ChatGPT sends under 1% of Google's traffic, and three new studies now agree on it

Three independent studies published within the same week now agree on the same uncomfortable number: ChatGPT sends software companies a sliver of the traffic Google's organic search does, and the gap in actual conversions is even wider than the gap in visits.

The headline figure comes from SEO consultant Marie Haynes, who pulled 12 months of Google Analytics 4 data from 100 websites — more than 145 million sessions — and compared referral traffic from ChatGPT against Google organic. ChatGPT sessions came to just 0.62% of the volume Google organic drove. Folding in referrals from every other LLM, including Gemini, Perplexity and Copilot, only pushed that figure to 0.73%, MarTech reported on October 22. Neither MarTech's writeup nor the original analysis specifies the exact reporting window, the industry mix of the 100 sites, or how each site defined a conversion, so the figures describe current scale rather than a precise, controlled experiment.

Conversions tell an even starker story

Among the 55 sites in Haynes' sample that tracked conversions, ChatGPT visitors converted at 13.8%, more than four times the 3.02% rate for Google organic visitors — the kind of number GEO vendors like to lead with. But volume swamps the rate: Google organic still produced roughly 27 times as many total conversions as every LLM referral source combined, simply because there are so few LLM sessions to convert in the first place. Haynes also found that ChatGPT referral growth across her sample stopped compounding around August 2025 and has been declining since.

"sessions from ChatGPT amounted to just 0.62% of the volume driven by Google organic." — Marie Haynes, via Search Engine Roundtable

She isn't the only one finding this

What makes the number hard to dismiss as one analyst's sample is that two other researchers landed on the same conclusion the same week, as Search Engine Roundtable rounded up on October 22. SEO Works measured how much traffic AI assistants send across its own client portfolio and found organic search still delivers roughly 39 times more sessions. Growth consultant Kevin Indig, checking separately across a sample of 12 sites for whether AI tools are quietly driving unattributed "direct" traffic to long-tail pages, reported he could not find a reliable relationship between AI visibility and branded or direct traffic. None of the three works for the same company, and none sells a GEO measurement product, which is part of why the agreement matters: there is no shared incentive pulling the number in one direction, and three different methodologies — referral analysis, client-portfolio comparison, and a direct-traffic correlation check — converged on the same order of magnitude.

Who this hits hardest

The teams most exposed are the ones that added a GEO or AI-visibility line item to next year's plan on the assumption that it was already a meaningful acquisition channel, not a future one. Content teams that shifted writers from keyword-targeted blog posts to "AI answer engine" formats, agencies pitching GEO audits as a standalone retainer, and marketing leaders who cited AI Overviews or ChatGPT citations in a board deck as evidence of channel diversification are the ones this data complicates most directly. Teams running GEO as a small, instrumented pilot alongside — not instead of — their existing SEO and paid search programs aren't really affected; the data doesn't say AI search traffic is worthless, only that it is still small relative to the channel it's often pitched as replacing.

What this means for a software marketing budget

This doesn't settle whether AI search visibility will matter more later. Haynes' own data shows LLM referral traffic converts better per visit, and a meaningful share of ChatGPT-driven traffic likely lands in GA4 as unlabeled "direct" rather than a tracked referral, which would understate these numbers somewhat. But at today's scale, a software marketing team that has shifted budget, headcount or quarterly OKRs toward AI visibility ahead of core organic and paid search is optimizing for a channel that, by three separate counts, still delivers a small fraction of what Google alone produces.

• Audit what share of marketing budget and content ops time went to GEO or AEO work this quarter against what share of trackable traffic and pipeline it actually produced
• Keep technical and content SEO funded at current levels before diverting spend to citation-tracking tools or "AI-first" content formats
• Check how much of your own "direct" traffic in GA4 might actually be unlabeled LLM referrals, since that changes how real this gap looks for your site specifically
• Revisit GEO vendor pitches that lead with conversion-rate comparisons rather than volume — a higher rate on a tiny base is not yet a growth channel