We flag suspicious audiences by cross-checking behavior, growth and reach patterns. Signals include sudden follower jumps, low viewer-to-follower ratios, abnormal like/comment distributions, repeat accounts and unreachable contact lists. We combine engagement scoring, audience geo/language mismatches and recent-post consistency to mark likely fraud.