13 Months of LLM Data: Traffic, Growth & Conversions Revealed
▼ Summary
– LLM referral traffic currently accounts for less than 2% of total referral traffic on average, making it a small but growing topic of interest.
– This traffic is growing rapidly, with an average 80% increase observed from the first to the second half of 2025.
– The sources (like YouTube or Reddit) that LLMs cite in their responses are shifting quickly, which can impact site traffic and content strategy.
– Despite low volume, LLM referrals convert at a very high rate (approximately 18%), outperforming other channels like SEO and PPC.
– Brands should establish dedicated monitoring for this volatile channel and optimize for the high-intent users it delivers to prepare for future growth.
Understanding the impact of Large Language Models on website traffic is a pressing concern for modern brands. While the conversation is dominated by speculation, the most effective strategy begins with analyzing concrete data to separate trends from hype. A detailed examination of referral traffic from LLMs like ChatGPT and Perplexity over thirteen months reveals critical patterns that can inform a smarter, more proactive digital strategy.
Our analysis of this dataset highlights four pivotal findings. First, LLM referral traffic currently represents a minor portion of overall site visits, averaging less than two percent of total referral volume. Traffic sources range from models like Gemini and Claude, contributing between 0.15% and 1.5%. This indicates that, for many businesses, this channel is not yet a primary driver of immediate revenue, despite its prominence in industry discussions.
Second, while the volume is small, the growth rate of LLM referral traffic is exceptionally rapid. Comparing the first and second halves of the period studied, we observed an average increase of 80%. Growth was not uniform, with some organizations seeing jumps as high as 300%. This accelerating trend underscores the importance of tracking not just current volume but the velocity of growth, as consumer adoption expands and underlying algorithms evolve.
A third key insight involves the dynamic nature of source citations. The platforms and content types that LLMs reference in their responses are undergoing significant shifts. Recent monitoring of thousands of prompts shows a notable rise in citations linking to YouTube and, previously, Reddit. These changes directly influence which sites receive traffic and should inform broader content and advertising planning. Since LLMs do not transparently report this data, accessing it requires specialized third-party tools.
Perhaps the most compelling discovery is the fourth: LLM referrals demonstrate a remarkably high conversion rate. Our data shows these visitors convert at approximately 18%, a rate that surpasses channels like paid search, SEO, and direct traffic. This suggests users arriving via an LLM are highly qualified, often having had their query validated before clicking through. However, this high-value traffic currently constitutes a volume roughly twenty-five times smaller than traditional SEO.
Given these findings, brands should take several actionable steps. Establishing dedicated monitoring is essential. This means tracking the growth velocity of LLM referrals and using available tools to see which sources are driving citations. Furthermore, businesses must capitalize on this high-value traffic by analyzing the user journey for these visitors. Understanding what content they land on and what intent they exhibit allows for optimization targeting this premium audience.
Finally, planning for future growth is non-negotiable. Developing a content strategy tailored for AI discovery is crucial. This involves creating material that positions your brand as an authoritative source LLMs will choose to cite and summarize, which differs from conventional SEO practices. Allocating a modest budget for tools and resources focused on this channel, even before it reaches critical mass, is a prudent investment.
LLM referral traffic is transitioning from an emerging curiosity to a strategic signal. Its current small share is counterbalanced by explosive growth, evolving source patterns, and outstanding conversion quality. The prudent approach is not to overreact but to monitor trends closely, understand citation origins, and observe audience behavior. By staying data-focused, organizations can strategically position themselves to leverage this fast-moving channel as it continues to develop.
(Source: Search Engine Land)
