Derek Chua9 min read

Google Gemini Has Overtaken Perplexity in Referral Traffic. Here's What That Means for Your GEO Strategy.

Gemini referral traffic doubled in two months while ChatGPT declined. Here's what the data means for your GEO strategy and how to check your own analytics.

Platform comparison showing Gemini's rapid growth in AI referral traffic relative to ChatGPT and Perplexity, 2025-2026

If you've been checking your GA4 referral sources lately and noticing "gemini.google.com" appearing more often, here's your answer. Google Gemini's referral traffic to websites more than doubled between November 2025 and January 2026. In the same period, ChatGPT's traffic fell. And Perplexity, which was sending three times more referral traffic than Gemini just five months ago, has been overtaken.

Key Takeaway: Gemini referral traffic grew approximately 115% in two months (Nov 2025 to Jan 2026) and has now surpassed Perplexity globally. The strategic implication: the same E-E-A-T and structured content signals that work for Google Search are what Gemini rewards. Your existing GEO work is already Gemini-ready.

Written by Derek Chua, digital marketing consultant and founder of Magnified Technologies. Derek has been tracking AI search referral patterns across client sites since early 2025.

This isn't a "Gemini is about to overtake ChatGPT" story. ChatGPT still controls roughly 80% of all AI referral traffic, and all AI platforms combined account for just 0.24% of global web traffic. But the direction of travel matters here, and this data has one clear implication for most SME marketing teams: your GEO content work is not platform-specific. It travels.

What the Data Actually Shows

SE Ranking tracked AI referral traffic across more than 101,000 websites with Google Analytics installed. Their study (published March 26, 2026) shows Gemini's referral traffic grew 51.5% in December 2025 and a further 42% in January 2026, totalling approximately 115% over two months.

For context: during the eight months before December, Gemini's traffic was essentially flat, growing at around 4% per month. The acceleration to roughly 47% monthly growth represents a 12x change in pace. That's not organic drift. Something structural changed.

Meanwhile, ChatGPT's referral traffic peaked in October 2025, then dropped 8.5% in November and 18.6% in December. There was a small recovery in January, but the overall trend is clear. ChatGPT's lead over Gemini narrowed from 22x in October to 8x in January. Still dominant. Less so than it was three months ago.

The Perplexity milestone is the most concrete data point. In August 2025, Perplexity was sending roughly three times more referral traffic than Gemini. By January 2026, Gemini was sending 29% more visitors than Perplexity globally, and 41% more in the US. That's a complete reversal in five months.

(SE Ranking data: 101,574 sites across 250 markets, November 2025 - January 2026. Note: SE Ranking sells AI visibility tracking tools — this data comes from their own Google Analytics-based dataset. Figures reflect referral traffic only, not total AI platform usage.)

Why Gemini's Growth Is Tied to Google's Own Products

The timing is not coincidental. Google released three Gemini 3 models in rapid succession: Gemini 3 Pro on November 18, Gemini 3 Deep Think on December 4, and Gemini 3 Flash on December 17, 2025. Flash became the default model in the Gemini app and, critically, in AI Mode for Google Search.

That last part matters. AI Mode is Google's AI-native search experience, and when it sends a user to your website, that visit registers in GA4 as a Gemini referral. The more people use AI Mode in Search, the more Gemini referrals show up in your analytics. Gemini's growth isn't only about standalone chatbot adoption. It's built into how people search on Google.

This is a structural difference from ChatGPT. ChatGPT sends referrals when users click links from within the ChatGPT interface. Gemini sends referrals from both the standalone app and AI Mode, which is embedded inside the world's most-used search engine. The distribution ceiling is fundamentally different.

What This Means for Your GEO Work

Here's the part that should simplify your strategy rather than complicate it.

Gemini uses Google's own knowledge graph, indexing infrastructure, and quality signals. The E-E-A-T principles Google has been reinforcing for years (Experience, Expertise, Authoritativeness, Trustworthiness) are the same signals Gemini uses when deciding what to surface and cite.

At Magnified, we've been tracking AI citation patterns across client websites since early 2025. The pattern holds consistently: pages that get cited in Google AI Overviews and AI Mode are not some separate category of "AI-optimised" content. They're the same pages that already have strong organic rankings, clear author attribution, structured headings, and FAQ sections that answer real user questions directly.

If you've been doing the GEO fundamentals, including proper schema markup, FAQ sections, clear author bylines, and structured content with declarative headings, your existing content is already positioned to receive Gemini referrals. You don't need a separate Gemini strategy. You need to continue and accelerate the content work you're already doing.

The implication runs the other way, too. Businesses that have been deferring GEO on the assumption that "AI search traffic is still tiny" may have been right about the scale. But Gemini's integration with AI Mode means Google's AI is already embedded in the searches your customers are performing. The best time to start building structured, citable content was last year. The second-best time is now.

The Three Platforms Worth Tracking (and What Each One Rewards)

Not all AI platforms respond to the same signals. Here's a practical breakdown.

Google Gemini (and AI Mode): Rewards the same signals as Google Search. E-E-A-T, structured content, FAQ schema, clear author attribution, factual accuracy. Content that performs well in organic search is already positioned for Gemini citations. Check GA4 for referrals from gemini.google.com.

ChatGPT: Referrals show up as openai.com in GA4. ChatGPT tends to surface well-known brands, frequently cited sources, and content that appears across multiple high-authority sites. Link diversity and brand mentions across the web matter more here than technical schema alone. Still the dominant platform at roughly 80% of AI referrals, though declining from its October peak.

Perplexity: Smaller share overall, but often sends higher-intent traffic (researchers, buyers comparing options). Content that performs well here tends to be comprehensive, well-cited, and direct. FAQ sections and structured answers help. Now sending less traffic than Gemini, but still worth having on your tracking list.

The practical takeaway: don't fragment your content strategy across three separate platform playbooks. The content qualities that help you rank in Gemini (structure, expertise, citation-worthiness, factual depth) also improve your visibility in ChatGPT and Perplexity. One strong content foundation serves all three.

How to Check If Gemini Is Already Sending You Traffic

This takes under five minutes in GA4.

  1. Log into Google Analytics 4.
  2. Go to Reports, then Acquisition, then Traffic Acquisition.
  3. Change the primary dimension to "Session source / medium."
  4. Filter for "gemini" in the search box.
  5. Set the date range to the last 90 days.

You're looking for sessions from gemini.google.com. Depending on your site's content and topic area, you may already be receiving traffic. If the number is small, that's expected. The point is to establish your baseline now so you can measure growth over the next 6 to 12 months.

If you want to go deeper, add a secondary dimension for "Landing page." This shows you which specific pages are already getting cited by Gemini. Those pages are doing the right things. Study their structure and apply it to your other content.

One more thing to check: if you're seeing growth in "direct / none" traffic over the same period, some of that may be AI-referred traffic that isn't being attributed correctly. Dark traffic from AI is a real measurement gap. But that's a separate audit.

For context on what AI Mode actually cites, our earlier analysis of Google's AI Mode organic citation patterns is worth reading alongside this. And if you want a practical framework for building the kind of expertise content Gemini rewards, our GEO positioning guide covers the approach directly. Both link back to the SEO and GEO work that underpins visibility across all of these platforms.

Frequently Asked Questions

Is Gemini now more important than ChatGPT for SEO? Not in total traffic volume. ChatGPT still drives about 80% of all AI referral traffic and sends approximately 8x more visitors than Gemini as of January 2026. But Gemini's growth rate is faster, and its integration with Google AI Mode gives it structural distribution advantages ChatGPT doesn't have. For businesses already focused on Google, Gemini is the AI platform most directly connected to your existing SEO and GEO work.

Do I need a separate content strategy for Gemini? No. Gemini uses Google's quality signals, so content that performs well in Google Search is already positioned for Gemini citations. The GEO fundamentals including structured headings, FAQ sections, clear author attribution, schema markup, and E-E-A-T signals apply to both. There is no separate Gemini optimisation checklist required.

Why did ChatGPT's referral traffic drop? SE Ranking measured drops of 8.5% in November and 18.6% in December 2025. The exact cause isn't confirmed, but the timing follows ChatGPT's October peak, which coincided with the GPT-5 launch. Some of that initial momentum normalised, and Gemini's product launches in November and December likely captured a portion of new-user growth. ChatGPT remains the dominant AI referral platform overall.

How much of my website traffic should I realistically expect from AI platforms? Currently, very little. All AI platforms combined account for about 0.24% of global web traffic, up from 0.15% in 2025. Organic Google Search still drives the vast majority of search-referred traffic. The reason to build for AI citation now is to establish position early, before the market becomes more competitive, not because AI is currently moving the needle on your traffic reports.

How accurate is this data? SE Ranking's dataset comes from 101,574 sites across 250 markets with Google Analytics installed. It's the largest publicly available dataset on AI referral traffic to date. SE Ranking does sell AI visibility tools, which is worth noting as a potential bias. That said, the directional findings (Gemini growth, ChatGPT decline from peak, Perplexity overtaken) are consistent with patterns reported by Similarweb using a different methodology. Take the precise percentages as indicative, not exact.

Work With Magnified

Ready to turn traffic into leads?

We help SMEs grow with AI-powered SEO, content marketing, and paid ads. If you're getting traffic but not leads — let's fix that.