Three studies find ChatGPT sends under 1 percent of Google's search traffic

A marketing team spending on generative engine optimisation this quarter has three new studies to put against its budget, and all three say the AI referral traffic those budgets chase is a tiny share of what Google still sends. Search Engine Roundtable's Barry Schwartz pulled them together on Thursday morning. The conclusion the three independent analyses converge on is the one that matters for a 2027 plan: AI visibility work is hard to justify from the numbers today.

The three studies are Marie Haynes' "GEO is a distraction," published 21 October 2026; The SEO Works' AI Referral Traffic 2026 benchmark, covering September 2025 to August 2026; and Kevin Indig's "Does AI drive more direct traffic?", published 19 October 2026 in Growth Memo. Each uses a different sample and asks a different question, and they still land in the same place.

The traffic gap

Haynes analysed 100 websites and 145 million sessions in GA4 over 12 months. ChatGPT accounted for 0.62 percent of the volume Google organic sent, and all large language models together reached 0.73 percent. Google organic drove 96.5 percent of conversions across every channel, and generated 27 times more conversions than every language model combined. ChatGPT visitors did convert better once they arrived, 13.8 percent against 3.02 percent for Google organic, but the volume gap swallows that gain.

The SEO Works studied 95 websites, with 63 of those in the conversion analysis. Organic search delivered 39 times more traffic than AI assistants, and AI referrals produced 3.2 conversions for every 100 from organic search. AI converted better on average, at 7.18 percent against 6.08 percent for organic, with B2C businesses benefiting most (1.51 times better) and B2B product companies worse off (0.62 times, so AI underperformed for that segment). The firm's conclusion: "a 5 percent improvement in organic search performance would deliver more additional conversions than doubling AI referral traffic."

The hidden-traffic question

The second argument for AI visibility work is that it drives untracked direct traffic, as readers see a brand in an AI answer and type the name into the address bar later. Indig tested that on 12 sites, mostly B2B, running 632 correlation comparisons between AI visibility metrics and both branded search and direct traffic. He excluded bot visits and common direct-visit pages. His result: direct traffic surged 140.9 percent from January to September 2026, but "engaged Direct changed negative 1.0 percent in monthly counts" once he filtered the automated share. "An effect could still exist. I just couldn't isolate it," Indig writes.

What a marketing budget does with this

Haynes' line is the sharpest. "While it's interesting to measure and try to influence whether your content is cited by ChatGPT, for the vast majority of sites this traffic is miniscule compared to Google organic visitors," she writes. Her sample is heavily weighted to small and mid-size sites, so an enterprise with a different traffic mix may still get different figures, and SEO Works noted that AI referral traffic rose sharply in May 2026 when ChatGPT began making brand names clickable inline inside answers. A team planning for 2027 can read the three studies the same way: measure the AI referral and conversion numbers from your own GA4 before you set a GEO budget, rather than inheriting a figure from last year's assumption that this traffic was about to arrive.

Source

Search Engine Roundtable: 3 New Studies Suggest AI Visibility & Search Not Worth Investing In, by Barry Schwartz. Underlying studies: Marie Haynes: GEO is a distraction, The SEO Works: AI Referral Traffic 2026 Benchmark, and Kevin Indig: Does AI drive more direct traffic? in Growth Memo.