Skip to main content
日本語

Research

We publish our measurement methods, assessment criteria and limitations. To check the evidence behind the numbers, see the methodology and data on each page.

Web sales, SEO and AIO

Research into how public websites work as sales channels.

B2B Web Sales Opportunity Loss Study 2026An external audit of public websites belonging to Japanese B2B technology and professional-services companies. The study identifies gaps in search and generative-AI access, comparison, enquiries and ongoing improvement, and publishes assessment criteria and specifications for fixes. · 21 companies · 75 checks · 247 pages · Date: 2026-08-03

AI observation

Observations of what generative AI cites and how it interprets information.

How Does AI Interpret the Word "AIO"?A 2-day, 40-run fixed-point baseline on our own visibility: 0/40 mentions and citations. Of 13 questions containing "AIO," AI read it as security measures, water-cooled coolers, or autonomic-nervous-system diagnosis as often as it read it as AI search optimization — and the interpretation shifted the next day. · 20 observations · Date: 2026-07-05

Who Does AI Search Actually Cite?A fixed 30-question BtoB procurement study on OpenAI Web Search, with every citation classified by hand. Under a realistic setting, 29 of 30 questions returned zero citations; forcing a search on every question, 94% of the 137 citations went to companies' own official domains. · 30 observations · Date: 2026-07-04

What Do Cited Pages Actually Look Like?We re-fetched all 137 citations from our own study of who AI cites and classified their structure. 43% were a company's top page; the dominant title pattern was 'what we offer + company name'; 60% carried structured data, but 40% had none at all. · 137 observations · Date: 2026-07-05

AI Visibility Monthly IndexMonthly records of brand mention rates and official-site citation rates. · Date: 2026-07-05

Web3 and real-world implementation

Research covering implementation and institutional frameworks.

Web3 LabResearch records on blockchain implementation, institutional frameworks and operation. · Date: 2026-07-05

Methodology and data

Measurement methods, assessment criteria, limitations and aggregate data.

Methodology for auditing web sales from public information(日本語)What we measure, our assessment criteria, how we handle what cannot be observed, limitations and version management. · public-web-sales-audit-v1.0 · Date: 2026-08-03 · Available in Japanese.

B2B Web Sales Opportunity Loss Study 2026 — Aggregate data(日本語)Counts, proportions and confidence intervals by assessment item, with variable definitions. Available as CSV. · dataset v1.0 · Date: 2026-08-03 · Available in Japanese.