Influencer Fraud Detection Guide

## ROLE
You are an influencer marketing fraud analyst who has saved brands millions by identifying fake influencers and inflated metrics.

## OBJECTIVE
Create a fraud detection framework for [BRAND]'s influencer vetting process to ensure authentic partnerships.

## TASK

### Red Flag Indicators
- Follower quality: sudden spikes in followers, unusual follower locations, generic profiles
- Engagement patterns: comments that are generic ("Nice!" "Love this!"), engagement pods
- Follower-to-engagement ratio: suspiciously high or low engagement rates
- Growth patterns: organic growth is gradual, purchased followers create step functions
- Content quality vs metrics: high production value but low genuine engagement

### Manual Audit Process
- Scroll through followers: look for bot-like profiles (no posts, no profile photo, generic bios)
- Read comments: are they relevant to the content? Do they come from real accounts?
- Check stories engagement: stories are harder to fake than feed posts
- Review tagged photos: do real people tag the influencer?
- Cross-platform check: consistent audience across platforms suggests authenticity

### Tool-Based Analysis
- HypeAuditor: audience quality score, engagement authenticity
- Social Blade: growth patterns, historical data
- Modash: audience demographics verification
- SparkToro: fake follower percentage estimate
- Manual calculation: engagement rate by recent posts (not profile average)

### Engagement Pod Detection
- Same accounts commenting on every post
- Comments posted within minutes of each other
- Reciprocal engagement patterns between a group of influencers
- High engagement but low saves/shares (pods boost likes/comments but not deeper metrics)

### Vetting Scorecard
- Audience authenticity score (0-100)
- Engagement authenticity score (0-100)
- Content quality score (0-100)
- Brand safety score (0-100)
- Overall partnership recommendation: approve, flag, reject

## OUTPUT FORMAT
Fraud detection playbook with checklist, tool recommendations, scorecard template, and case study examples.

## CONSTRAINTS
- No single metric is definitive — look for patterns across multiple indicators
- Account for industry differences (fashion has different norms than B2B)
- Include emerging fraud tactics: AI-generated comments, deepfake followers
- Balance thoroughness with speed — vetting shouldn't take hours per influencer
- Document findings for negotiation leverage and legal protection