Let me introduce you to Milena Traikovich, a seasoned expert in demand generation who has dedicated her career to helping businesses craft impactful campaigns that attract and nurture high-quality leads. With her deep knowledge of analytics, performance optimization, and lead generation strategies, Milena offers invaluable insights into the evolving landscape of digital marketing. Today, we dive into a fascinating discussion about recent revelations regarding ChatGPT link interactions, exploring what the data tells us about user behavior, the potential of AI-driven traffic, and how marketers can adapt to these emerging trends.
How did you first come across the data on ChatGPT link interactions, and what stood out to you the most?
I stumbled upon this leaked OpenAI file through industry discussions, and I was immediately struck by the stark contrast between the massive number of link impressions and the incredibly low click-through rates. Seeing over 610,000 impressions for a top-performing page result in just over 4,000 clicks, boiling down to a 0.69% CTR, was eye-opening. It really highlighted how visibility in AI platforms like ChatGPT doesn’t necessarily translate to engagement.
What do you make of the variation in click-through rates across different pages, especially with some showing as low as 0% or 0.01%?
The variation is quite telling. A top page hitting 1.68% CTR shows there’s potential, but the majority hovering at 0.01% or even 0% suggests that most links are either irrelevant to users or poorly positioned. It could be due to the context in which the links are presented—perhaps the content around them doesn’t compel users to click, or the links themselves don’t seem trustworthy or valuable. It’s a signal that placement and relevance are critical, even in AI-driven environments.
Can you break down how OpenAI tracks impressions and clicks across different zones like the main response, sidebar, and citations?
From the data, it’s clear OpenAI is meticulous about tracking where links appear and how users interact with them. They categorize impressions and clicks by zones like the main response area, sidebar, citations, and even search results. Each zone has its own CTR calculation, which helps paint a picture of user behavior in specific contexts. For instance, the main response area racks up huge impressions but dismal clicks, while sidebar and citations fare better despite fewer views. This granular approach shows they’re trying to understand engagement at a very detailed level.
Why do you think the main response area, despite having the most visibility, struggles with such a low click-through rate?
I think it comes down to user intent. When people see links in the main response, they might feel the AI has already summarized the key information, so there’s little incentive to click through. It’s almost like the content is too self-contained. Plus, users might not trust or prioritize links embedded in a generated response as much as they would in a more curated or familiar setting like a search engine results page. It’s a behavioral shift we’re seeing with AI tools.
How do you interpret the higher CTR of 6–10% in sidebar and citation links compared to other areas?
Sidebar and citation links likely perform better because they’re perceived as more credible or intentional. Users might see citations as a direct source of the information provided, giving them a reason to click for verification or deeper reading. Sidebars might also stand out visually, drawing attention away from the main text. Even with fewer impressions, these zones seem to tap into a user’s curiosity or need for authority, which boosts engagement.
What insights can we draw from the almost nonexistent impressions and zero clicks for links in search results within ChatGPT?
It’s intriguing, isn’t it? This suggests that ChatGPT’s search results feature isn’t a primary touchpoint for users, or perhaps it’s not as prominently displayed or intuitive to navigate. If users aren’t even seeing these links, it could be a design or functionality issue. Alternatively, users might be so focused on the conversational aspect of ChatGPT that they bypass search-like features altogether. It shows that not all zones are created equal in terms of driving attention.
Based on this data, how realistic is it for publishers to view ChatGPT as a significant traffic source to replace losses from traditional organic search?
Honestly, the data paints a pretty sobering picture. With an average CTR of less than 1% for most pages, ChatGPT isn’t poised to replace Google organic search traffic anytime soon. AI-driven traffic is growing, no doubt, but it’s still just a tiny fraction of overall traffic for most publishers. The behavior is different—users are looking for quick answers, not necessarily to explore further. Publishers shouldn’t bank on this as a primary channel right now; it’s more of a supplementary opportunity.
How should marketers and content creators adjust their strategies given these low click-through rates on ChatGPT links?
Marketers need to pivot toward optimizing for visibility and relevance within the AI context. First, focus on creating content that answers very specific, conversational queries since that’s how users interact with ChatGPT. Second, consider how your links are framed—can they be positioned as authoritative sources or citations? Finally, don’t rely on AI platforms for traffic; instead, use them to build brand awareness and ensure your content is part of the conversation, even if clicks are minimal. It’s about playing the long game.
Looking at the broader picture, how do you see AI-driven traffic evolving for publishers over the next few years?
I believe AI-driven traffic will grow as more users adopt tools like ChatGPT for everyday queries, but it won’t mirror traditional search behavior. It’ll likely remain a smaller, niche source focused on instant answers rather than exploration. However, as AI platforms refine their interfaces and link placements, we might see incremental improvements in CTR. Publishers who adapt early—by crafting content tailored to AI responses and building trust within these ecosystems—will be better positioned. What’s your forecast for AI-driven traffic in the digital marketing space?
