category,metric,scope,surface,value,unit,note
research_design,total_queries,study,,500,count,Includes 480 main-panel queries and 20 branded controls
research_design,ai_surfaces,study,,4,count,"ChatGPT, Gemini, Google AI Mode, Google AI Overview"
research_design,replicates,study,,3,count,"Three complete, separate runs"
research_design,total_observations,study,,6000,count,500 queries × 4 surfaces × 3 runs
surface,unprompted_brand_visibility,main_panel,ChatGPT,54.93,percent,n=1440; controls excluded
surface,unprompted_brand_visibility,main_panel,Gemini,64.10,percent,n=1440; controls excluded
surface,unprompted_brand_visibility,main_panel,Google AI Mode,74.93,percent,n=1440; controls excluded
surface,unprompted_brand_visibility,main_panel,Google AI Overview,50.49,percent,n=1440; controls excluded
finding,brand_divergence_min,pairwise_surface_comparisons,,55.88,percent,Minimum across six brand comparison pairs
finding,brand_divergence_max,pairwise_surface_comparisons,,62.20,percent,Maximum across six brand comparison pairs
finding,source_divergence_min,comparable_source_pairs,,82.77,percent,Google AI Mode excluded from source comparison
finding,source_divergence_max,comparable_source_pairs,,91.27,percent,Google AI Mode excluded from source comparison