The global online retail market was valued at USD 4.25 Billion and is projected to expand at a compounded annual growth rate (CAGR) of 9.4% from 2020-2027. Online retail companies work on the business model of selling goods or services over the internet. It includes business-to-business (B2B) and business-to-customer (B2C) segment. These companies tailor their business model to capture internet sales by creating websites, applications, promoting the product on social media, building distribution centers, shipping centers, and warehouses.

The growing availability of different product varieties, price comparison, and growing usage of smartphones are some of the factors driving the growth for online retail companies. In addition, increased internet penetration especially in remote areas has revolutionized the online retail industries, allowing the business to reach the last mile customer. Some of the major players operating in the global market include Amazon.com, eBay, Alibaba, Flipkart, Ikea, and Warby Parker. Among these, Amazon.com is the market leader in terms of market share and customer base.

The graph below indicates the US E-Commerce market. The market has seen a spike during the COVID-19 pandemic as most of the users are shopping from home due to lockdown conditions in different countries. The second quarter of the year 2020 has seen a rise of 16.1% in online shopping compared to the previous quarter.

Figure 1: Online Retail Industry Analysis during Covid-19 (Accessible in PDF Version)   

Problems Surrounding Internet Fraudulence

One of the major restrainers in the growing online retail market is increased internet frauds. It has impacted the growth of online retail companies negatively. Due to increased internet frauds, companies are not only facing financial losses. They are losing customer’s trust, which has resulted in increasing customer churn rate and reducing customer loyalty. Fraud is an intentionally false representation of a fact, which is done to deceive someone to obtain profit. The profit can be in the form of money, sensitive information, or goods. Some of the common frauds affecting the online retail industries are:

1. Credit Card Frauds: Credit card is the most common type of fraud. This is generally done from a stolen card of a legitimate customer to make an online purchase. There are some red flags used to identify such frauds such as:

• Different credit/debit card, same shipping/billing address
• Same card different address
• The amount of purchase is large compared to previous purchases
• A large quantity of the same product
• Increase in the frequency of purchase

This type of fraud not only affects the customers but also the merchants, as the customer might place a refund request for the order he/she have not purchased.

2. Phishing: In a phishing fraud, information of a legitimate customer such as credit card number, mobile number, address, and ATM Pins is stolen via email, phone, or any other type of social engineering methods by the fraudster for any financial or personal gain. The user generally gets a link via email or phone, which contains malware, the moment the user clicks on the link; malware is installed in the system and captures all the information, which is then used by the fraudster for personal benefits.

3. Fake Accounts: Fraudsters create a fake account of a legitimate user and start purchasing. The information present in the account is real, hence making it difficult for the merchant to detect the fake accounts.

4. Triangulation Fraud: In this type of fraud, the fraudster creates a fake online shopping website offering products at a cheaper price compared to other online stores. This is usually done to gather the payment-related information; once the fraudster gathers all the details, he orders the same product from the original website and sends that to the customer. 

Resolving Online Frauds Using Latest Technologies

Machine Learning: Online retail companies can detect credit card fraud by leveraging machine-learning algorithms. The machine-learning model created is trained using the backlog data and then deployed in the transaction log of the ERP and CRM systems maintained by the online retail companies to detect the fraudulent activity in advance. The model depicted below has an accuracy of 98.9% of detecting frauds. It is fast enough to detect anomalies and classify them as a fraudulent transaction.

Figure 2: Credit Card Fraud Detection System (Accessible in PDF Version)  

AWS GaurdDuty: AWS GuardDuty is another very well known application of AWS, designed to continuously monitor security and detect any threats. It analyzes billions of transactions and detects any variation from the normal threshold. It can be deployed on the AWS infrastructure, without affecting the normal mode of production and performance. It not only detects the frauds in advance but also provides recommendations on preventing the frauds from happening in the future. It is easy to use, highly scalable, low-cost model for fraud detection.

SIEM Tools: These tools provide a holistic view of the information security system of an organization. These tools can be deployed in two types – SIM – Security Information Management, which analysis the collected data from transaction logs to detect any discrepancies and SEM (Security Event Management), which analyzes real-time data and informs network administrators of any suspicious activities.

Developing a System of Online Responsibility

1. Keep the financial data separate: Building a separate database for online transactions and payments would help in reducing fraud. Using the principle of least privilege, only the authorized users in the company can access the financial database preventing the chances of human errors.

2. Customer Awareness Programs: Customer awareness programs can be run in the form of ads on the websites. Besides, emails can be sent to all the registered customers making them aware of the common types of fraud. Customers can also be informed both via email and message regarding the type of order they have purchased, data, and amount of their purchase. This model can also be deployed within the organization to aware employees of data breaches.

3. Secure the online website with HTTPS: HTTPS is a secured internet protocol deployed to prevent fraudsters from accessing the websites.

4. Two-Factor Authentication: Multi-factor authentication is used to grant access to the payment website only after authorizing the user 2-3 times. The user is first authorized via OTP on email, then again with different OTP via phone, and at last via text message.

Future of Online Payment System

Online Payment Ecosystems are seamless digital payment systems where transactions can be done through virtual and mobile channels. Online payment ecosystems have reduced the cost of rendering financial services. Using these platforms, users can not only transfer funds but also pay for services such as utility bills, cab payments, insurance premiums, DTH, flight/train fare, buy goods from online retail stores, and movie tickets. The use of online payment systems has increased to 64% globally in the last 5 years. Some of the factors driving the growth of online payment systems are Government initiatives for cashless or low cash economy, increasing digital literacy, user-friendly APIs, 24/7 customer service, and interoperability of digital banks on a single application.

Advantages of online payment systems are listed below:

• Reduction in time of transfer
• Increased digital literacy among users
• Improved monitoring of transactions using transaction history
• Improved financial inclusion, especially in the remote areas
• Rewards and cashback offer to the customer using online payment

Figure 3: Business Strategy for Preventing Frauds (Accessible in PDF Version) 

Way Forward

Online retail stores are giving stiff competition to traditional stores. In the last 10 years, these stores have captured a larger market share. The online payment system has proved an increase for online retail industries, easing the payment collection and invoice generation process. However, cybercrimes have increased immensely as most of the transactions are happening online. Companies are leveraging machine learning and analytics to detect frauds and help financial institutions to assess risks in advance.

Steps taken to improve the security of these platforms include KYC, which helps in easily tracing the frauds. Tokenization is another method applied to reduce the risk of fraud. These are some of the precautionary measures if adopted properly would help online retailers to prevent frauds and improve customer experience.