Arrow AI research library
These market studies assess public website signals. Each has its own sample, observation date and method. A website score does not measure how often an AI answer mentions, recommends or cites a company. Inspect the evidence available for each study before drawing a conclusion.
100 yachting companies covered: 95 scored and 5 not scannable in the historical collection. The rounded mean of the 95 scores is 57/100; the maximum is 81/100. Original values are preserved, with the legacy method and its limits explained.
Observed 2026-08-19 · 100 covered · definitions clarified 2026-09-07
25 med spas and aesthetic clinics scored across five families of public website signals. This measures AI visibility readiness, not clinical quality.
Updated 2026-09-01 · 25 providers
Interpret the evidence
Use the citation measurement methodology to separate technical source checks, provider API responses and observations from consumer AI products. Different methods or dates need separate series; a new score should not silently replace a historical one.
The platform buyer framework helps assess what a tool actually measures. For the yacht study, the yacht visibility guide connects source facts with buyer questions. Find the wider reading path in the GEO resource library.
Define the sector, location and buyer questions. We can scope a study with explicit source checks and, where available, separately recorded AI answer observations.
