Fake followers
Fake followers are purchased, bot or inactive accounts inflating audience size. From outside a platform they can only be estimated, never proven: the working signals are engagement far outside the tier's plausible band, generic comment walls, vertical growth steps without explaining events, and audience geography unrelated to the content. Single signals have innocent explanations; three converging justify walking away. Any tool claiming definitive detection is overclaiming.
Worked example
40k followers, 90 likes per post, comments in a language the creator does not publish in, growth in two vertical jumps — a pass, on converged evidence.
Related: Engagement rate · Audience demographics
Full context in the guides, or measure it yourself with the free tools.