[Original Survey] Is the Era of "Comparison & Recommendation Articles" Over in the Age of AI Search?

In the field of SEO and content marketing, where professionals are considering how to adapt to Google AI Overview (AIO), the following opinions are often heard regarding comparison and recommendation articles:

"Are comparison articles already saturated?"
"In the age of AI search, aren't comparison and recommendation articles already obsolete?"
"Q&A formats seem more likely to be cited by AI."
"Comparison articles are already saturated; isn't it too late to invest in them now?"

EXIDEA Inc. (Headquarters: Chuo-ku, Tokyo; Representative Director: Takuma Ogawa) independently analyzed 19,826 Japanese URLs cited by Google AI Overview, categorizing them by content type. The results revealed trends that contradict these common perceptions.

This release shares four key facts that can help SEO and content professionals design content for the AIO era.

Key Findings of This Survey

Even in the AI search era, comparison and recommendation articles remain a valuable SEO asset that is readily cited by AIO.

However, the key is not the format of "Top X Recommendations" but the substance, such as numerical evidence, comparison axes, pros and cons, and practical advice.

For future SEO, it is crucial to review existing comparison articles to make them "easily citable by AI."

Survey Overview

Item: Content
Survey Target: URLs cited in AI Overview responses on Google Search
Number of Articles Analyzed: 19,826
Classification Used: 10 content type categories (automatic classification by LLM)
Analysis Methods: ① Comparison of AIO citation rates by category ② Multivariate logistic regression (Elastic Net regularization) of 40 features including structural elements of comparison articles ③ Cross-tabulation of industry × content type
Objective Variable: Whether the URL was cited by AIO (binary classification)
Industry Classification: Query classification by LLM (Informational/Commercial/Transactional/Navigational × 14 industry categories) linked to article URLs
Data Collection Period: February to March 2026
Survey Entity: EXIDEA Inc.

① "Comparison/Ranking Articles" Top AIO Citation Rate (46.0%) Among 9 Categories

Looking at AIO citation rates by content type, the "Comparison/Ranking" type ranked first at 46.0%, approximately 2.4 times higher than the lowest-ranked "Q&A/Forum" type (19.3%).

AIO Citation Rate by Content Type (n=19,826 articles). "Comparison/Ranking" is first at 46.0%, approximately 2.4 times that of "Q&A/Forum" (19.3%). "General Articles" are a default category for unclassifiable articles and are treated as a reference value.

Rank: 1, Content Type: Comparison/Ranking, Number of Articles: 772, AIO Citation Rate: 46.0%
Rank: 2, Content Type: Glossary/Dictionary, Number of Articles: 622, AIO Citation Rate: 44.2%
Rank: 3, Content Type: How-to/Tutorial, Number of Articles: 461, AIO Citation Rate: 36.9%
Rank: 4, Content Type: Product/Service, Number of Articles: 1,956, AIO Citation Rate: 33.4%
Rank: 5, Content Type: News/Press Release, Number of Articles: 529, AIO Citation Rate: 29.7%
Rank: 6, Content Type: LP/Conversion, Number of Articles: 500, AIO Citation Rate: 27.0%
Rank: 7, Content Type: Official Reference, Number of Articles: 303, AIO Citation Rate: 26.1%
Rank: 8, Content Type: Case Study/Interview, Number of Articles: 136, AIO Citation Rate: 25.0%
Rank: 9, Content Type: Q&A/Forum, Number of Articles: 616, AIO Citation Rate: 19.3%
Reference: General Article, Number of Articles: 13,931, AIO Citation Rate: 47.7%

Note that "General Articles" are a default category for all articles that do not clearly fit into other categories and are not a meaningful comparison target, so they are treated as a reference value.

Comparison/Ranking Articles Rank First

The most powerful content type for being cited by AIO is "Comparison/Ranking Articles," which can be expected to have 1.5 to 2.4 times the citation efficiency compared to Q&A or LP types. The assumption that "comparison and recommendation articles will weaken in the AIO era" is not supported by the data.

② The "Structure of Content" Matters More Than the "Comparison Article" Label

What is important is not just that an article is a comparison article.
Articles that are easily cited by AI share commonalities in the substance of their comparisons.

Effect of Features Related to Comparison/Ranking Articles on AIO Citation (Univariate Difference vs. Multivariate Coefficient Controlling for 40 Items. OR is for "Present vs. Absent").

Rank: 1, Feature: Practical Advice, Univariate Difference (pt): +16.5pt, OR (Present vs. Absent): 1.25
Rank: 2, Feature: Numerical Evidence, Univariate Difference (pt): +10.4pt, OR (Present vs. Absent): 1.13
Rank: 3, Feature: Pros and Cons, Univariate Difference (pt): +12.1pt, OR (Present vs. Absent): 1.10
Rank: 4, Feature: Balanced Discussion, Univariate Difference (pt): +9.9pt, OR (Present vs. Absent): 1.07
Rank: 5, Feature: Tabular Data, Univariate Difference (pt): +11.2pt, OR (Present vs. Absent): 1.07
Rank: 6, Feature: Comparison/Ranking Structure, Univariate Difference (pt): +2.6pt, OR (Present vs. Absent): 1.03

Interestingly, the "substance of the comparison" (practical advice, numerical evidence, pros and cons, tables) shows a larger positive association after controlling for other features than the "comparison/ranking article label." The most explanatory feature was "practical advice," where articles with this element had approximately 1.25 times the odds of AIO citation compared to those without.

Simply listing "Top X Recommendations" is insufficient. Articles that compare with numerical data, present both pros and cons, and provide practical advice tend to be more readily cited by AI.

③ In Education, Lifestyle, and IT, AIO Citation Rates for Comparison Articles Exceed 60%

Looking at the data by industry, there are clear sectors where the effect of comparison/ranking articles is particularly strong. The top 10 AIO citation rates for comparison articles, limited to industries with 5 or more articles, are as follows:

Heatmap of AIO Citation Rate (left) and Sample Size (right) by Industry × Content Type. Citation rates exceed 60% for "Comparison/Ranking" × "Education/Learning," "Lifestyle/Living," and "IT/Technology," with "Finance/Insurance" also high at 58.4%. "General Articles" are excluded as a reference value.

Rank: 1, Industry: Education/Learning, Number of Comparison Articles: 27, AIO Citation Rate: 66.7%
Rank: 2, Industry: Lifestyle/Living, Number of Comparison Articles: 69, AIO Citation Rate: 65.2%
Rank: 3, Industry: IT/Technology, Number of Comparison Articles: 186, AIO Citation Rate: 62.4%
Rank: 4, Industry: Finance/Insurance, Number of Comparison Articles: 113, AIO Citation Rate: 58.4%
Rank: 5, Industry: Marketing, Number of Comparison Articles: 54, AIO Citation Rate: 55.6%
Rank: 6, Industry: Beauty/Fashion, Number of Comparison Articles: 31, AIO Citation Rate: 54.8%
Rank: 7, Industry: Travel/Tourism, Number of Comparison Articles: 11, AIO Citation Rate: 54.5%
Rank: 8, Industry: Food & Beverage/Gourmet, Number of Comparison Articles: 21, AIO Citation Rate: 52.4%
Rank: 9, Industry: Housing/Renovation, Number of Comparison Articles: 12, AIO Citation Rate: 50.0%
Rank: 10, Industry: HR/Career, Number of Comparison Articles: 37, AIO Citation Rate: 45.9%

Please note that for some categories like Travel/Tourism and Housing/Renovation, the sample size is small, so the figures should be viewed as reference values for understanding trends.

In three industries—Education/Learning (66.7%), Lifestyle/Living (65.2%), and IT/Technology (62.4%)—over 60% of comparison/ranking articles are cited by AIO. Finance/Insurance also showed a high rate of 58.4%, close to this level. These are industries where users frequently search in a context of "choosing from multiple options," and AI tends to actively summarize comparative information.

On the other hand, in industries where decision-making is more consultation-based, such as HR/Career (45.9%), the citation rate for comparison articles is more modest compared to other industries.