How Do E-commerce Sites Stand Out in AI Mode Results?

We conducted a detailed analysis of the reference sites returned in AI Mode results to identify factors that could influence being cited as a reference. To understand how SEO processes are evolving, let’s examine the results of this study together.

Key take aways;

• Objective: Measure how well on-page elements in AI Mode referenced URLs align with search intent and which elements most increase the chance of being referenced.
• Scope: 200 keywords, about 3,931 rows and nearly 4,000 URLs. USA location, English language.
• Method: Computed a single Overall Alignment Score from text similarity and rule-based checks across title, meta, H1, H2, schema, and slug.
• Core metrics: Average Overall Alignment 0.5583. Keyword presence rates Title 0.815, Meta 0.815, H1 0.151, H2 0.006, Schema 0.000, Slug exact phrase 0.015.
• Biggest lifts: Keyword in Title and Meta adds about +0.6786 to the score. H2 present +0.2831, Slug exact phrase +0.2858, H1 present +0.2762.
• Slug: Average token overlap 0.296. Clean, intent-reflecting slugs are helpful.
• Schema intent match: Overall 0.295. Best segments Commercial investigation 0.5311 and Informational how-to 0.4891. Local transactional is weak at 0.097.
• Most frequent schema types: Organization 1765, BreadcrumbList 1457, Article 666, WebSite 633, Product 402. ItemList and HowTo are underused where expected.
• Conclusion: Prioritize rewriting Title and Meta to match intent, enforce a single aligned H1, use H2s for sub-intent coverage, apply the correct schema types per intent, and adopt a concise slug policy.

Background and Objective

We reviewed reference results fetched via Google AI Mode for USA location and English language. Each row represents a URL and the page’s headline and structural elements. Our objective is to see how well the on-page elements in results referenced by AI Mode align with search intent and to measure which elements increase alignment the most. This helps content and technical teams prioritize and deliver quick wins.

Here, 200 different keywords and nearly 4,000 reference URLs were examined.

Detailed data table

https://docs.google.com/spreadsheets/d/1fyNK-m_gRX7VA1txawo23B4Sq15H1zZpnmL0g6UmZ6k/edit?usp=sharing

Tools Used

• Semust
• DataforSEO
• n8n
• ChatGPT
• Apps Script

Business Questions

• Does using the keyword in title, meta, H1, H2, schema, and slug increase alignment
• Do schema types match search intent, and are there repeatable correct schema patterns
• Do slugs provide meaningful alignment with the query

Data Summary

• Row count: 3931
• Unique keyword count: 2563
• Keyword presence rates by field: Title 0.815, Meta 0.815, H1 0.151, H2 0.006, Schema 0.000, Slug exact phrase 0.015
• Average vector similarity scores: Title 0.8153, Meta 0.8153, H1 0.4984, H2 0.1327, Schema 0.0016
• Average slug token overlap: 0.296
• Average Overall Alignment Score: 0.5583

Top 10 Most Frequent Sites

Method Summary

• Bag-of-words cosine similarity was used for text alignment.
• A composite Overall Alignment Score was computed. Weights: Title 0.30, H1 0.25, Meta 0.20, H2 0.10, Schema 0.05, Slug 0.10. These were measured across all rows.
• Schema @type values were extracted and matched to keyword intent classes. Intent rules were defined as Transactional, Commercial investigation, Informational how-to, Informational definition, Local transactional, General informational.

Key Findings

What is the contribution of Title and Meta usage

• The presence of the keyword in these fields increases the Overall Alignment Score by about 0.6786 points.
• This shows that intent-consistent title and meta writing is critical.

What is the contribution of H1 and H2

• H1 presence rate is 0.151. When present, the average contribution is 0.2762 points.
• H2 presence rate is 0.006. When present, the contribution is 0.2831 points.
• For H2, semantic subheadings that break down the intent are more meaningful than exact repetition.

What is the alignment between slug and keyword

• When the exact phrase appears in the slug, the contribution is 0.2858 points.
• Average token overlap is 0.296. Clean, intent-reflecting slugs are useful.

How is schema intent alignment

• Overall schema intent alignment is 0.295.
• Highest matches: Commercial investigation 0.5311 and Informational how-to 0.4891.
• Transactional 0.444, General informational 0.2059, Local transactional 0.097.
• Most frequent types: Organization 1765, BreadcrumbList 1457, Article 666, WebSite 633, Product 402.
• ItemList and Product are not as common as expected in the relevant intents. How-to pages often lack the HowTo type.

Do Backlinks Influence Being Cited in Google AI Mode

• The average DR of referenced domains is 69.39.
• The average number of referring domains is 427.
• The average total backlink count is about 85,000, but the distribution is skewed. Some sites have no backlinks, and very large sites inflate the average.
• Across 3,900+ URLs, 1,891 have no backlinks. 2,427 URLs have fewer than 10 referring domains.

In our AI Mode analysis, backlinks seem helpful for appearing as a reference, yet they are not strictly required. The data confirms this.

Why We Did This Study

• We wanted to answer whether alignment between pages referenced by AI Mode and search intent affects visibility and click performance.
• We aimed to see, in numeric terms, which fields deliver the fastest and highest return when optimized.
• By measuring schema intent alignment, we sought to highlight the importance of correct @type and required fields instead of simple keyword repetition.