Fraud

Gaming an ML Algorithm to buy low quality Impressions

Picture of By Sven Hezel
By Sven Hezel
AdFraud
3-minute read
Article Topics

How Fraudsters Game CPM Algorithms Through Fake Conversions: The Missing Link Between Impression and Conversion Fraud

Impressions-Fraud

In the ever-evolving landscape of digital advertising, Cost Per Mille (CPM) remains a prevalent pricing model. Under this model, advertisers pay for every thousand impressions their ads receive. However, this model isn’t without its vulnerabilities. One of the key concerns has been the manipulation of the CPM bidding algorithm by fraudsters, particularly through generating manual conversions. Surprisingly, marketers often overlook the connection between impression fraud and conversion fraud, considering them separate entities. This article aims to shed light on how these two forms of fraud are intertwined and how fraudsters exploit this loophole to their advantage.

Understanding CPM Algorithms

Before diving into the tactics employed by fraudsters, it is important to understand how CPM algorithms work. These algorithms are designed to determine the quality of a site as a source of traffic. When an ad generates a higher number of conversions, the algorithm considers the site to be a good source of quality traffic, often leading to higher bids for ad placements on that site.

The Tactic: Manual Conversions

Here’s how fraudsters manipulate the system:

  1. Fake Impressions: They start by generating fake impressions to attract advertisers.
  2. Manual Conversions: Then, they manually create conversions associated with these impressions.
  3. Algorithmic Misdirection: These manual conversions are detected by the CPM algorithms, which then interpret the site as a quality source of traffic.
  4. Increased Bids: As a result, the algorithm bids higher for ad placements on these fraudulent sites.
  5. Increase Volume: Driving low quality traffic to the site and reselling it at a higher rate.

Why It Works

The deception lies in making the algorithm believe that the conversion is a result of a genuine human interaction. Most marketers use conversions as a key performance indicator (KPI), making it a perfect target for fraudsters to manipulate.

The Overlooked Connection: Impression and Conversion Fraud

Most marketers are vigilant about conversion fraud, but few understand its connection to impression fraud. They often view them as isolated problems to be solved independently. The reality, however, is far from it.

  1. Interdependency: Impression fraud sets the stage for conversion fraud. An inflated number of impressions makes a site appear attractive placement for advertising, which then becomes fertile ground for faking manual conversions.
  2. Mutual Amplification: Each form of fraud amplifies the effectiveness of the other. High impressions attract high-value ads, which when subjected to conversion fraud, attract even higher bids.

Countering the Fraud

Awareness is the first step toward counteraction. Marketers need to understand the relationship between impression and conversion fraud and adapt their detection mechanisms accordingly.

Multi-layered Fraud Detection: Opt for solutions that can identify and flag anomalies in both impressions and conversions.

Conclusion

As the digital advertising space evolves, so do the tactics employed by fraudsters. While it’s crucial to keep pace with these advancements, understanding the foundational relationship between different types of fraud can provide a robust base for building stronger, more effective detection algorithms. Therefore, marketers must consider impression and conversion fraud as interconnected issues that require an integrated approach for effective mitigation.

Marketers should not overlook the symbiotic relationship between impression and conversion fraud. Acknowledging this link can pave the way for more comprehensive fraud detection mechanisms and a safer advertising environment.

AdSecurity Platform Learn more
Share this Article
Facebook
Twitter
LinkedIn

Want to get the latest from 24metrics ?

Subscribe to our newsletter below to receive updates on features, filters, new types of fraud etc.