I hear two opposite beliefs from business owners right now. One says, “Everyone uses ChatGPT now, Google is finished.” The other says, “ChatGPT sends us almost no traffic, so it doesn’t matter.” Both sound reasonable. Both are only half right.
The truth sits in the gap between two numbers: how much searching happens inside ChatGPT, and how much traffic ChatGPT actually sends to websites. Those numbers are very far apart, and that gap should shape how you split your time and budget.
In this article, let me explain both numbers in plain words, show you why they differ so much, and then walk you through how to measure the same ratio on your own site. I will finish with how I would use the result to decide where your next marketing dollar goes.
How big is ChatGPT as a search tool, really?
Let me start with raw volume. According to a summary of OpenAI’s own research published by the United Nations University’s C3 blog in September 2025, ChatGPT was handling more than 2.5 billion messages a day from over 700 million weekly users, with data running to mid-2025.
Google is bigger. In March 2025, Google said it already sees more than 5 trillion searches a year. As Search Engine Land reported it:
“We already see more than 5 trillion searches on Google annually.”
— Google, as quoted by Search Engine Land, Google now sees more than 5 trillion searches per year
Five trillion a year works out to roughly 13.7 billion searches a day. So on raw counts, ChatGPT messages are about 18% of Google searches. But that comparison is unfair, because many ChatGPT messages are not searches at all. People use it to write emails, fix code or talk through a problem.
The same OpenAI research split messages into three kinds of intent: Asking (49%), Doing (40%) and Expressing (11%). It also found that a large share of use is personal, not work:
“Over 70% of ChatGPT messages are for non-work purposes.”
— UNU C3, What over 2.5 billion daily messages reveal about how people use ChatGPT
Three ways to count a “search” inside ChatGPT
This is where most headlines go wrong. How big ChatGPT looks depends entirely on what you call a search. Patrick Stox at Ahrefs worked through this in his February 2026 analysis, using OpenAI’s study of 1.5 million conversations. Here is a simplified version of his math.
| Definition of “search” | Share of ChatGPT use | Search-like prompts per day | As a share of Google’s daily searches |
|---|---|---|---|
| Pure information seeking only | 24% | About 600 million | 4.38% |
| All “Asking” intent | 51.6% | About 1.29 billion | 9.42% |
| Stox’s own classification (things people used to do on Google) | 65% | About 1.625 billion | 11.86% |
That last row is where the “12% of Google” headline comes from. Stox is open that it is his own judgment call. Google also handles things like calculations and translations, so he argues the narrower definitions undercount what ChatGPT replaces.
In brief: ChatGPT’s search-like volume is somewhere between 4% and 12% of Google’s. That is big. It is bigger than Bing, which Ahrefs estimates at about 1.2 billion searches a day. But Google still has roughly 5.5 times more daily activity than all of ChatGPT combined.
How much traffic ChatGPT sends compared with Google
Now the second number. Ahrefs looked at traffic across 76,000 websites in its own web analytics product. Google sent about 39.98% of all traffic to those sites. ChatGPT sent about 0.21%. That makes Google’s referral traffic roughly 190 times larger.
A separate Ahrefs study from June 2025, covering 81,947 websites, found that AI traffic as a whole made up about 0.25% of total site traffic on average, with ChatGPT sending more than 80% of that AI traffic. Different samples, similar picture.
So here is the key comparison:
- ChatGPT has up to about 12% of Google’s search-like volume.
- ChatGPT sends about 0.5% of the traffic Google sends (0.21 divided by 39.98).
Stox summarizes the difference neatly:
“The data shows two different business models: Google connects people to websites, ChatGPT keeps them in conversation.”
— Patrick Stox, Ahrefs, ChatGPT Has 12% of Google’s Search Volume but Google Sends 190x More Traffic to Websites
Why ChatGPT users click out so rarely
Ahrefs estimated click-through rates by dividing each platform’s traffic share by its daily queries. These are modeled estimates, not measured clicks, so I treat them as direction, not hard science.
| Platform and definition | Estimated click-through rate |
|---|---|
| Google (all searches) | 29.2% |
| ChatGPT, pure search prompts | 3.5% |
| ChatGPT, all Asking prompts | 1.6% |
| ChatGPT, Stox’s classification | 1.3% |
| ChatGPT, all usage | 0.84% |
Depending on the definition, ChatGPT’s click-out rate is 88% to 97% lower than Google’s. Why? Let me explain the main reasons in simple terms.
- The answer is the product. ChatGPT writes a complete reply. For many questions, there is nothing left to click for.
- Links are small and optional. Sources often sit as small citations or in a side panel. On Google, the links are the page.
- Many prompts are tasks, not searches. If someone asks ChatGPT to rewrite an email, there is no website to visit.
- Follow-up questions stay in the chat. Instead of opening a second site, people just ask another question.
- Some clicks are hidden. Copy-pasted links and app traffic can arrive with no referrer and land in “Direct.” So some ChatGPT traffic is real but invisible.
Google is not immune either. Pew Research found that people who saw an AI summary on Google clicked a traditional result in 8% of visits, versus 15% without one (March 2025 data). The difference is that Google still starts from a page of links.
Influence without clicks: where ChatGPT’s value hides
If ChatGPT sends so little traffic, why care? Because a mention in an answer can still shape a buying decision. Someone asks ChatGPT, “Which project management tool is best for a small architecture firm?” They read three names. A week later they type one of those names into Google. That visit shows up as branded organic search or direct, not as ChatGPT.
This is the part most reports miss. Here are the signals I look at for this hidden influence:
- Branded search impressions in Google Search Console, trended month by month.
- Direct traffic to deep pages, such as a pricing page, which people rarely type by hand. I explain this in my guide to what direct traffic really means.
- “How did you hear about us?” answers on forms and sales calls. Add ChatGPT and “AI assistant” as options.
- Brand mentions in AI answers, tracked with a prompt-tracking tool.
None of these prove that ChatGPT caused anything. If you want to test that properly, my guide on proving AI search impact with holdout and pre/post tests covers the method. And if you want to know whether the few ChatGPT visitors you do get are worth more per visit, read my article on testing whether AI referral traffic is higher quality.
How to work out your own ChatGPT-to-Google ratio in GA4
Industry averages are a start. Your own ratio is what matters. Here is how I would do it in about 20 minutes.
- Open GA4 and go to Reports → Acquisition → Traffic acquisition.
- Set the dimension to “Session default channel group.” Google’s documentation now lists an AI Assistant channel for traffic from tools such as ChatGPT, Gemini, Copilot and Grok. It notes that this channel excludes Google’s AI Overviews and AI Mode, which count as organic search.
- Switch to “Session source / medium” and filter for chatgpt.com to isolate ChatGPT from other assistants.
- Pick a long date range. At least 90 days, because AI traffic is small and noisy week to week.
- Calculate the traffic ratio: ChatGPT sessions ÷ Google organic sessions.
- Calculate the value ratio: ChatGPT key events ÷ Google organic key events. This matters more than sessions.
- Write both numbers down with the date. Repeat every quarter.
If your setup doesn’t show AI traffic cleanly yet, follow my step-by-step guide to tracking ChatGPT and AI traffic in GA4, and clean up referral noise using my guide to referral traffic in GA4.
Worked example: a hypothetical B2B software site
Let me show the math with made-up numbers. Say a small software company sees this over 90 days:
| Metric | Google organic | ChatGPT | Ratio (ChatGPT ÷ Google) |
|---|---|---|---|
| Sessions | 60,000 | 480 | 0.8% |
| Demo requests (key events) | 300 | 6 | 2.0% |
| Demo rate | 0.5% | 1.25% | 2.5 times higher |
In this example, ChatGPT sends less than 1% of Google’s traffic, which is a bit above the Ahrefs average of about 0.5%. But because those visitors convert better, ChatGPT produces 2% of Google’s demo requests. Add some unknown share of branded search that started in ChatGPT, and its true influence is higher still. That is the kind of picture I would expect for many B2B sites, but only your data can confirm it.
Splitting budget between Google SEO and AI visibility
So what do you do with these numbers? Here is the decision table I use as a starting point. It is not a formula. It is a way to have a sensible conversation.
| Your situation | Suggested focus | Why |
|---|---|---|
| ChatGPT under 0.5% of Google traffic, and no lift in branded search | Mostly Google SEO, with light AI visibility work | Google still drives nearly all measurable value. |
| ChatGPT 0.5–2% of Google traffic, with higher conversion rates | Google SEO first, plus a steady AI visibility program | Small traffic, but valuable visitors and likely hidden influence. |
| ChatGPT above 2%, or customers often say they found you through AI | Treat AI visibility as its own channel with its own budget | Your buyers clearly research in assistants. |
| Falling Google clicks on question keywords | Protect and adapt Google content, and invest in brand mentions | Clicks are moving into answers on both platforms. |
The good news is that much of the work overlaps. Clear pages, strong reviews and mentions on trusted sites help on Google and in ChatGPT. Our SEO services cover both for that reason. If you are worried about losing Google clicks specifically, my article on adapting when you lose traffic to AI answers gives a step-by-step response, and my guide to estimating AI Overview click loss by keyword helps you forecast it before it happens.
Industries where ChatGPT matters more than its traffic suggests
The 190x gap is an average across thousands of sites. In some markets, I would give ChatGPT more weight than its traffic share suggests. I have no single study that ranks industries here, so this is my reasoning, not hard data:
- B2B software and services, where buyers compare options and ask for shortlists.
- High-consideration purchases, such as travel, education or expensive equipment, where people research in many steps.
- Technical and developer products, because developers already use AI assistants daily.
- Health and wellness information, where people ask private questions they might not type into a public search box. Accuracy rules apply here, so be careful.
Where I would give ChatGPT less weight: urgent local services (“plumber near me”), quick purchases and anything that depends on maps, opening hours or live stock. Google Maps and the local pack still own those moments. To check your own industry, use the GA4 ratio above and compare it with the Ahrefs average. Then look at my guide to AI referral traffic benchmarks for what is normal.
How the gap between ChatGPT and Google may change
I won’t predict exact numbers, because nobody can. But here are the signals I watch to see if the gap is closing:
- Link design in ChatGPT. If citations become bigger and more clickable, click-out rates rise.
- Shopping and booking features. If more purchases finish inside the chat, clicks may fall even as influence grows.
- Google’s own AI features. As AI Overviews and AI Mode grow, Google’s click rate may move toward ChatGPT’s, not the other way round. Google, for its part, said in August 2025 that total organic click volume had been relatively stable year-over-year.
- Which index ChatGPT uses. Changes here affect which sites get cited. I cover this in my article on which search index ChatGPT uses.
My advice is simple: measure your own ratio every quarter, and let the trend in your data move your budget, not headlines.
Questions about ChatGPT vs Google traffic
Does ChatGPT really have 12% of Google’s search volume?
Only under one definition. If you count only pure information-seeking prompts, it is closer to 4.4%. The 12% figure uses Ahrefs author Patrick Stox’s broader classification of prompts that replace Google searches.
Why does ChatGPT send so little traffic?
Because it answers the question directly, shows links as small citations and handles many tasks that don’t need a website at all. Ahrefs estimated its click-out rate at 88% to 97% below Google’s.
Is ChatGPT traffic worth anything if it is so small?
Often yes. Visitors may convert at higher rates, and mentions can drive branded searches and direct visits that GA4 won’t credit to ChatGPT.
Where do AI Overviews and AI Mode show up in GA4?
Under organic search. Google’s GA4 documentation says the AI Assistant channel excludes AI Overviews and AI Mode, and Google’s Search Central guidance says that traffic is included in overall search data in Search Console.
Should I move my SEO budget to ChatGPT?
Not wholesale. For most sites, Google still sends far more traffic. Shift budget gradually as your own ChatGPT ratio and conversion data justify it.
What I would do with your ChatGPT and Google numbers this month
Here is my short plan. Pull your ChatGPT-to-Google ratio for sessions and key events. Write it down with today’s date. Add “ChatGPT or another AI assistant” to your “How did you hear about us?” field. Then decide, using the table above, whether AI visibility deserves a small program or its own budget line.
If you would like a second pair of eyes on your numbers, or help building visibility on both Google and ChatGPT, our team does exactly this. You can request a free quote here, or, if you want to see options first, request a custom quote based on your goals.
Further reading on ChatGPT vs Google traffic
- How Much Traffic Do AI Overviews Take? Estimating Click Loss by Keyword
- AI Referral Traffic Benchmarks: How Fast It's Growing and What's Normal
- AI Mode vs AI Overviews: Same Answers or Different Games?
Sources for this ChatGPT vs Google traffic analysis
- Ahrefs: ChatGPT Has 12% of Google’s Search Volume but Google Sends 190x More Traffic to Websites
- Ahrefs: AI traffic study (June 2025)
- Search Engine Land: Google now sees more than 5 trillion searches per year
- UNU C3: What over 2.5 billion daily messages reveal about how people use ChatGPT
- Google Analytics Help: Traffic-source dimensions and default channel group
- Google Search Central: AI features and your website
- Google: AI in Search is driving more queries and higher quality clicks
- Pew Research Center: Google users are less likely to click on links when an AI summary appears
