AI Visibility · 9 July 2026

ChatGPT vs. Google: Which Drives More Organic Traffic?

Ask which channel drives more organic traffic, ChatGPT or Google, and you will get a confident answer from your analytics: Google, by a mile. Google Search still sends more visits in a single day than ChatGPT sends most sites in a year.

The problem is that the number you are reading is wrong, and not by a small margin: it is wrong in a way that systematically flatters Google and hides ChatGPT, because a large share of the traffic ChatGPT actually sends you never gets labelled "ChatGPT" in the first place. It lands in your reports as Direct, or as a branded Google search, and the tool quietly credits the wrong channel.

So the useful questions are which numbers you can trust, and how to measure the two fairly. This guide does both: it reframes the comparison around the measurement gap, gives you a step-by-step way to see ChatGPT versus Google traffic properly in GA4, and puts real, sourced figures against the scale of each.

Which actually drives more organic traffic, ChatGPT or Google?

On raw, measured volume, Google wins comfortably and will for a while. In June 2025, Similarweb estimated that all AI platforms combined sent 1.13 billion referral visits to the top 1,000 websites worldwide, while Google Search sent 191 billion in the same month.^[1] That is roughly a 170-to-1 gap. If the contest is total clicks, it is not close.

Two things sit underneath that headline. The AI number is growing fast: those AI referrals were up 357% year on year, and ChatGPT alone accounted for more than 80% of them.^[2] The measured ChatGPT figure is also an undercount, which matters more for your own reports. The 1.13 billion is what tracking tools could attribute. A meaningful slice of ChatGPT's real influence never carries a ChatGPT label at all, and that is where most owners get the comparison wrong.

The short version

• Google drives far more measured traffic: 191 billion referrals in June 2025 against 1.13 billion from all AI platforms combined.^[1]
• ChatGPT's real contribution is undercounted. Many ChatGPT visits arrive with no referrer, so analytics files them under Direct rather than as a ChatGPT referral. The comparison you see is skewed before you start.
• The growth curve is steep: AI referrals rose 357% year on year, ChatGPT making up over 80% of them, and referrals to news sites alone jumped 770%.^[2]
• Google's own baseline is shrinking. Ahrefs analyzed 300,000 keywords and found the presence of an AI Overview correlated with a 34.5% lower click-through rate for the top-ranking page.^[3] You are comparing ChatGPT against a Google number that is eroding from the inside.
• Per visit, ChatGPT traffic holds up well. In one study of 304 sites, ChatGPT traffic engaged at 63.42%, slightly ahead of organic search at 61.64%.^[4]
• You can fix the measurement with a custom GA4 channel group and a few cross-checks, covered step by step below.

Why does ChatGPT traffic show up as Direct, not as a referral?

Because in most cases ChatGPT hands the browser a link with no referrer attached, and analytics has a firm rule: a visit with no referrer and no campaign tag goes into the Direct bucket. That single rule is responsible for most of the missing ChatGPT traffic.

When someone clicks a source link inside a ChatGPT answer, the click often opens from an in-app browser or under a referrer policy that strips the originating URL. Your site sees a visitor appear with no referring page. GA4 has nothing to attribute the visit to, so it records it as Direct, the same bucket as someone typing your address by hand. The visit was earned by ChatGPT; the label says otherwise.

A lot of ChatGPT's influence does not produce an immediate click at all. Someone asks ChatGPT "who does emergency boiler repairs in Sheffield", sees your name in the answer, and then does what people do: they open a new tab and Google your business by name, or they type your address in directly a day later. That visit lands in your reports as branded organic search or as Direct. The AI answer started the journey, but Google or Direct takes the credit. No analytics fix fully recovers this part of the gap, and it is why measured ChatGPT traffic will always understate the real figure.

So when you look at a report that says ChatGPT sent 40 visits and Direct sent 900, some unknown share of that 900 is misfiled ChatGPT. The two channels are not being measured on the same terms.

What gets recorded where

A quick map of how the same underlying journey lands in different buckets:

Three of these five journeys were driven by ChatGPT, and only one of them will show up in your reports with ChatGPT's name on it.

How do you measure ChatGPT vs Google traffic in GA4?

You will never recover every misfiled visit, but you can get from a wild undercount to a defensible estimate with four moves, in rough order of effort.

1. Build a custom AI channel group

Out of the box, GA4 does not treat chatgpt.com as its own channel, so AI referrals with an intact referrer get lumped into a generic Referral line. Create a custom channel group (Admin, then Channel groups, then create new) and add a channel that matches AI sources by referrer or source. A source regex along the lines of chatgpt|openai|perplexity|gemini|copilot|claude pulls the visits that did keep their referrer into a single "AI Assistants" line you can compare against Organic Search directly. This alone surfaces the ChatGPT traffic that was hiding inside Referral.

2. Use deep landing pages as a Direct proxy

The misfiled ChatGPT visits sitting in Direct can be spotted by their landing page. Real direct traffic (someone typing your URL) almost always lands on your homepage. A visitor landing straight on a deep internal page, a specific service page or a blog post, with Direct as the source, rarely typed that URL from memory. Filter Direct traffic by landing page: a spike of Direct sessions hitting a niche page that answers a specific question is very likely ChatGPT (or another assistant) that lost its referrer. It is an estimate rather than proof, but it puts a floor under the real number.

3. Watch branded search in Search Console

Because so much AI influence resurfaces as a branded Google search, your own brand name is a measurement instrument. In Google Search Console, track impressions and clicks for queries containing your business name. A steady climb in branded search, especially when it tracks with your visibility in AI answers, is a signal that ChatGPT and its peers are sending people who then look you up. It will not appear in any "ChatGPT" report, but it is real demand the AI created.

4. Ask people directly

The lowest-tech check is the most reliable one for the branded-search leak. Add "How did you hear about us?" to your contact or booking form, with ChatGPT or AI assistant as an explicit option. Self-reported attribution is messy, but it catches exactly the journeys analytics cannot: the customer who found you in an AI answer, sat on it for a week, then called. When those replies start naming ChatGPT, you know the channel is working regardless of what GA4's Direct bucket says.

How big is ChatGPT's referral traffic?

Big enough to stop ignoring, though Google is in no danger yet. The measured figure, 1.13 billion AI referrals across the top 1,000 sites in June 2025, is a fraction of Google's 191 billion, but it is a fraction that multiplied more than fourfold in a year.^[1][2]

The distribution matters as much as the total. ChatGPT is not spreading its traffic evenly; it concentrates on the sites it trusts to answer a given question. Similarweb found referrals to news and media sites specifically were up 770% year on year, far faster than the overall rate.^[2] The engine is getting more willing to send people off-platform to a source that clearly answers the question, which is the behaviour a well-built business page can benefit from.

Put the growth rate next to Google's own trajectory (next section) and the trend lines are moving in opposite directions: the AI number is climbing steeply from a low base, while the value of a Google ranking is being chipped away from the top.

Is Google's organic traffic shrinking?

Yes, in the way that matters: the clicks a top Google ranking earns are being absorbed by Google's own AI answers. So when you compare ChatGPT against Google, you are comparing it against a baseline that is steadily eroding.

The number that pins it down comes from Ahrefs. Analyzing 300,000 keywords, it found that when an AI Overview appeared in the results, the top-ranking page saw a 34.5% lower click-through rate than comparable informational keywords without one. The average position-one CTR on those keywords fell from 0.073 in March 2024, before the US rollout, to 0.026 a year after.^[3] The ranking is intact; the click it used to earn is being absorbed by the answer box above it.

That reframes the whole comparison. The pile of "Google traffic" is not fixed: Google is shrinking its own referral output from the inside, by answering more questions on the results page. Your No.1 position is intact; the clicks it used to guarantee are leaking into a summary box. For the fuller picture of how a top ranking and an AI recommendation diverge, our companion piece on why ranking No.1 on Google won't get you into ChatGPT digs into the source overlap data.

Does ChatGPT traffic convert better than Google traffic?

Per visit, it tends to hold its own and often edges ahead, because the people who arrive have already been pre-qualified by the AI. Siege Media measured engagement across 304 client properties from July 2023 to July 2025 and found ChatGPT traffic engaged at 63.42%, slightly ahead of organic search at 61.64%.^[4]

Someone who clicks through from a ChatGPT answer has usually already read a summary, seen you named against alternatives, and clicked specifically to get the detail. They are past the browsing stage and into evaluation. That is a smaller stream than Google sends, but it arrives warmer, which is why "ChatGPT drives less traffic" is true but "ChatGPT drives worse traffic" is not.

For a business: do not judge ChatGPT purely on session count against Google. A channel that sends a tenth of the volume but at higher intent, and is growing 357% a year, deserves a different scorecard than a raw-clicks comparison gives it.

What does this mean for a local service business?

It means the channel most likely to be sending you your best-fit customers is also the one your analytics is worst at seeing, so you may be underinvesting in it without knowing.

Picture a two-van plumbing firm in Leeds. In GA4 it sees 1,100 Direct sessions a month and shrugs them off as "people who know us". But a chunk of those are people who asked ChatGPT for "a reliable emergency plumber in Leeds", got the firm named, and came straight to the site with the referrer stripped. The owner has no idea the AI is working, because every one of those visits is wearing a Direct label. Meanwhile the firm keeps pouring budget into Google Ads for the same search, competing for a click that an AI answer is increasingly giving away for free.

That blind spot has a flipside. The AI recommendation the firm cannot see is not gated behind a big advertising budget or a famous name; it turns on whether an assistant finds the business easy to quote, how many real reviews customers have left it, and whether its service area and contact details agree across the web. Those are levers a small local operator can actually pull. Most have not, and not because it is hard: their own reporting never surfaced the channel in the first place. The demand is already landing, mislabelled, from customers the firm is crediting to nobody, while the few adjustments that would grow it sit untouched.

So which one should you focus on?

Both, but measure them accurately before you decide the split. Google is still where the volume is, and a readable, well-structured site remains the foundation that AI engines read from too. ChatGPT is the smaller, faster-growing, higher-intent channel that your reports are actively hiding from you. Writing one off because of a raw-traffic comparison, when that comparison is built on a broken measurement, is the mistake this whole piece is trying to prevent.

The first move is to fix your instruments: set up the AI channel group, sanity-check your Direct traffic, and start asking customers how they found you. Once you can see the real numbers, the investment decision is a straightforward one.

Where do you start?

Start with two numbers your dashboard cannot give you: how often an AI engine names your business, and how much of the traffic it sends is already sitting mislabelled in your reports. The measurement fixes earlier in this piece cover the second number. For the first, run the four checks from the GA4 section for a month and pair them with a quick look at what the engines actually say when someone asks for your service.

Seeing the traffic is only part of the job. Even once you can measure ChatGPT properly, you still need to know why it does or does not name you, and that comes down to signals a traffic report never shows: whether AI crawlers can read your site, whether your content is specific enough to quote, whether your off-site presence and business details line up across the web.

A free QBiz Leads AI visibility check closes that second gap. It scans your website in about thirty seconds and returns a clear pass or fail on the signals that decide whether AI tools can find and recommend you, so you get a prioritised list of what to fix instead of a vague sense that the AIs have not noticed you.

Get your AI Visibility audit →

Frequently asked questions

Does ChatGPT or Google drive more organic traffic?

Google, by a wide margin on measured volume: 191 billion referrals to the top 1,000 sites in June 2025 against 1.13 billion from all AI platforms combined. But ChatGPT's real contribution is undercounted, because many of its visits arrive with no referrer and get filed as Direct rather than as a ChatGPT referral.

Why is my ChatGPT traffic showing as Direct in Google Analytics?

Because the click usually arrives with no referrer attached, often from an in-app browser or a referrer policy that strips the source URL. GA4 sends any visit with no referrer and no campaign tag to the Direct channel, the same place it puts someone typing your address by hand, so earned ChatGPT visits get mislabelled.

How do I track ChatGPT traffic separately in GA4?

Build a custom channel group that matches AI sources by referrer (chatgpt.com, perplexity.ai, and similar) so intact referrals land in their own line. Then estimate the misfiled portion by filtering Direct traffic to deep landing pages, watch branded queries in Search Console, and add an "How did you hear about us?" option to your forms.

Is ChatGPT traffic worth more than Google traffic?

Per visit it often is, because the visitor has already seen a summary and clicked for detail. One study of 304 sites put ChatGPT engagement at 63.42% against 61.64% for organic search. It sends far less volume than Google, but it arrives at a later, higher-intent stage.

Is Google search traffic declining because of AI?

The clicks it sends are. Google now answers many searches with an AI summary at the top of the page, and Ahrefs found that across 300,000 keywords the top-ranking page saw a 34.5% lower click-through rate when an AI Overview was present. Your ranking can be unchanged while the traffic it earns quietly falls.