Retail, CPG and Logistics

Types of ecommerce fraud and how to detect and prevent them in 2023

The ease, convenience, and timesaving in shopping with a few mouse clicks and having the items delivered to the doorstep has ensured the steady growth of e-commerce. But there's always a catch. And in this case, it's been an equally steady increase in e-commerce fraud. According to a Juniper Research report, e-commerce fraud loss in 2022 amounted to 41.4 billion dollars and is expected to rise further.

E-commerce fraudsters keep changing their modus operandi, making preventing and detecting fraud increasingly difficult. Let's examine the most common e-commerce fraud types and how e-commerce organisations can prevent and detect occurrences.

Transaction/Credit Card: The swindler obtains details of a credit card and uses it to make an online payment. When the payment goes through, the actual cardholder reports this unauthorised transaction to the bank, and chargeback processing is started to regain the money. The e-commerce vendor must refund the amount and shell out an admin fee to the card company.

Chargeback/Friendly: Here, the cardholder themselves are the cheats. They order and receive the item but ask for a chargeback claiming that the delivery wasn't made. This type of fraud is termed friendly as genuine customers make it.

Account Takeover:  An account takeover (ATO) happens when fraudsters access customers' account details and login with malicious intent. They can access payment information which can then be used for transaction fraud, get other sensitive data, and modify contact details so that the account holder is not notified of purchases which they then get delivered to their addresses.

Triangulation: This fraud type has a bona fide online store and customer, a fake online shopping site, or a marketplace vendor page run by the swindler who has stolen card details. A genuine customer makes an online purchase with the phony seller, who then buys the item from the authentic online store using stolen card details and has it shipped to the genuine customer. The cardholder (of the stolen credit card) initiates a chargeback against the fraudulent transaction. The chargeback, the admin fee to the card company, and the loss of product must be borne by the bona fide online store. The customer who bought the product will have no idea of the fraud going on with the use of their transaction.

Affiliate: The loyalty partners of the e-commerce store send spam users to the website to increase their commission earnings. They create fake user accounts or generate many bogus clicks on affiliate links to inflate traffic statistics.

Return: The customer misuses the return option by returning used or damaged products or empty boxes and then claims a refund. Such activity is policy abuse, but the loss to the retailer is the same as fraud.

Fraud detection and prevention is a tightrope walk between dealing with emerging threats and risks while delivering an undisrupted customer experience. AI technologies like deep learning analyse large volumes of transaction data to detect behavioural patterns that indicate fraud. These intelligent solutions improve accuracy over time. Traditionally, there were pre-defined, rigid rules which raised a significant number of false positives while being ineffective over time with newer methods of fraud.

AI-based anomaly detection is used to track card fraud and considers behaviour of the majority as well as the individual user. If a generally risky action is in accordance with the regular pattern of a particular user, then it isn't tagged as possible fraud, thereby minimising false positives. If any behaviour is marked as fraudulent, the account or transaction is suspended based on the business policy and subject to further investigation. Anomaly detection usually uses a supervised learning model trained using previously classified historical data. AI systems can also identify emerging threats through unsupervised learning. Here the system discovers new patterns and data interrelations that can be classified as suspicious, and it differentiates between various types of deceptive actions.

In the case of ATOs, the AI system detects atypical behaviour that does not match historical patterns for the user and flags the account. The AI system identifies the patterns associated with chargeback transactions and differentiates between valid and invalid activity. It blocks or suspends transactions where the user behaviour matches that of fraudulent chargebacks. AI algorithms can pinpoint fake ids by tracking the number of user ids linked to an IP address. Spam traffic is detected by checking for deviation from regular traffic sources and website visitor behaviour. AI solutions identify and blacklist affiliate fraudsters by detecting fake ids and spam traffic.

According to the Juniper Research report, cumulative losses due to online payment fraud between 2023 and 2027 will be more than 350 billion dollars. An optimal combination of verification solutions at the necessary points in the customer journey must be employed to contain online fraud. Additionally, AI solutions trained on a wide range of data will prove more effective in prevention and detection than an on-premises solution trained on data limited to just a single vendor.

* For organizations 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 organizational 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 organizations 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 organizations that are innovating collaboratively for the future.

Recent Posts