Study: ChatGPT Sends 0.62% of Google’s Organic Traffic

Study: ChatGPT Sends 0.62% of Google's Organic Traffic
Study: ChatGPT Sends 0.62% of Google's Organic Traffic
Affiliate disclosure: This post contains affiliate links. If you buy through them, I may earn a commission at no extra cost to you. Full disclosure

ChatGPT sent 0.62% as many sessions as Google organic search across 100 websites over 12 months, according to an analysis SEO consultant Marie Haynes published October 21.

For a high-ticket store owner reading Ecommerce Paradise, this decides where the next hour and the next dollar go: Google Shopping feed work or AI visibility work. In her sample, Google organic produced 964 conversions to 36 for all AI chatbots combined. ChatGPT visitors still converted at 13.8% against 3.02% for Google organic.

Below are the numbers with their baselines, the Shopify and eMarketer data that points the other way, a rough budget model, and five steps to run this week. I covered the ChatGPT tracking bug that hid AI referrals for a week yesterday, and it changes how far you can trust any of these counts.

Whatever channel mix you land on, put the store inside an LLC you formed the right way. Northwest Registered Agent is the best LLC formation service I would point an ecommerce owner to, and unlike cheap formation sites that upsell, it forwards physical mail to operators who run their stores from outside the formation state or country. Form your LLC with Northwest →

ChatGPT Drove 0.62% of Google’s Organic Sessions in 100-Site Study

Haynes analyzed Google Analytics 4 data from 100 websites of varying sizes, covering 12 months and more than 145 million sessions, Search Engine Land reported on October 21. ChatGPT sessions equaled 0.62% of Google organic volume. Adding every other large language model raised the figure to 0.73%. Neither number is a share of all site traffic.

Fifty-five of the 100 sites tracked conversions. On those, ChatGPT visitors converted at 13.8% and Google organic visitors at 3.02%, which Haynes describes as more than 4.5 times the rate. Google organic still drove 964 conversions to 36 from all LLMs combined, or 27 times as many, per her write-up.

Haynes says exponential growth in ChatGPT referrals ended in August 2025, around the GPT-5 release. In her words, “ChatGPT has been sending fewer and fewer visitors to websites since then.” Her conclusion, as quoted by Search Engine Land: “The AI search that is actually sending people to businesses today isn’t ChatGPT. It’s Google.” She calls generative engine optimization, or GEO, “a distraction.”

The limits are visible in the write-up. Haynes calls it a small dataset. It gives no industry mix, no exact reporting dates and no definition of a conversion across sites, and 45 of the 100 sites tracked no key events at all. Per Search Engine Land’s summary, the analysis did not measure changes in Google organic traffic, and it has no ecommerce breakdown or revenue per session.

A second Search Engine Land piece the same day tested whether AI hides inside Direct traffic. Growth advisor Kevin Indig checked 12 mostly B2B sites across August and September and found that Direct sessions with identifiable AI traces averaged 0.003% of Direct traffic. Much of the Direct growth looked like bots. He could not detect an AI effect, though he says one may exist.

Shopify Says AI Orders Tripled While Haynes Sees Decline

The study lands against loud claims in the other direction. On Shopify’s Q2 call, reported by Retail TouchPoints on August 5, the company said AI-driven traffic and orders tripled year over year. President Harley Finkelstein added that new buyer orders from AI channels arrive at nearly twice the rate of other channels. Revenue rose 34% to $3.6 billion and GMV rose 32% to $116 billion. The report gives no absolute baseline for the tripling.

eMarketer’s June 22 report, using Grips Intelligence data, counted 20.5 million US ecommerce AI referral visits from ChatGPT in Q1 2026, up 210.5% year over year. ChatGPT held 92.2% of AI referral traffic, Gemini 5% and Perplexity 1.9%. Visibility Labs data cited in the same piece put ChatGPT ecommerce traffic at a conversion rate 31% above non-branded organic search in 2025.

The same eMarketer report shows how misleading percentages can be at the edges. Copilot referral visits rose 3,918% to 44,720, and Claude visits rose 3,798.6% to 294,267, but eMarketer notes the report gives no prior-period base for either, so the growth rates are hard to read. Large jumps from small bases look bigger than the absolute volumes. Generative AI engines also account for 3.3% of US online discovery time per Comscore, which measures attention, not referral traffic or sales.

These sources disagree, and they do not measure the same thing. Haynes compares ChatGPT sessions to Google organic across mixed sites and sees a decline since August 2025. Grips counts US ecommerce referral visits against Q1 2025. Shopify reports a tripling against Q2 2025 with no base figure. The conversion premium runs from 31% in the Visibility Labs ecommerce data to more than 4.5 times in Haynes’ mixed sample. eMarketer also flags that the 31% figure uses 2025 data from a different source than the Grips traffic numbers.

Haynes points store owners toward agentic checkout instead, naming Google’s Universal Cart, the Universal Commerce Protocol and WebMCP as the opportunity, though she offers no measured results for them. I covered Google Universal Cart going live for Shopify stores. I also covered the audit where no Shopify store in 122 passed all nine AI-agent checks.

What 0.62% Means for Your High-Ticket Ad and SEO Budget

My read: this does not say AI search is a dud. It says that for the sites in this sample, Google is where the volume is, and ChatGPT is a small channel with unusually motivated visitors. Both facts can be true, and a year-over-year jump can sit next to a decline from a late-2025 peak.

Here is hypothetical math, not data from any store. Say your store gets 10,000 Google organic sessions a month. At the study’s 0.73% all-LLM ratio, AI chatbots send about 73 sessions. At 3.02%, your Google sessions produce about 302 conversions. At 13.8%, the 73 AI sessions produce about 10. The visitors are better, and Google still delivers 30 times the volume.

Conversion in Haynes’ data means whatever each site set as a key event. For a high-ticket store that might be a quote request, a call or a completed order, and those can look very different from each other. Run the same math with your own numbers before you move a dollar.

High-ticket buyers add a wrinkle. Someone spending several thousand dollars on a sauna, a log splitter or a mobility scooter does not decide in one session. They research for days, compare models, and often ask an assistant which one to trust. If ChatGPT shapes that research and the buyer later types your brand into Google or your URL into the browser, GA4 credits Google or Direct and the assistant gets nothing. The 0.62% likely understates how often AI touches a sale, though no one in this data can say by how much.

That is why I would not read the Haynes ratio as permission to ignore AI, and why I would not read the Shopify tripling as a reason to rebuild your strategy around it. Both are reported by parties with something to explain. Haynes argues against GEO as a discipline, and Shopify sells the platform that benefits from agentic shopping. Your own data is the only tiebreaker that has no angle.

My thresholds, not industry standards: if AI referrals in your GA4 are under 1% of your Google organic sessions, treat AI visibility as maintenance and keep the budget on Google. Between 1% and 3% with a conversion rate at or above Google’s, give it its own monthly line item. Above 3%, run it as a channel with its own targets.

The 96% of conversions that sit with Google come from work you already know: a clean Merchant Center feed, product pages that answer a $4,000 buyer’s questions, and campaigns that convert. I walked through that in how to turn Google Shopping clicks into sales. Google is also folding AI into the same surfaces, since AI Mode ads run through your existing campaigns.

If you are tempted to buy ChatGPT traffic, I covered measured estimates putting ChatGPT ad clicks at six times Google’s cost. High-ticket stores live on margin, so before shifting spend read my piece on how to scale a high-ticket store without cutting prices.

Tracking is the weak link. ChatGPT’s referrer problems, Safari 27 blocking email-to-ID ad trackers, and Indig’s finding that AI traces in Direct are nearly undetectable all point the same way: your GA4 undercounts the thing you are trying to measure. The fix is not a better pixel. It is asking buyers where they found you.

If your store runs on Shopify, its analytics reports break sessions down by referrer, which gives you the raw material for the ratio. Doing the feed work, the tracking and the campaign tuning on top of running a store is a lot of surface area. That is the job my team takes on in the turnkey done-for-you service, where we build and run the high-ticket store for you.

New to the model? Start with what high-ticket dropshipping is, then come back to the numbers once you have a store that gets traffic.

Run your AI-versus-Google ratio past other store owners and me inside the community. Join the Skool community →

5 GA4 and Feed Steps to Run on Your Store This Week

Run these five steps this week. If you want a second set of eyes on your GA4 setup, book a call through my discovery page.

  1. Export 12 months of GA4 sessions by source. Filter for referrers such as chatgpt.com, perplexity.ai, gemini.google.com and claude.ai, then compare the total against Google organic. Paste the export into Claude and ask for a month-by-month ratio. ChatGPT does the same job.
  2. Define what a conversion is. Haynes found 45 of 100 sites tracked no key events. Mark orders, quote requests, form fills and phone clicks as key events in GA4, and log every lead’s source in HubSpot. My setup guide is how to set up HubSpot for a high-ticket store.
  3. Ask every buyer how they found you. Add a required “How did you hear about us?” field to quote forms, with ChatGPT, Google, YouTube and a friend as options. Send a one-question post-purchase email through Klaviyo to catch the rest.
  4. Fix the Google side first. Audit your Merchant Center feed titles, GTINs and images, then check your conversion tracking against why high-ticket Google Ads need a complete conversion system. Watch for the Performance Max asset group testing toggle as Google rolls it out.
  5. Track your top 20 product keywords in SEMRush so you can see Google rankings move, then set a review date after Black Friday traffic gives you a bigger sample. Stress-test the store first with my high-ticket Black Friday checklist.

Frequently Asked Questions

Should I stop working on AI search visibility?
No. Haynes calls GEO a distraction, but Shopify reports AI-driven orders tripled and eMarketer counted 20.5 million US ecommerce visits from ChatGPT in Q1 2026. Keep it small, measured and behind the Google work. Bing’s new report shows which AI queries cite your rivals instead of you.

Why did ChatGPT visitors convert better than Google visitors?
Haynes does not say. My guess, and it is only a guess, is that ChatGPT visitors arrive after a conversation that already narrowed their choice, and a smaller audience skews toward people ready to act. Treat it as a hypothesis to test in your own data.

Can I trust a 100-site sample for my store?
Only partly. Fifty-five sites tracked conversions, the industry mix is unknown, and Haynes herself calls it small. Run step one on your own GA4 before you act on anyone’s ratio.

Where does ChatGPT traffic show up in GA4 when the referrer is missing?
Usually in Direct. Indig found AI traces averaged 0.003% of Direct sessions in his sample, so GA4 alone will not settle it. A post-purchase survey in Omnisend catches what the analytics miss.

Should I test ChatGPT Ads?
Wait for your own AI referral number first. ChatGPT Ads adds creative enhancements on October 30, and I broke down how they rewrite your copy and URLs and how to opt out.

How do I handle AI-referred leads on a high-ticket sale?
A $4,000 buyer needs a call, not just a cart. My guide to high-ticket sales follow-up covers the sequence.

I am still picking a niche. Does this change anything?
Not yet. Pick the niche first with my free niches list, then set up tracking before your first ad dollar.

GA4 setup, feed work and Google campaigns are exactly what my team handles when we build and run your high-ticket store. See the turnkey done-for-you service →

Run step one before you read anything else on AI traffic this week, because your own ratio beats anyone’s sample, mine included. Subscribe to the YouTube channel for daily breakdowns. More breaking news coming through the day.

Related Articles

If this was useful, these go deeper: