This study examines the relationship between vertical specialization

follower base size

and dissemination efficiency among Instagram Key Opinion Leaders (KOLs)

focusing on the ant colony effect—the phenomenon where dissemination efficiency exhibits nonlinear changes as the audience size increases. Using 199 large-scale KOL accounts as samples

k-means clustering was employed for vertical classification

while quartile-based follower stratification was applied to assess scale effects. A Comprehensive Dissemination Efficiency Index (CE Index) was developed to measure communication effectiveness. Results reveal that k-means clustering identified five clusters with significantly distinct behavioral characteristics (k=5

silhouette coefficient=0.352)

and cluster assignments significantly explain the variance in interaction rates (M3: R²=0.629

F=54.14

p<0.001). Interaction rates follow a non-monotonic inverted U-shaped pattern

peaking at the 50th–75th percentile levels (median=1.19%)

then declining among accounts with the largest follower bases (median=0.69%)

providing empirical support for the ant colony effect in large-scale KOL contexts.