GEO Series #2 | Getting Into AI Overview Boosts CTR by 35%. A 5-Point Preview of GEO Conference 2026.

GEO Conference 2026 early-bird tickets close April 20th.

June 18th, Washington DC. The third — and final — specialized GEO (Generative Engine Optimization) conference. OpenAI, Google, Anthropic, Adobe, and other core AI search players will be there.

Whether you can attend or not, read this article.

I’m going to preview the five discussion points the conference is likely to center on, with the latest data. This is GEO Series #2, continuing from “Even #1 on Google Won’t Get You Cited by AI”.

The AI search landscape changed just in the past month. CTR numbers updated. New measurement concepts emerged. New tools appeared. I’ve organized everything so you can grasp the full picture in 30 minutes.

CTR Drops to 0.61% With AI Overview — But Citation Entry Reverses That

Numbers first.

In the previous article I cited Ahrefs’ data showing a 58% CTR drop when AI Overviews appear. New data has since come out.

Look at Seer Interactive’s 3,119-keyword study. Organic CTR when AI Overview appears: plummeted from 1.76% to 0.61% (Seer Interactive). A 65% drop. Ad CTR also fell 68%, from 19.7% to 6.34%.

What does “0.61%” actually mean? Search results appearing on page 1 — and fewer than 1 in 100 searchers clicks.

Japan shows a slightly different picture. Per Ahrefs’ 863,000-keyword study, Japan’s AI Overview CTR is 1.8% — slightly above the global average of 1.6%. But still a 38% drop from the pre-AI Overview baseline of 2.9% (Ahrefs Blog).

Here’s where you should focus: the reversal effect.

If your URL is included in the AI Overview citation: organic CTR rises 35%. Paid ad CTR rises 91% (Search Engine Land).

So the situation is: if you’re not cited in AI Overview, CTR craters. If you’re cited, you get more clicks than before. The world has shifted to one where “whether you’re in the citation or not” determines everything.

One more data point: of the URLs cited in AI Overviews, 88% are outside Google’s top 10. The “ranking-citation split” I described last time is now backed by numbers.

High search rank but no AI citation. Low search rank but AI cites you. That asymmetry is the foundation for every discussion at GEO Conference 2026.

What is GEO Conference 2026?

GEO Conference is a specialized conference focused on Generative Engine Optimization (official site).

First held in 2024, second in 2025. This third edition has been officially announced as the final one. Past conferences had limited video sharing; this final conference prioritizes in-person attendees receiving the information first.

Basic Info

• Date: June 18, 2026 (1-day event)
• Venue: National Union Building, Washington DC
• Tracks: Technical + Marketing (2 tracks)
• Past attendees: OpenAI, Google, Anthropic, Adobe, Etsy, L’Oreal, Comcast, Accenture and more
• Special event: Potomac River sunset cruise (2 hours, light refreshments and wine included)
• Early-bird deadline: April 20, 2026

The “final conference” aspect is notable. The organizers said “3 conferences, then our job is done.” That may indicate GEO has grown beyond what a single specialized conference can contain — and is being integrated into larger frameworks.

In 2025, the conference shifted the industry’s understanding of JSON-LD’s importance (Fire&Spark Conference Recap). The knowledge that structured data directly connects to AI citations spread from that event.

What will the final conference debate? The full agenda hasn’t been published. But from the trajectory of the past two conferences and industry movements in early 2026, five discussion points can be read ahead.

A 5-Point Preview of What Will Be Discussed at the Conference

New Performance Metrics After CTR Collapse

In SEO, “ranking” and “click-through rate” were the output metrics. In the GEO era, these are losing function.

ALM Corp’s data captures this sharply: organic share dropped 11–23 percentage points across major industries, while text ads rose 7–13 points (ALM Corp).

The era of “get #1 ranking” as the goal is ending. So what do we measure? Answering that question will likely be the conference’s central theme.

Three candidate metrics (Averi AI):

• Citation Rate: percentage of queries where your site is cited
• Share of Voice: your share of AI search citations vs. competitors
• AI-referred pipeline: conversions attributable to AI search referrals

Especially “AI-referred pipeline” is a fresh concept — tracking not “where did the traffic come from?” but “did AI’s involvement lead to a purchase or signup?” Achievable with GA4 segment setup combined with CRM. It could change how we think about conversion measurement.

The “Next Stage” of Structured Data

I covered JSON-LD importance extensively in the previous article. The 2025 Conference spread the knowledge that structured data directly connects to AI citation visibility.

At the 2026 Conference, “beyond the basics” is what’s likely to be discussed. Article, Person, Organization schema basics have been widely adopted. The focus will shift to advanced FAQPage schema usage and the relationship between HowTo schema and AI citations.

In my xhack.net experiments, articles with FAQPage schema show visibly higher ChatGPT citation frequency. FAQs in the form “What is X?” and “What’s the difference between X and Y?” are especially easy for AI to use as answer material.

My personal data: about 2 weeks after implementing JSON-LD, the target keyword’s ChatGPT citations went from once per week to 3–4 times per week. Small sample, but the directional signal is there.

Machine-Readable E-E-A-T

E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) has long been Google’s quality evaluation standard. In the GEO era, this concept is shifting from “human judgment” to “machine reading.”

Linking Person schema to LinkedIn. Adding DUNS numbers to Organization schema. Structuring author credentials with hasCredential properties. This “making E-E-A-T machine-readable” will likely be the technical track’s central topic.

GEO Auditor Blog also mentions semantic structuring of E-E-A-T (GEO Auditor Blog). Whether you can provide author information in a form AI can “read” is where citation diverges. A major opportunity for individual bloggers.

Proving the Quality Advantage of LLM Referrals

In the previous article I cited Amsive data showing LLM referral conversion rates of about 18% — highest of any channel. This trend is accelerating.

Metricus’ 2026 report is interesting: cases where rankings didn’t change but traffic dropped 35% are being reported (Metricus). Maintaining search position, yet clicks are dropping. This “silent drain” is flowing toward AI search.

The conference will likely feature multiple case studies quantitatively demonstrating the quality advantage of LLM referrals. The marketing track’s focus: how to grow a channel that’s “small in volume but wins on quality.”

For Japanese e-commerce operators: check right now whether “perplexity.ai,” “chatgpt.com,” or “claude.ai” appear in GA4 referral sources. Many will show zero — but when these numbers start appearing, LLM referral becomes “a channel worth considering."

"The Last Conference” and GEO’s Integration

The third and final conference. This fact itself signals GEO’s future.

My read: GEO is past its peak as an independent concept and will be integrated into SEO, content marketing, and brand strategy. Put differently: “marketers who don’t know GEO can’t survive” is approaching.

When a specialized conference becomes unnecessary, it’s evidence that the knowledge has become common. Just as there’s no longer an HTTP/2 specialized conference, GEO will eventually become “basic literacy” for web marketing. Reading that timing is the significance of the final conference.

“Share of Synthesis” — A New Way to Measure

Among the five points, what I’m watching most closely is the “new performance metrics” — specifically, a concept called Share of Synthesis.

Traditional SEO used “Share of Voice” (exposure share in search results) as the basis for competitive analysis. In the GEO era, this evolves into “Share of Synthesis.” The question becomes: when AI “synthesizes” an answer, how much does your content serve as the raw material? That’s the percentage to measure.

How to measure it practically:

Citation Rate

The most fundamental metric. Ask AI questions across a defined set of keywords and measure the probability of your site being cited.

LLMs typically cite only 2–7 domains per answer (Frase.io). Compared to Google’s “10 blue links,” the available slots are dramatically fewer. Whether you make it into that narrow space is the GEO-era competition.

My xhack.net weekly tracking: 5 target keywords × 3 AI search tools (ChatGPT, Perplexity, AI Overview) = 15 queries. Citations ÷ 15 = Citation Rate for that week.

Share of Voice (AI Search)

For comparing yourself vs. competitors:

Your citations ÷ (Your + competitor citations combined) × 100

In highly competitive B2B spaces, category leaders reportedly capture 30–50% SOV (Discovered Labs). If your Citation Rate is 15% but competitors are at 45%, you’re losing 3-to-1.

Current State of Measurement Tools

Honestly, tooling is still maturing.

Siftly offers cross-platform citation tracking across ChatGPT, Gemini, Claude, and Perplexity (Siftly). Peec AI differentiates explicit brand mentions from “silent citations” — where your content was used but not named.

But every tool costs hundreds of dollars monthly — a barrier for individual bloggers and small media.

That’s why I recommend manual tracking. Weekly 20-minute manual checks let you track Citation Rate variation well enough. Rather than waiting for perfect automation, taking action now is more valuable.

The Step 2 AI citation check from my previous article — repeat it weekly — and you’ll develop an intuitive feel for Share of Synthesis changes. Start there.

The State of the Japanese GEO Market

Conclusion: Japanese-language GEO is in a state of “introductory content saturated, practical content absent.”

Search “GEO とは” and you’ll find 15+ Japanese articles: C-NAPS, Tokyo SEO Maker, CINC, Devo — SEO media has published comprehensive introductory coverage. Entering that space now makes differentiation difficult.

But there’s a massive gap in Japanese practical-level GEO content.

Articles explaining AI Overview CTR changes with data: about 10. Of those, ones that actually cite Ahrefs data and mention Japan’s CTR (1.8%): a handful including AIO Lab by GMO.

Even more striking: the “ranking-citation split” framing I introduced in the previous article. Articles that explicitly frame this concept in Japanese: essentially zero.

“Share of Synthesis” and “LLM citation monitoring” at the practical implementation level: nothing found in Japanese.

This is an opportunity.

In the English-speaking world, January–March 2026 saw rapid deepening of GEO discussion. Conference 2025 knowledge diffused, measurement tools appeared, case studies accumulated. Japanese is 6 months behind and staying there.

That’s why I keep writing this GEO series: translating the latest English-language knowledge into actionable Japanese-language format. Almost nobody else is doing this.

The Conference runs in June, knowledge spreads to the English-speaking world July–August, Japanese translation happens after that. There’s no time to wait. Readers who get ahead now gain a significant advantage.

3 GEO Preparation Steps You Can Start Today Even Without Attending

Step 1: Measure Your Own Citation Rate (20 minutes)

1. Open ChatGPT
2. Choose 5 keywords relevant to your site
3. Enter a question for each keyword
4. Check if your site is cited in the answer
5. Run the same 5 questions in Perplexity
6. Citations ÷ total queries = your Citation Rate

My current xhack.net Citation Rate: approximately 13% (2 citations in 15 queries). Still low — but zero before JSON-LD implementation, so the direction is visible.

Step 2: Execute the 3-Step Practical Guide from the Previous Article (65 minutes)

The 3 steps from “Ranking-Citation Split” — if you haven’t done them yet:

• JSON-LD basic implementation (15 minutes)
• AI citation check (20 minutes) — this doubles as Step 1’s Citation Rate measurement
• Structural rewrite of existing articles (30 minutes/article)

These 65 minutes are the shortest path to standing on the same ground as Conference attendees.

Step 3: Record Your Citation Rate Weekly (20 minutes/week)

Repeat Step 1 every week. That’s it.

Set up a spreadsheet with 4 columns: “Date,” “Keyword,” “ChatGPT Citation,” “Perplexity Citation.” Update every Monday morning in 20 minutes. After 4 weeks, you’ll see your Citation Rate trend.

Rising: GEO optimization is working. Flat: content or structured data needs improvement. Declining: competitors may be taking your position.

No expensive tools required — just ChatGPT and Perplexity accounts (both free tier available) and one spreadsheet. Zero startup cost.

Summary

GEO Conference 2026 is June 18th. Early-bird deadline April 20th. The final conference’s five likely discussion points:

• AI Overview CTR crashes to 0.61%. But being cited drives +35% reversal
• 88% of AI-cited URLs are outside Google’s top 10. “Ranking-citation split” confirmed by data
• 5 conference discussion points: new metrics / next-stage structured data / machine-readable E-E-A-T / LLM referral quality / GEO integration
• “Share of Synthesis” emerging as a new success metric — Citation Rate × Share of Voice captures your competitive position in AI search
• Japanese GEO market: “intro content saturated, practice content empty” — first movers gain now

Even without attending the Conference, what to do is clear: measure Citation Rate, implement JSON-LD, record weekly. Start those three in April, and when the industry moves post-Conference in June, you’ll already have months of data.

The essence of GEO isn’t about “appealing to AI.” It’s about delivering content that solves readers’ problems — in a form AI can understand. That’s the same as the SEO origin. What changed is the rules of “delivery.”

Next GEO Series #3: I plan to cover “how to measure Share of Synthesis and a weekly improvement cycle” — including a Python script (with Gen’s help) for automatically tracking AI citation rates for your own articles.

First — use this article to grasp the full picture. Thirty minutes gets you through Citation Rate measurement. Only those who do it stand on the starting line of the AI search era.