Skip to content
Get a Proposal
All articles

Analytics & CRO

Gemini Dropped Its UTM Tags. Is Its Traffic Now in Direct?

2 November 2026

6 min read

person using macbook pro on table
Photo by Myriam Jessier on Unsplash

Google added UTM tags to Gemini links in September, then removed them by Sept 28. See what it means for your GA4 reports and how to track Gemini visits now.

Did your Gemini traffic look better in September, and then quietly slide back into the "Direct" pile?

If you've been watching AI referrals in GA4, that's not your imagination. Google started adding UTM parameters to the links Gemini cites in early September 2026, and by around September 28 they were gone again.

Nobody announced either change. Marketers found out by spotting it in their data, and Search Engine Roundtable's Barry Schwartz said the tags "vanished overnight" after growing for about three weeks.

This guide explains what happened, why it matters for the way you report on AI search, and what to set up this week so Gemini visits stop hiding.

Key Takeaways

  • Google began adding utm_source=gemini to cited links in early September 2026 and stopped around September 28, according to Search Engine Roundtable.
  • Desktop Gemini still passes a gemini.google.com referrer, and that referral count stayed flat through the whole UTM change.
  • The Gemini iOS app opens links in its own WebView without a referrer, so GA4 records those visits as Direct.
  • One Reddit user put referrer pass-through from the Gemini mobile app at around 9 percent, but that figure is anecdotal and unverified.
  • Google hasn't documented any of this, so treat Gemini numbers in GA4 as a floor, not a total.

On October 1, 2026, Search Engine Journal reported that Google had quietly added UTM parameters to outgoing Gemini links. A Reddit user noticed about 24 hours after it appeared, and Google's John Mueller said he'd seen it too.

The idea was simple. A UTM tag tells your analytics exactly where a visit came from, no guessing needed.

Then, per Search Engine Roundtable, the tags stopped around September 28. That report comes from a chart shared by an analyst called Natzir, who tracked the tags appear and then disappear.

Mueller's own reaction was measured. He said it "would be great to have these retained (in addition to the UTM-tagging)", meaning the referrer data as well, and asked for reproducible examples he could pass to Google's team.

Person analysing a website traffic report on a laptop
AI referrals are small today, but they're the visits most likely to be mislabelled. Photo by Shoper on Unsplash

Why Gemini Visits Land in Direct

GA4 files a visit under Direct when it has no referrer and no UTM data to tie it to a source. As SEJ put it, Direct is "the fallback used when there is no referrer or UTM data to explicitly tie a visit to a referrer."

Here's where Gemini gets tricky.

What the reports suggest:

  • Desktop Gemini still sends a gemini.google.com referrer, so those visits show up as Gemini referrals.
  • The iOS app opens links in its own WebView with no referrer. Natzir found these visits show as Direct, with only a "GeminiiOS" user agent in server logs as evidence.
  • Android and the mobile assistant were described by a Reddit user as often failing to pass referrers, though that is one person's observation.

So the UTM tags were quietly doing a lot of work. While they were on, they could rescue app visits that a referrer couldn't. Now that they're off, those visits are back to being anonymous.

What This Means for Your Reporting

If you report on AI search to a boss or a client, your Gemini line is probably too low. And your Direct line is probably too high.

That matters for decisions. A rise in Direct can look like brand strength when part of it is people clicking a cited link inside an AI app. We covered the same blind spot for ChatGPT in our post on AI traffic hiding in your GA4 Direct channel, and Gemini adds a second source to the pile.

The honest way to report it is as a range. Show the referrals you can measure, then flag that app traffic is under-counted.

Gemini Tracking in Action: A Quick Example

Take a service business that sees 40 visits a month from gemini.google.com in GA4. That's the measured floor.

If Direct jumped in September and fell back in October, with no campaign or email send to explain it, some of that swing may be the UTM window opening and closing. You can't prove it from GA4 alone, but you can build the case from other signals.

A simple check takes three steps:

  1. Compare weekly Direct sessions from August to October and note the weeks that moved without a campaign.
  2. Filter server logs for the "GeminiiOS" user agent on those weeks.
  3. Look at the landing pages: if the visits hit pages Gemini would cite (guides, FAQs, pricing), that's a hint they were AI-driven.

It's not perfect. But it turns a mystery into a pattern you can watch.

Developer reviewing server logs on a monitor
Server logs hold the user agent that GA4 never sees. Photo by Compagnons on Unsplash

How to Track Gemini Traffic This Week

1. Set up a Gemini segment now

Create a GA4 view or custom channel group that catches any session with a gemini.google.com source or a utm_source of gemini. Keep both rules, because the UTM may return without notice.

2. Check your server logs

Logs hold what GA4 can't see. Search for the "GeminiiOS" user agent and count the hits per page. Ask your developer or host for a weekly export if you don't have access.

3. Annotate the dates

Mark early September (tags on) and around September 28 (tags off) in your reporting notes. Whoever reads the report in six months will thank you.

4. Make your cited pages worth the click

Tracking is only half the job. The pages Gemini cites should lead somewhere useful: a clear next step, a form that works on a phone, and a fast load. Our GEO optimisation and web analytics setup work covers both sides.

Where This Breaks Down

Don't over-read any of this. Google hasn't documented when Gemini adds UTMs, and SEJ notes it may only happen when an answer directly references a website. The "9 percent" referrer figure came from a single Reddit user, not a study.

And AI referrals are still small for most sites. If Gemini sends you a handful of visits a month, a perfect attribution setup won't change your strategy. A quick segment and a log check is enough; a full rebuild isn't.

It's also possible the tags come back. Build your rules so they work either way.

Person tapping a link on a smartphone at a desk
App visits are the ones most likely to arrive without a referrer. Photo by Christian Wiediger on Unsplash

FAQs

Does Gemini add UTM parameters to links now?

Not at the moment. Google added utm_source=gemini to cited links in early September 2026 and stopped around September 28, per Search Engine Roundtable. Google hasn't documented the behaviour, so it could return. Set up tracking that works with or without the tag.

Why does Gemini traffic show as Direct in GA4?

GA4 uses Direct when a visit has no referrer or UTM data. The Gemini iOS app opens links in a WebView without a referrer, so those visits arrive unlabelled. Desktop Gemini still sends a gemini.google.com referrer, so it is tracked as a referral.

How can I find Gemini visits that GA4 missed?

Check your server logs for the "GeminiiOS" user agent, which Natzir found on iOS app visits. Then compare those hits with unexplained jumps in Direct sessions and with the landing pages involved. It gives you an estimate, not an exact count.

Is Gemini traffic big enough to worry about?

For most sites it's still small, so don't rebuild your analytics for it. But it's growing as a research channel, and mislabelled visits skew your view of what works. A simple segment plus a monthly log check is a sensible minimum.

Conclusion

The takeaway: Gemini traffic in GA4 is a floor. Some of your Direct traffic is likely AI-driven, and Google can change what it sends without telling anyone.

This week, build a Gemini segment, pull one server-log export and annotate September in your reports. That's about an hour of work and it gives you honest numbers.

If you'd like help making your AI-search reporting trustworthy, see our channel attribution service or get a free quote and we'll take a look.

Sources

David Banjo

About the author

Founder & CEO, Binatec Agency

David Banjo is an entrepreneur and web developer, and the Founder & CEO of Binatec Agency, where he helps businesses scale through custom software, digital products, and e-commerce growth strategies. He is also the creator of ProposelyAI, an AI-powered platform designed to streamline and automate proposal generation for freelancers and agencies. Recognized for his work in web development, AI workflow automation, and growth marketing, David equips brands with the modern tools needed to optimize performance and expand their digital reach.

Discussion

unknown user

John Doe

Supreb Information. Thanks

15 July 2014 View replies . 1

replylike

Kazzim Abdulmojeed

Very Helpful Information. But i need more clarity about lorem

15 July 2013 View replies . 2

replylike

Leave a comment

See your site the way Google does

Scores, missed opportunities and the quick wins worth fixing first. Ready in about a minute.

  • Free
  • No sign-up
  • Instant results
See all free tools