面向营销人员的反移动作弊指南_您需要知道的一切（英文版）_45页_4mb

The Marketer's Guide to Mobile Fraud Summary

Core Content Overview

The document outlines the challenges and strategies for combating mobile fraud in the context of increasing mobile ad spend and the evolving landscape of digital marketing. It emphasizes the importance of understanding various types of fraud, their impact on ROI and campaign performance, and the tools and methods available to detect and prevent them.

Main Fraud Types and Key Information

Basic Fraud

• Forced, fake or random clicks, installs, and activities.
• Click Fraud:
  • Fraudsters "fish" for installs by sending random clicks.
  • Forced Clicks: No X to close windows hiding content.
  • Click Redirect: App store opens instead of the desired content.
  • Misreporting Impressions: Reporting video views as clicks.
  • Bot Clicks: Automated clicks generated by bots.

Basic Install Fraud

• Simulated bots / farms create both a click and an install.
• Bot/Server Installs: Fraudulent installs generated by automated systems.

Basic Engagement Fraud

• Bots follow pre-programmed engagement scripts to improve retention and revenue.
• CPA Bots: Bots that mimic user behavior to manipulate campaign performance.

Emerging Engagement Fraud

• Manipulation of SDK messages by exploiting open source SDK vulnerabilities.

Man-in-the-Middle Attacks

• Device-based malware intercepts SDK-server communication to validate fraudulent sessions or in-app events.

Retransmission Fraud

• Repeating valid messages to the server to inflate engagement metrics.

Basic Protection Measures

• IP, user agent filtering (blacklisting).
• Basic distribution modeling (time-to-install).
• Install & receipt validation.
• Secured SDK message encryption & deduplication.

Advanced Fraud and Protection

Targeting Your Attribution

• Advanced fraud can skew ROI data and misattribute installs.

Click Flooding

• Definition: Criminals send a massive number of clicks to capture the last click before an install.
• Protection: Look for even or highly linear patterns in Click to Install Time (CTIT) and low conversion rates.

Install Hijacking

• Definition: Malware hidden in apps detects an app download and sends fraudulent clicks to capture attribution at launch.
• Protection: Monitor extremely short CTIT to identify suspicious activity.

Click Hijacking

• Definition: Malware detects real clicks and sends fake clicks to manipulate attribution.
• Protection: Use raw data timestamps to identify very short time intervals between clicks.

DeviceID Reset Fraud

• Definition: Fraudsters reset their DeviceIDs between installs to bypass real-time fraud detection.
• Protection: Look for high install rates from new devices or those with Limited Ad Tracking and block known fraudulent SiteIDs.

Limit Ad Tracking Exploitation

• Fraudsters enable Limit Ad Tracking to hide install fraud and repeatedly install apps.

Fighting Fraud: Tools and Strategies

The Marketer's Fraud Toolkit

• Active Protection:

  • Real-time blocking of fraudulent traffic.
  • Blacklists of offending SiteIDs, DeviceIDs, and IP addresses.
  • Blocks very short CTIT.
  • Install & receipt validation.

• Validation:

  • Real-Time Filtering to ensure traffic meets targeting and fraud standards.

• Alerts:

  • Real-Time Notifications when certain fraud KPIs increase.

• Deep Reporting:

  • Real-Time Insights to understand fraud protection ROI and identify anomalies.

Pre-Campaign and Live Campaigns

• Pre-Campaign: Include enforceable fraud terms in your IO using tools like IO Builder: Fraud Appendix.
• Live Campaigns: Enable real-time rejected install postbacks and transparent reporting in fraud dashboards.

Post Campaign

• Reconciliation: Review fraud performance data with network partners during payment reconciliation.

Success Stories

• Challenge: A fast-growing e-commerce brand targeted by advanced fraudsters.

• Impact: Over 200 fraud sources blocked, $500k in average monthly savings from fraud.

• Challenge: A large Asian digital marketplace faced fraud in mobile growth and attribution.

• Impact: 24% savings from fraud protection, $100B in mobile fraud revenue.

• Challenge: A large bank noticed discrepancies between reported installs and actual performance.

• Impact: 22% lift in organic installs, 51% of registrations and 71% of revenue events were identified as fraud.

Total Protection Strategy

• PROTECT360 is a comprehensive solution that combines:
  • Scale: Large volume of data processing.
  • Machine Learning: Advanced algorithms for fraud detection.
  • AI: Intelligent systems for real-time and post-install fraud analysis.

Conclusion

• Mobile fraud is a major threat to app marketers and requires a layered approach for effective protection.
• The document advocates for real-time protection and deep reporting to combat both basic and advanced fraud types.
• AppsFlyer provides tools and strategies to ensure transparency, accuracy, and ROI optimization in mobile campaigns.