Study: Are Higher Google Organic Rankings More Likely to Be Mentioned or Cited in AI Overviews?

Published: 2026-08-29 by Heysho

SEO and GEO/AI search are often described as closely related.

But what I personally wanted to understand was: How closely are they actually connected?

For example, if improving organic search rankings through SEO also increases the likelihood of appearing in AI Overviews, how high do you actually need to rank?

Is reaching Rank 1 important? Is being in the Top 5 enough? Or does simply making the Top 10 significantly increase the likelihood of being mentioned or cited in an AI Overview?

I decided to explore these questions using actual search data.

Specifically, this study examines the relationship between Google organic search rankings and mentions and citations in AI Overviews.

I selected product-category keywords with search demand across five industries in Japan and created approximately 1,000 “recommendation”-style queries.

For each query, I collected Google organic search results and, where an AI Overview appeared, analyzed:

• Whether brands and pages ranking highly in organic search also appeared in the AI Overview
• How Mention Rate and Citation Rate changed from Rank 1 to Rank 2, Rank 3, Rank 4, and beyond
• How the results differed between the Top 5, Ranks 6–10, and Ranks 11–15

The goal was to use actual data to examine a simple question:

How important is ranking highly in Google Search in the age of AI search?

Disclosure: This study was sponsored by SearchApi, which provided API credits for data collection. The study design, data processing and analysis, interpretation of the results, and article content were carried out independently by the author.

1. Research Findings

1.1. The Relationship Between Google Organic Rank and AI Overview Mention Rate

First, let’s look at the AI Overview Mention Rate.

Here, “Mention Rate” refers to the percentage of cases in which a brand appearing in Google organic search results was also mentioned in the body of the AI Overview for the same keyword.

The results are shown below.

The brand ranking #1 in Google organic search was also mentioned in the AI Overview in 20.6% of cases.

Although there is some variation across individual positions, the overall trend is clear: the higher a brand ranks in organic search, the more likely it is to be mentioned in an AI Overview.

The table below compares Mention Rates across three ranking groups.

These results suggest that reaching Google’s first page, roughly the Top 10, is not the whole story. Being in the Top 5 appears to make a substantial difference in the likelihood of being mentioned in an AI Overview.

1.2. The Relationship Between Google Organic Rank and AI Overview Citation Rate

Next, let’s look at the AI Overview Citation Rate.

Here, “Citation Rate” refers to the percentage of cases in which the exact same URL appearing in Google organic search was also included as a reference link in the AI Overview.

AI Overviews typically display links to Web pages used as supporting sources for the generated answer. In this study, I examined whether those reference links included the same URLs that appeared in the organic search results.

The results are shown below.

The URL ranking #1 in organic search was also cited in the AI Overview in 52.3% of cases, meaning that more than half of Rank 1 URLs were cited.

The rate fell to 45.3% at Rank 2 and 36.7% at Rank 3. It declined further to 23.9% at Rank 5, 6.4% at Rank 10, and just 1.4% at Rank 15.

While there is some variation between individual positions, the overall pattern is striking: pages that rank higher in organic search are substantially more likely to be cited in AI Overviews.

The results by ranking group are as follows.

The Citation Rate for Ranks 1–5 was 38.0%, which was approximately 3.3 times higher than the 11.4% observed for Ranks 6–10.

For Mention Rate, the corresponding difference between Ranks 1–5 and Ranks 6–10 was approximately 2.2 times.

In this dataset, therefore, exact URL citations showed a stronger relationship with organic search rank than brand mentions did.

1.3. Comparing Mention Rate and Citation Rate

Taken together, these results suggest that, from the perspective of earning citations in AI Overviews, there may be meaningful value not only in reaching the Top 10, but also in pushing into the Top 5 and potentially even higher positions.

I also calculated rank correlation coefficients using the Mention Rate and Citation Rate for each position from Rank 1 through Rank 15.

The correlation coefficient was -0.94 for Mention Rate and -0.99 for Citation Rate.

Because a higher rank number represents a lower search position, these negative values indicate that Mention Rate and Citation Rate tend to increase as organic search ranking improves.

Note: These correlations are based on 15 rank-level aggregate observations and show association only; they do not demonstrate a causal relationship.

2. Insights From the Study

2.1. Search Rank Shows a Strong Association With Citation Rate

One of the clearest findings from this study is the strong association between organic search rank and citation rates in AI Overviews.

The higher a page ranked in Google Search, the more likely it was to be cited as a reference link in an AI Overview.

In particular, the page ranking #1 in organic search was also cited in an AI Overview in 52.3% of cases. The rate fell to 45.3% at Rank 2 and 36.7% at Rank 3, showing a substantial decline as rankings decreased.

2.2. Brand Mentions Also Showed an Association With Search Rank

Brand Mention Rate showed a similar pattern: brands ranking higher in organic search tended to be mentioned more frequently in the body of AI Overviews.

However, the relationship with search rank was weaker than it was for Citation Rate.

Mention Rate was 12.9% for Ranks 1–5, falling to 5.7% for Ranks 6–10 and 2.9% for Ranks 11–15.

Personally, I found these rates lower than expected.

This suggests that factors beyond organic search rank may also play an important role in determining whether a brand is mentioned in an AI Overview.

From my own observations of AI search results, some brands that are not particularly strong in Google Search still appear frequently in AI-generated answers.

These often seem to include brands that are frequently featured by influencers on platforms such as YouTube or Instagram, or that generate discussion on social media.

Conversely, there are also brands that rank highly in Google Search but have relatively limited visibility in AI search.

At this stage, this is only an observation-based hypothesis. However, signals beyond search rankings, such as social media exposure, video content, and broader brand visibility across the Web, may also be associated with brand mentions in AI search.

I would like to investigate this further with data in future studies.

2.3. Hypothesis 1: Google Search Is Part of AI Overview Answer Generation

Why might organic search rank be so strongly associated with AI Overview Citation Rate?

One possible explanation is that Google Search itself plays a role in how AI Overviews and AI Mode gather information.

According to Google’s documentation, AI Overviews and AI Mode may use a technique known as Query Fan-out, in which multiple related searches are performed across subtopics and data sources.

For example, a query such as 「Tシャツ おすすめ」 (“recommended T-shirts”) might lead to related searches such as:

• 「20代 男性 Tシャツ おすすめ」 (“recommended T-shirts for men in their 20s”)
• 「30代 女性 Tシャツ おすすめ」 (“recommended T-shirts for women in their 30s”)
• 「カジュアル Tシャツ おすすめ」 (“recommended casual T-shirts”)
• 「フォーマル Tシャツ おすすめ」 (“recommended formal T-shirts”)

In this way, the original query can be expanded into multiple related searches, with information gathered from different perspectives before an answer is generated.

In other words, AI Overviews are not entirely separate from Google Search. Search appears to play a role in retrieving information from the Web during the answer-generation process.

Given this mechanism, websites that consistently rank highly across a range of related searches may have more opportunities to be retrieved and referenced through Query Fan-out, potentially increasing their likelihood of being cited in AI Overviews.

However, even URLs ranking #1 in organic search had an AI Overview Citation Rate of only 52.3% in this study.

If AI Overviews simply used the search results for the user’s original query as-is, we would expect a much stronger overlap between organic rankings and citation sources.

One possible explanation for this gap is that AI Overviews do not rely solely on the original query. Instead, they may gather information through multiple related searches generated through Query Fan-out.

From a GEO perspective, this suggests that ranking #1 for a single keyword may be less important than consistently appearing prominently across a broader set of related search intents.

2.4. Brand Mentions Are Not Determined by Search Rank Alone

Brand mentions, however, may need to be viewed somewhat differently.

This study found that brands with higher Google organic rankings were more likely to be mentioned in AI Overviews, but the association was weaker than it was for URL citations.

This may suggest that the process of selecting which Web pages to use as sources differs from the process of deciding which brands to mention or recommend in the final answer.

Rather than simply choosing brands that appear in a particular set of search results, an AI-generated answer may combine learned knowledge, including training data, with information gathered from multiple sources and related searches.

For example, if Query Fan-out triggers multiple related searches and the same brand is repeatedly recommended across many different sources, that brand may have a greater chance of being recognized as relevant to the topic.

With that in mind, increasing brand mentions may require more than simply getting your own website to rank highly in Google Search. It may also involve:

• Being recommended in comparison articles
• Being featured in media coverage
• Appearing in videos on platforms such as YouTube
• Generating discussion on social media
• Being mentioned positively across a variety of websites

In other words, it may become increasingly important to build broader awareness and reputation across the Web so that a brand is consistently associated with a particular topic or category.

To summarize: if SEO is primarily about “making your own pages easier to discover in search results,” GEO/AI search may also require “building a presence in which your brand is consistently recommended across the Web.”

2.5. Hypothesis 2: SEO and AI Search May Favor Similar Types of Information

Another possibility is that the websites and brands most likely to perform well in both Google Search and AI search share certain characteristics.

Google Search and generative AI do not operate in exactly the same way. However, both ultimately aim to surface information that is relevant and useful in response to a user’s question or search intent.

As a result, websites and brands with characteristics such as the following may be more likely to perform well in both environments:

• Providing extensive, expert information on a specific topic
• Being highly relevant to users’ search intent
• Presenting information in a clear and well-organized way
• Being consistently mentioned across multiple websites and media channels
• Having established a certain level of visibility and recognition within their field

In other words, the relationship may not simply be:

Higher Google ranking → greater likelihood of appearing in AI search

There may also be an underlying structure in which:

Shared qualities that make information easier to evaluate → greater visibility in both Google Search and AI search

From this perspective, SEO and GEO may be better understood not as either “the same thing” or “completely separate disciplines,” but as sharing a common foundation while also having distinct mechanisms for information retrieval, selection, and presentation.

3. Study Design and Analysis Method

From here, I will explain how this study was conducted.

3.1. Keywords Studied

This study covered the following five industries:

• Beauty
• Fashion
• Home appliances
• Sports and outdoors
• Furniture

For each industry, I selected 200 product-category keywords in descending order of search volume and combined each one with the Japanese term 「おすすめ」 (“recommended” / “best”).

For example:

• 「Tシャツ」 (T-shirts) → 「Tシャツ おすすめ」 (“recommended T-shirts”)
• 「ドライヤー」 (hair dryers) → 「ドライヤー おすすめ」 (“recommended hair dryers”)

With 200 queries across five industries, the initial dataset contained 1,000 keywords.

However, some keywords appeared in more than one industry, resulting in 972 unique keywords.

I collected Google search results for all 972 keywords using SearchApi. Of these, 700 keywords for which an AI Overview was successfully retrieved and analyzed were included in the main analysis.

3.2. Google Search Collection Conditions

Google Search and AI Overview conditions settings were as follows:

Organic search results were collected through Page 2.

However, Google does not always display exactly 10 organic results per page. In many cases, only eight or nine organic results appeared, meaning that the number of observations decreased substantially at Ranks 17, 18, 19, and 20.

To keep the sample sizes across ranking positions as consistent as possible, the main analysis was therefore limited to Ranks 1–15.

3.3. Using SearchApi to Collect Search Results

https://www.searchapi.io/

For this study, I used SearchApi to collect Google organic search results and AI Overview data.

SearchApi is an API service that provides structured SERP data, including Google organic results and AI Overviews. For this study, I collected organic search result fields, AI Overview content in Markdown format, and AI Overview reference links using the settings gl=jp, hl=ja, and Desktop.

Because SearchApi is an API-based service, using it generally requires some programming knowledge.

However, coding AI tools such as Codex and Claude Code have made API-based data collection much more accessible than it was in the past.

I expect that data analysis will increasingly involve collecting and processing data directly through a combination of APIs and coding AI tools, rather than relying solely on SaaS dashboards. If this type of workflow interests you, it is well worth experimenting with.

3.4. Definition of Mention Rate

The first metric examined in this study was Mention Rate.

Mention Rate measures whether a brand appearing in Google organic search results was also mentioned in the body of the AI Overview for the same keyword.

One challenge in measuring brand mentions was variation in how brand names are written.

For example, the same brand may appear in different forms depending on:

• English vs. Katakana spelling
• Official names vs. abbreviations
• The presence or absence of spaces
• Uppercase vs. lowercase letters

As a result, the same brand may appear under slightly different names in organic search results and AI Overviews.

To account for these variations, I created a dictionary of brand-name patterns and performed entity resolution so that different name variants were treated as the same brand when appropriate.

3.5. Definition of Citation Rate

The second metric was Citation Rate (Exact URL Citation Rate).

AI Overviews display links to Web pages that serve as supporting sources for the AI-generated answer.

For this study, I compared those AI Overview reference links with the URLs appearing in Google organic search results.

More specifically, Citation Rate measures:

“Whether the exact same URL appearing in organic search was also cited in the AI Overview”

A match was counted only when the same page appeared in both places after URL normalization. A citation to a different page on the same website was not considered a match.

4. Future Outlook

I plan to continue conducting research in this area.

This study focused on the relationship between Google organic search results and AI Overviews, but I would like to broaden the scope in future research.

For example, by comparing AI Overviews with other AI search and generative AI services such as Google AI Mode, ChatGPT, and Gemini, I hope to investigate which types of brands and websites are more likely to be mentioned or cited by each platform.

I would also like to look beyond Google Search and examine other channels that may influence brand mentions, such as YouTube and social media, to better understand how broader Web visibility relates to AI search visibility.

5. Notes

5.1. Data Usage and Citation

Please refrain from reproducing or redistributing text, figures, tables, analytical content, or other materials from this article for which the author holds copyright or other rights, except as permitted by applicable law.

Quoting or referencing this article in reports, articles, social media posts, presentations, and other materials is welcome. When doing so, please include a link to this article and credit “Heysho” as the author.

Please note that factual information, including numerical data, may not itself be protected by copyright. Nothing in this section is intended to restrict quotation or other uses permitted under applicable copyright law.

If you would like to reproduce figures or tables, redistribute a substantial portion of the dataset, use the data at scale, or reuse the material for commercial purposes, please contact me in advance.

5.2. Study Notes

This study is an exploratory analysis of Japanese-language, product-category “recommendation”-style queries conducted in the Japan region on desktop devices at the time of data collection.

Google search results and AI Overviews may vary over time and depending on factors such as region, device, and user environment.

This study examines the association between organic search rank, Mention Rate, and Citation Rate. It does not establish a causal relationship in which higher search rankings directly cause brands or pages to be mentioned or cited in AI Overviews.

This study was conducted independently of Google LLC and does not represent Google’s approval, sponsorship, or official views.

5.3. Acknowledgements

Finally, I would like to thank SearchApi once again for sponsoring this study and providing API credits for data collection.

Conducting an analysis of Google search results and AI Overview data at the scale of approximately 1,000 keywords required a stable way to collect SERP data. SearchApi’s support made it possible to carry out this research at that scale.

The study design, data processing and analysis, interpretation of the results, and conclusions were all developed independently by the author.