In October 2025, ChatGPT had more than 800 million weekly users; yet, most analytics dashboards don’t even list it as a traffic source.

There’s an entirely new visitor channel quietly growing faster than Google Search, sending higher-intent users who convert at rates that make traditional search traffic look inefficient.

In this guide, you’ll learn what makes ChatGPT traffic unique, how to track it in your analytics, and strategies to increase both direct and indirect visits from this AI assistant.

chatgpt-vs-google.com)
  • Growth trajectory: ChatGPT’s traffic share grew 3× between January and September 2025 (source: chatgpt-vs-google.com)
  • Monthly growth: ChatGPT grows 14.1% every month while Google shrinks 3.2% monthly (source: chatgpt-vs-google.com)
  • Industry variation: ChatGPT shows 32× higher penetration in Finance (0.97%) compared to Autos (0.03%), indicating extremely uneven adoption across industries (source: chatgpt-vs-google.com)
  • User base: 800 million weekly active users (source: techcrunch.com)
  • Daily usage: 2.5 billion prompts daily (as of July 2025) (source:openai.com)
  • Web presence: chatgpt.com is the 9th most visited website globally with an estimated 586.4M organic visits each month (source: ahrefstop.com)
  • Search interest: The term “ChatGPT” gets searched on Google an estimated 326 million times globally each month (source: Ahrefs’ Keywords Explorer)
    • ~0.5% of visits but accounted for ~12.1% of sign-ups. That’s 23x higher rate than traditional organic search visitors in our case.

    And we’re not the only ones with higher conversion rates from LLM traffic. Recently, Buffer’s Director of Growth, Simon Heaton, reported similar results:

    Bar chart titled 'Aggregate Conversion Rate Comparison' showing LLM-driven traffic converts at 20.15%, which is 185.4% higher than organic search traffic's 7.06% conversion rate.

    Even some people in the comments chimed in similar data:

    Screenshot of a LinkedIn comment from Vignesh D confirming similar results of fewer visits but higher Sales Qualified Lead (SQL) conversions from AI-driven discovery like ChatGPT.

    Not everyone will get the same results, though. This study by Amsive suggested that LLM conversions depend on the business model. And there’s even a massive, 61-page research providing evidence that ChatGPT referrals convert worse when it comes to e-commerce.

    There’s mixed research on conversion rates across different websites, but visitors coming from ChatGPT might convert at much higher rates than those from regular search traffic. Either way, it’s worth running your own tests to see how it performs for your site.

    detailed guide by Amsive).

    An easier option is to use tools like Ahrefs’ Web Analytics, which already have AI traffic sources built in and automatically categorize these visits.

    Screenshot of an Ahrefs Web Analytics 'Traffic sources' report showing 'AI search' as a distinct channel with 225 visitors, indicated by a purple star icon and an orange arrow.

    Monitor the same metrics you track for other channels

    Once you’re tracking ChatGPT traffic, measure its performance using the same analytics framework you already use for Google Search or social media. The key difference: pay attention to how ChatGPT visitors behave compared to other channels.

    All of these metrics can be tracked in Ahrefs’ Web Analytics. Chances are, the tool you’re already using can do it too—as long as it supports event tracking and can identify LLM-driven traffic.

    Metric category What to track Why it matters
    Volume metrics (establish baseline) Total visits from ChatGPT; Unique visitors; Week-over-week and month-over-month growth rate Shows how much traffic you’re getting and whether it’s growing. Even small numbers matter since ChatGPT traffic typically represents just a fraction of total traffic but is growing rapidly.
    Engagement metrics (validate quality) Time on page; Bounce rate; Pages per session Reveals whether ChatGPT visitors are genuinely interested in your content. AI-referred visitors often spend more time and explore more pages because they arrive pre-qualified with clear intent.
    Conversion metrics (measure impact) Goal completions (leads, downloads, signups); E-commerce transactions; Conversion rate vs other channels Connects ChatGPT visibility to business outcomes. AI traffic often converts at dramatically higher rates—for Ahrefs, ChatGPT traffic was ~0.5% of visits but ~12.1% of signups (roughly 23× higher conversion rate than organic search).
    Content performance (reverse-engineer what works) Pages receiving ChatGPT traffic; Top landing pages from AI referrals; Topics generating the most visits Identifies patterns in what ChatGPT cites so you can replicate success. Helps you prioritize which content types and topics deserve more investment.
    AI search visibility tool, Ahrefs Brand Radar:

    1. Go to Brand Radar and enter your niche or market, then leave everything else blank and click Explore.
    2. Next, open the Cited pages report and look for patterns. For example, pages that explain new terms, offer how-to guides, or provide reviews.
    3. You can also switch between different AI indexes using the top filter to see how citation preferences vary between models.
    Ahrefs Brand Radar 'Cited pages' report screenshot filtered by the niche 'crm,' showing a list of external pages like TechRadar and Wikipedia that are frequently cited by ChatGPT for this topic.

    You can narrow the analysis down to a single competitor (or a few competitors) and find their best-performing content in ChatGPT:

    1. Enter your competitor’s brand name, and make sure to add their website’s address, too.
    2. Go to the Cited pages report.
    3. Limit the domain scope to your competitor’s domain.
    4. Look at the results.
    Ahrefs Brand Radar 'Cited pages' report screenshot for a competitor, Mailchimp, demonstrating how to filter the results to a single domain to analyze their top-performing content in ChatGPT.

    Ahrefs’ Keywords Explorer to pull data from Google’s much larger keyword database, giving you plenty of new ideas. Since what people search on Google often overlaps with what they ask ChatGPT, it’s a great way to spot additional opportunities.

    Here’s how:

    1. Enter a few broad topics related to your business (you can use the built-in AI for ideas or your preferred AI assistant).
    2. Next, open the Matching Terms report.
    3. Apply these filters: Volume: up to 500 (you can adjust this later for better results), Target: your domain, show ranking positions.
    Ahrefs Keywords Explorer screenshot of the 'Matching Terms' report, showing how to filter for low-volume keywords and a target domain to find long-tail content opportunities.

    If you see an “x” under ranking position, it means you don’t currently have a page targeting that keyword, and you can consider creating one. If the ranking position is greater than 10, you can likely improve your existing content (click the SERP icon on the right to see which URL is ranking).

    Recommendation

    There’s a simple trick to see which queries ChatGPT uses when it searches the web for a prompt—check out this short video to learn how it works.

    ChatGPT usually performs query fan-out on longer prompts, breaking them into several related searches to find the best-fitting information. So when you get an idea—say, from a social media comment or customer feedback—try this:

    1. Enter the prompt into ChatGPT.
    2. Use the query fan-out hack to see which search queries it generates.
    3. Take those queries and search them on Google to see what’s ranking, or plug them into a tool like Keywords Explorer to check their search demand.
    AI assistants like ChatGPT strongly prefer recently updated content.

    Bar chart comparing the average age of content cited by Organic SERP and AI assistants, showing that AI assistants (average 909 days since last updated) strongly prefer more recently updated content than organic search.

    Here are the types of pages worth updating regularly:

    • Key guides and tutorials—add new examples, visuals, or data.
    • Tool pages—highlight the latest features and improvements.
    • Industry statistics or benchmark reports—update figures and insights.

    Refreshing these pages helps you attract more traffic from both LLMs and Google. While Google Search doesn’t heavily favor recency, updating content is still a proven way to maintain first-page rankings.

    Moreover, if you find that ChatGPT cites an outdated study in an important topic for you, there’s a chance to replace it with your own or be cited next to it.

    If you know a popular study made by your competitor, simply plug it into Site Explorer and see if it gets cited in any of the 6 AI indexes we track.

    Ahrefs Site Explorer screenshot for a specific blog post, demonstrating how to check the number of AI citations, showing this page has 134 ChatGPT citations.

    Further reading

    AI Crawler Access Checker.

    Screenshot of the Ahrefs AI Crawler Access Checker tool for ahrefs.com, showing that crawlers like GPTBot and OpenAI are specifically set to 'Allowed' in the site's robots.txt file.

    Then, you need to ensure ChatGPT can easily find and understand your content. Here’s how:

    • Strengthen internal linking so AI crawlers can navigate your site structure.
    • Use clean formatting and semantic HTML (proper heading hierarchy, descriptive lists, clear sections).
    • Reduce JavaScript-heavy elements since most AI crawlers don’t render them well.
    • Avoid paywalls on key content if you want AI tools to reference it.
    • Use SEO-friendly URLs to indicate page content.
    we tested it). And actually, ChatGPT is the one that hallucinates the most.

    Bar chart comparing estimated 404 rates by referral channel, showing that ChatGPT has the highest 404 error rate at 1.01%, indicating it hallucinates broken links more often than other referrers like Perplexity and Gemini.

    Start by monitoring your 404 errors to spot issues like:

    • Hallucinated URLs. Pages ChatGPT “made up” based on your site’s URL structure.
    • Outdated URLs. Links AI models recall from old training data.
    • Misspelled or variant URLs. Common typos or slight variations of your real pages.

    In Ahrefs Web Analytics, there’s a dedicated report for potential 404 errors hallucinated by AI. Pages in this report may still receive traffic, but their title includes “404,” indicating that users (or AIs) are trying to access pages that don’t exist.

    Ahrefs Web Analytics 'Possible 404' report screenshot, showing a graph of visitors and a table listing unique visitors to pages titled '404,' which helps identify AI-hallucinated broken links that still receive traffic.

    If you don’t have a closely related page to redirect that traffic yet, you can create one so visitors (and search engines) land somewhere useful instead of hitting a dead end.

    However, prioritize URLs that are generating meaningful traffic. If a hallucinated URL only gets a handful of visits, it may not be worth the effort to create a new page for it (you can redirect to your resources directory instead).

    how to make programmatic content work, including how to find keywords with traffic potential that scale to populate dozens of pages.

    Here are two extra points to keep in mind with this approach:

    • Check if ChatGPT uses web search to gather information on your topic. If it does, your chances of being cited increase.
    • Prioritize time-sensitive data. Topics that change regularly (like yearly stats or pricing) are great candidates for programmatic content, since LLMs tend to favor fresh, up-to-date information.

    For example, if you ask “cost of living in different cities,” ChatGPT will pull information from a webpage it finds through a web search, rather than relying on built-in data. It will also note that the information is for 2025 and may change the following year.

    Screenshot of a ChatGPT response providing a 'Global overview' of cost of living based on Numbeo's 2025 index, demonstrating the AI's preference for current and time-sensitive data gathered via web search.

    check out our complete guide to E-E-A-T.

    relies on real-time searches instead of memorization. If others follow, traditional SEO becomes even more important for AI citations.
  • Younger audiences use AI assistants for information search more than others. As they gain purchasing power, conversational search will become the norm (full study).
  • Agentic AI will act instead of just answering. Future AI could complete tasks via APIs and databases directly, reducing web traffic while creating new opportunities for “AI-friendly” services.

    • Got questions or comments? Find me on LinkedIn.

    Similar Posts

    Trezor Safe 7 Impressions: Quantum Proof?

    Trezor Safe 7 Impressions: Quantum Proof?

    ByFast Domain

    Pre-order your Trezor Safe 7 here: https://cyberscrilla.short.gy/Safe7 Credits for videos I used:https://x.com/AltCryptoGems/status/1980601547336691761https://x.com/fecoxa/status/1980595013886091305https://x.com/hodler_investor/status/1980568172169241011https://x.com/thegaboeth/status/1980579649236475962https://x.com/0xmars_/status/1980604443595858178