Retail, CPG and Logistics
Key industry best practices to prevent ecommerce fraud in 2023 and beyond
E-commerce has become a way of life in all urban and semi-urban areas. But in its wake has risen the dark side of online fraud. In 2023 alone, global e-commerce fraud losses are estimated to touch 48 billion dollars. The conveniences brought in by e-commerce are also extended to committing theft – anytime, anywhere, and anonymously. The ease of physical effort and evasion makes it an attractive opportunity for swindlers worldwide.
There are different kinds of e-commerce fraud. The most well-known is credit card fraud, where stolen card information is used to buy a product or service online. The ultimate victim is the merchant, who must refund the purchase. Affiliate fraud is earning commission with fake activity like dropping cookies to visitors' devices which results in a commission if the visitor later goes to the merchant's website and makes an unrelated purchase. In account takeovers, a user's account is hacked by sending a phishing email or using other means to make unauthorised purchases and steal confidential information. Triangulation fraud involves creating a fake storefront to get credit card details. The fraudsters dispatch the goods added to the fake store order to the customer by purchasing them through an authentic store. The scamsters then use the credit card information to make other purchases. Friendly fraud happens when goods are bought and received, but a false refund claim is then made to the merchant.
Fraud detection has traditionally relied on predefined rules that decide the conditions for a transaction decline. Rule-based systems are based on rigid assumptions and do not take in previous behaviour or all the factors making up a transaction context. Therefore, they often are inaccurate and raise false positives leading to reputational damage and customer churn. AI-enabled fraud prevention solutions use both supervised and unsupervised machine learning. Large volumes of data from multiple vendors consisting of millions of transactions are analysed, and the algorithm looks for patterns in the data. It can spot anomalies and suspicious behaviour without using pre-established rules.
AI applications analyse data to identify specific patterns and red flags in previous fraud attempts to calculate a risk score that will be applied to every transaction. Advanced solutions combine behavioural analytics and behavioural biometric data with transactional data because deeper insights are needed to mitigate these threats as they get further sophisticated. Behavioural analytics uses details like the user's website or app browsing speed, viewed items and view frequency, conversion path traversal, and credit card number copy or paste actions. Behavioural biometric data records individual traits like the user's username/password typing style, touch pressure on the phone, and swipe direction to create a biometric signature. This signature can prevent an account takeover by asking for additional authentication when patterns are distinct from previously observed. Many solutions will flag an alert if a new customer raises a high-value order for a popular product and chooses overnight delivery. Instead of directly rejecting a possibly legitimate order, the customer journey should be analysed to check for pre-transactional browsing to reveal the intent and to check if the behaviour is low-risk. Establishing behavioural analytic signatures can identify granular details of user patterns, anomalies, and fraudster workflows. This signature prevents the denial of authentic orders and the consequent loss of customer base.
Many systems combine AI and rule-based decisions. If there is a unique or one-time occurrence, rules might be the antidote, as there isn't sufficient data or time for the AI system to learn. Businesses within the same industry sector face similar kinds of fraud, so defining industry-specific models as a baseline helps have a functioning model immediately. Layers specific to the organisation can be added for optimum results. AI solutions can examine a range of data points for every transaction and compare against patterns and discrepancies derived from billions of transactions in minuscule time. The evaluation speed allows all AI fraud decisions to be made in real time. AI tools constantly scan for new emerging patterns and learn and update the risk score and associated red flags. The adaptiveness of AI makes it an ideal deterrent for evolving fraud schemes.
E-commerce organisations must collaborate with a stable and secure data network to configure their models based on the latest data pool. Most merchants might find the fraud prevention process overwhelming, so that a trusted third-party payment processor might be the optimal strategy. Organisations that leverage technology and data collaboration to ensure a smooth customer journey with minimal fraud loss will be at the forefront of e-commerce retail growth, expected to reach 8.1 trillion in sales by 2026.
*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.
