The evolution of fraud and its detection in the retail industry
According to a recent report, e-commerce retailers faced 350% increase in fraudulent orders during the holiday season. Another reputed research estimates that merchants will lose $206 billion to payment fraud between 2021 and 2025. These statistics clearly call for an analysis of the types of fraud in the market and intelligent systems that organisations must adopt for retail fraud detection and mitigation.*
Types of fraud in the retail and eCommerce industries
Scams in the retail sector are constantly evolving, which requires businesses to engage in robust data processing and engineering steps to compare a customer’s past buying patterns with a possible fraudulent transaction. Let’s look at some of the common types of retail fraud:
- Transaction fraud:
The fraudster makes an online transaction with stolen credit card details. The card owner notices the transaction, notifies the bank, and requests a chargeback. But this money goes from the pocket of the retailer. There is also a hefty admin fee for the card network.
- Triangulation fraud:
- The customer makes a purchase in an online marketplace.
- The seller is a fraudster who receives the request and places an order in a legitimate online store.
- The fraudster uses a stolen credit card and gives you the buyer’s address for delivery.
- The online store delivers the item to the buyer.
- The owner of the credit card notices the transaction and requests a chargeback.
- The store has to pay the chargeback amount whereas the fraudster keeps the original customer’s money.
- Friendly fraud:
- Return fraud:
- Chargeback guarantee fraud:
It involves a genuine buyer, an online store, and a fake online store run by a fraudster:
This happens when a customer whose card was not stolen but still claims a refund. It could be because of genuine ignorance where the customer is unaware of the transaction or an opportunistic friendly fraud where the customer wants to return the product just because of remorse, disagreement with the policy, or wardrobing (purchasing a piece of cloth to wear for an event before returning it). Lastly, it could be a purchase where the customer refuses to take the delivery of the item but claims a refund.
Customers use loopholes in the online store’s return policies in combination with transaction fraud. One such example is purchasing an item from a store and returning it opened with the intent to re-purchase it at a lower price under the open box policy.
Some e-commerce fraud detection solutions offer to pay the admin fee in case you continue to get chargebacks even after blocking friendly frauds. This causes the solution provider to be extra careful and sometimes even block genuine customers.
Challenges in detecting fraud
Retail and e-commerce is a fast-moving industry with a large number of transactions on a daily basis. Detecting and isolating frauds is like finding a needle in a haystack. But you have to be fast and accurate before such cases cause severe damage. Some of the challenges companies face with fraud detection in the retail industry are:
- Large number of attributes:
- False positive rate (FPR):
- Rule-based decision-making and data processing:
- Constantly changing ways of fraudsters:
Fraud detection involves pre-processing and quantization of data that generates a large number of attributes. These attributes create a customer’s profile based on past buying patterns and events. A successful fraud detection model captures these attributes and works in multi-dimensional space to find fraudulent transactions.
False positive refers to an alarm raised by mistake after a genuine transaction. This could lead to an upset or an insulted customer as well as higher costs involved in resolving the issue. On the other hand, no alarm raised on an actual fraud could lead to a disaster.
Current fraud detection systems in different industries use a rule-based approach. This includes several steps of data pre-processing and feature engineering.
Fraudsters find newer ways to hoodwink companies. The system may not have sufficient historical data to detect these new frauds or abuse patterns.
Solutions to detect fraud
An effective e-commerce fraud solution finds suspicious patterns in customer data. But since fraudsters evolve, your solution needs to quickly adapt and circumvent the risk rules. It is important to anticipate new attack vectors proactively through methods such as:
- Social media lookup:
- Data enrichment:
- Device fingerprinting:
Although fraudsters can obtain data of various credit cards, what they can’t do is create legitimate social media profiles that match the names and details of every stolen credit card. Therefore, this is a great way to conduct background checks on social media to see if the user is genuine or not.
Build a complete profile of your customers based on a single data point, such as a phone number or an email address. Then, you can verify the email address and check if it has been opened with a temporary domain service.
This technique scans customers connecting to your site and spots suspicious logins via proxies, emulators, or VPNs. It also helps connect the dots between various suspicious accounts and helps flagging the entire network at once.
Other methods include the encoder-decoder structure, machine learning, generative adversarial network, and more.
How can Infosys BPM help?
Infosys BPM deploys fraud management solutions for the retail and e-commerce industry. Our team has the domain expertise, analytical skills, and technology to help you prevent frauds, protect the revenue and brand image, and increase customers’ trust. Our services include e-commerce fraud management, PoS fraud management, and product counterfeit management.
*For organisations on the digital transformation journey, agility is key in responding to a rapidly changing technology and business landscape. Now more than ever, it is crucial to deliver and exceed on organisational expectations with a robust digital mindset backed by innovation. Enabling businesses to sense, learn, respond, and evolve like a living organism, will be imperative for business excellence going forward. A comprehensive, yet modular suite of services is doing exactly that. Equipping organisations with intuitive decision-making automatically at scale, actionable insights based on real-time solutions, anytime/anywhere experience, and in-depth data visibility across functions leading to hyper-productivity, Live Enterprise is building connected organisations that are innovating collaboratively for the future.