SEO Is Not Dead, but GEO Is Not Optional

I've been reading two data-driven Japanese articles published within days of each other, and they seem to flatly contradict each other.

The first, by Mr. Tsuji of Faber Company, presented at the Web Tan Forum conference (Web Tan Forum is one of Japan's leading digital marketing publications, run by Impress Corporation), draws on over 10 years of tracking 5–15% of all Japanese search queries (roughly 200,000 search terms). His conclusion: organic search traffic has barely declined. 99% of businesses can afford to wait and see on AI search. SEO is not dead.

The second, published by AMP, paints a very different picture: AI Overviews are halving click-through rates, zero-click searches have surged from 56% to 69% in a single year, and a projected $33.7 billion GEO market is emerging by 2034. Ranking number one no longer means sales.

I wrote previously in my article on the Google Zero myth that the "death of Google" narrative is overblown. So which of these two is right?

Both of them.

The case for "SEO is broken"

The AMP article's numbers are hard to dismiss:

• AI Overviews went from 6.49% to 13.14% of all search queries in just two months, which is a 72% jump
• Click-through rates halve when an AI summary is shown: 8% versus 15% on regular results
• Even being cited inside an AI Overview gets you only about 1% CTR
• Zero-click searches surged from 56% to 69% in one year
• Global referral traffic from search dropped 6.7%, from 12 billion visits to 11.2 billion
• The GEO services market is projected to grow from $848 million in 2025 to $33.7 billion by 2034 at a 50.5% CAGR

These are structural numbers, not noise. The zero-click trend in particular aligns with what I covered in Part 1 of the AJSA seminar report. The zero-click era is real and accelerating. Search engines are becoming answer engines. The user never leaves.

The case for "SEO is not dead"

Mr. Tsuji's data tells a different story, and it's backed by an unusually large dataset. His team has been continuously tracking a meaningful slice of all Japanese searches for over a decade. As of 2025, the verdict: organic search traffic has "barely declined."

He goes further. 99% of businesses can take a wait-and-see approach to AI search. The panic is premature.

But here's what makes his presentation genuinely interesting. He also showed that over half of the sites that ranked well 10 years ago have disappeared from the top results. SEO has always been volatile, and this kind of churn is not a new phenomenon caused by AI. It has been happening for years.

The real shift his team identified is that Google now evaluates page-level usability, not just site authority. Even Japanese government sites (some of the most trusted domains in the country) dropped roughly 40% in keyword rankings after a 2025 core update, because their pages had poor UX. Meanwhile, government sites with clear layouts, good headings and visual aids held steady.

And as I pointed out in the Google Zero article: ChatGPT drives only about 2% of Google's total traffic volume. The structural shift is real, but the scale is still small.

Both are right, because they're measuring different things

Here's the reconciliation: Mr. Tsuji is measuring the water level. AMP is measuring the riverbed.

Mr. Tsuji's data tracks current traffic volume. The river is still flowing. If you depend on organic search today, the water hasn't dried up. AMP's data tracks structural change in how search works, and the riverbed is clearly shifting course. The water will follow eventually, but it hasn't fully redirected yet.

This is not a contradiction. It's a time-lag.

But the most interesting insight is hiding in plain sight in both articles: user value is the constant that ties them together.

Mr. Tsuji's core conclusion is that Google now measures post-click user satisfaction in real time and reflects it directly in rankings. If users click through to your page and bounce right back to the search results, you drop. If they stay and are satisfied, you rise. The era of gaming rankings with backlinks and keyword density is over. What matters is: did you actually help the person?

AMP's GEO optimisation advice, stripped of the buzzwords, says essentially the same thing: answer questions clearly upfront, back your claims with data, structure your content so it's easy to extract useful information. Be clear. Be credible. Be useful.

These are not opposing strategies. They are the same discipline applied to two different interfaces. Google's algorithm and AI engines are converging on identical criteria: does this content genuinely serve the user? Both sides arrive at the same destination from different directions.

What to actually do

Practical steps, because this is what matters:

1. Don't panic and tear up your SEO. Organic traffic is still flowing. The 99% wait-and-see figure from Mr. Tsuji's data is reassuring if your current strategy is fundamentally sound.

2. Start structuring content for AI citation now. Clear answers upfront, data-backed claims, FAQ-style sections. This is not speculative. It is measurable, and the data backs it up. I covered the practical how-to in detail in Part 2 of the AJSA seminar report.

3. Track whether AI engines reference you, not just where you rank. The metric is shifting from "position on page one" to "cited in the answer." This is the new visibility.

4. Diversify your channels. Don't depend on any single platform, whether that's Google, ChatGPT, or anything else. I argued this in the Google Zero article and nothing since has changed my mind.

5. Focus on what you can prove. Firsthand data, original research, real case studies. This is what both Google's algorithm and AI engines reward. Generic content that could have been written by anyone (or by an AI) is precisely the content that gets pushed out first.

The question isn't whether SEO is dead or alive. It's whether your content is worth citing, by humans or by machines.