear to visit a product page, assess it briefly, and then leave if it does not match expectations.
Early Stage Behavior Shifts Where Sales Get Counted
The findings propose an explanation related to where the sale is finalized. The study notes evidence of “trust and verification behavior,” where users review information found through ChatGPT but switch to familiar channels like Google Search before checking out. That behavior shifts last-click attribution toward traditional channels, even when ChatGPT initiated the buying interest.
Performance Improving Gradually
ChatGPT referrals improved over time on two measures. The researchers reported increasing conversion rates and revenue per session through the twelve-month period. Trend projections indicated continued growth but not enough to match organic search performance within the next year according to the paper’s forecasts. The gap remains substantial because search continues to control most purchase-driven demand.
Scale of Impact Still Limited
Across all findings, ChatGPT contributed visible but small commercial value. The working paper focuses on a channel in early development and recognizes data sparsity challenges in ecommerce categories that receive few AI-driven visits. It also highlights that more integrated buying features inside AI platforms could change referral patterns, but the current dataset does not yet reflect that shift.
Consistency With Other Available Results
The conclusions match what other researchers have observed. SALT.agency’s measurement reported that organic search sessions show higher engagement or goal-completion rates in most sectors, with exceptions in categories such as health and careers. Another analysis from Amsive found no statistically significant difference in conversion between LLM sessions and organic traffic, while showing that LLM traffic represented less than one percent of total traffic across tracked sites. These independent findings align with the same pattern: low scale and lower commercial intent relative to search engines.
Notes: This post was edited/created using GenAI tools. Image: DIW-Aigen.
Read next:
• AI Bots Spark Conversation but Don’t Change Posting Habits[2]
• Erase.com Explains the Hidden ROI of Online Reputation Management[3]
References
- ^ analyzed (papers.ssrn.com)
- ^ AI Bots Spark Conversation but Don’t Change Posting Habits (www.digitalinformationworld.com)
- ^ Erase.com Explains the Hidden ROI of Online Reputation Management (www.digitalinformationworld.com)