Collusion Fraud: The Art of Gaming the System with Complicity

How platform business models are at an increased risk of fraud when two or more separate parties collude.
By Shahnawaz Backer (additional contributions by Yiing Chau Mak)
April 14, 2021
4 min. read

Companies like Uber, Airbnb, PayPal, and others with platform business models have flourished in the past few years by matching up service providers (such as restaurants and drivers) to consumers and hiding the complex, behind-the-scenes processing (like payments) from users The rapid adoption of this business model has brought it into the cross hairs for fraudsters, who are always scheming to game the system and illegally monetize legitimate business processes. F5 Labs has found that attackers are defrauding digital systems by colluding with other participants who serve different roles on the platform. This article discusses this phenomenon, which we call collusion fraud.

What Is Collusion Fraud?

Collusion fraud occurs when two or more participants conspire to defraud another participant in a digital business transaction that involves multiple participant groups. This type of fraud is growing in prominence as more digital businesses pivot to become platforms that serve more than just one purpose. For example, an online e-commerce provider’s digital platform allows a consumer to select items from a seller of their choice and have it delivered. A single business transaction on that platform provides online processing, payment, preparing goods, logistics, and delivery. Completing these activities require services from multiple providers specializing in different areas, which at times takes the processing out the platform’s control. The collaborative act to complete these multistep business transactions provides an avenue for malicious players. Fraudsters design these hacks so they can quickly make money and target returns that are generated as by-products of the main transaction, such as a cashback reward or gratuity. These by-products are usually managed separately from the main transaction and are often hard to reclaim post–fraud detection if the consumer or other participant has already used them.

Collusion Fraud in Action

Collusion fraud can happen in any industry vertical. F5 Labs and Shape Security researchers followed two cases of fraud that revealed collusion in action. Fraudsters made gains in these cases in the form of gratuities and cashback rewards points.

Case One: Leading Food and Beverage Company

The first case involved a leading food and beverage (F&B) company in which collusion fraud manifested as gratuity, or tip, abuse. The company’s digital platform provides a convenient service to its customers by bringing together the restaurant outlet, logistics requirements, and online payments. Figure 1 explains the legitimate process in completing an online transaction that includes a tip.

A box-and-arrow diagram, titled "Legitimate Order Flow on Food & Beverage Platform". The steps in order read, "1. Consumer places an order on F&B online platform selecting the outlet of choice (Say Outlet X) using a credit card including a tip for driver 2. F&B platforms passes the order to Outlet X and delivery request to the logistics service partner 3. Food outlet prepares the order and the platform updates the logistics service provider. 4. Logistics service provider assigns the delivery person. 5. Food is delivered to the consumer, completing the delivery person's order 6. Delivery person claims the gratuity from the logistic provider 7. Logistic provider pays the delivery person and charges the tip amount from the F&B platform 8. F&B platform settles the costs with the food outlet and logistics service provider".
Figure 1. Order flow on a food and beverage provider platform.

In this specific case, the consumer-fraudster and a delivery person collude to monetize stolen credit cards. They achieve the collusion fraud through the following steps:

  1. Using a stolen card, a fraudster places an expensive order (more than $300) that includes a generous tip (usually more than 30 percent).
  2. The order goes through the standard, legitimate lifecycle, as described in Figure 1.
  3. The credit card owner detects the transactions and disputes the charges with the bank. This leads to a charge-back on the F&B platform for the full amount of the order, including the tip.
  4. The tip amount that was paid out for the delivery service cannot be recalled.

Shape’s security data reveals that in a period of three months this particular F&B online platform received almost 3,000 collusion fraud orders with a cumulative value of USD $1.5 million and gratuity/tip amounting to about USD $350,000.

Case Two: Leading Online Payment Wallet

The second case involved a leading online payment wallet that suffered in the form of cashback rewards due to collusion. Figure 2 documents the legitimate flow of a transaction initiated by a consumer.

A box-and-arrow diagram, titled "Consumer Using Reward To Settle Purchase". The steps in order read, "1. Consumer places an order with Merchant X's online platform and chooses to pay with the online payment wallet 2. Merchant X initiates the transaction with payment wallet platform 3. Payment wallet platform charges the user, transfers payment to the merchant, and rewards user with a percentage of money spent (cashback) 4. Consumer places an order on Merchant B's online platform and uses the cashback reward from the online wallet 5. Payment wallet settles the transaction with cashback reward".
Figure 2. Consumer flow for earning and spending cashback rewards.

In this case, the consumer and Merchant X conspire to defraud the payment wallet platform of rewards points in the following manner:

  1. Consumer purchases goods from Merchant X using a payment wallet platform.
  2. As shown in Figure 2, the consumer earns rewards points, which are then used to purchase goods from Merchant B. Once the cash back reward is consumed, Merchant X refunds the original sum to the user, citing reasons such as unavailability of stock.
  3. The payment wallet refunds the original sum to the user, but the cashback rewards are not recoverable.


As the adoption of digital services and platform businesses grows, consumers will be enticed by various incentives beyond cashback rewards. Fraudsters will find a way to collude to steal these incentives, resulting in greater varieties of collusion fraud.


Detecting collusion is difficult and will require artificial intelligence to weed out such transactions at scale. F5 recommends the following security controls:

  • Use artificial Intelligence-powered analytical models, clustering groups and transaction to detect collusion fraud.
  • Train and retrain the AI models as fraud techniques evolve. 
Join the Discussion
Authors & Contributors
Shahnawaz Backer (Author)
Yiing Chau Mak (Contributor)
Director of Data Science, Shape

Read More from F5 Labs

2023 Identity Threat Report: The Unpatchables
Top Risks
2023 Identity Threat Report: The Unpatchables
11/01/2023 report 80 min. read
Sensor Intel Series: Top CVEs in March 2024
Top Risks
Sensor Intel Series: Top CVEs in March 2024
04/30/2024 article 7 min. read
2024 Bad Bots Review
Bots and Automated Attacks
2024 Bad Bots Review
03/14/2024 article 15 min. read