Synergy of data analytics & alternative data in fintech: Combating fraud and managing credit risk

By: Mohan Ramaswamy, Co-Founder & Chief Executive Officer, Rubix Data Sciences

Financial fraud is a growing concern in India, with the Reserve Bank of India (RBI) reporting a 48 percent increase in the number of fraud cases from 9,097 in FY22 to 13,530 in FY23. While the overall value of financial fraud has decreased by almost 50 percent from Rs 59,819 crore in FY22 to Rs 30,252 crore in FY23, the instances of financial fraud have increased. This indicates that fraudsters are becoming more active and innovative in finding new ways to cheat and deceive. 

In B2B settings, transactions often involve substantial sums of money, multiple stakeholders, and high stakes, hence fraud and credit risk are not new challenges for banks and other financial institutions or companies. However, the rapid growth of the fintech sector has also been marked by the quick evolution in the tools and methods of perpetrating fraud. 

Therefore, banks, financial institutions, and fintechs need to be vigilant to protect themselves from fraudulent activities while making informed decisions about extending credit to partners or customers. In this context, data analytics and alternative data can play a crucial role in combating fraud and managing credit risk in B2B settings.

Unleashing the power of data analytics

Traditional methods of fraud protection have been labour intensive and often slow, relying heavily on highly skilled personnel. However, modern fraud protection relies significantly on data analytics to identify and prevent fraud. Data analytics is multidisciplinary in nature; it harnesses a broad spectrum of data analysis techniques, including statistics, mathematics, and computing to generate insights from large data sets. The increase in computing power in recent times has made data analytics even more powerful. 

Modern fraud analytics systems can perform risk monitoring activities in seconds by sifting through vast datasets to identify patterns, anomalies, and potential threats in real time, enabling quick decision-making. According to the Association of Certified Fraud Examiners (ACFE), a typical fraud lasts 12 months before being detected and causes a median loss of $117,000 per case. Deploying fraud analytics tools drastically reduces detection time, thereby minimising losses, and safeguarding businesses. 

Data analytics is applied in fraud detection and prevention in B2B settings in the following ways:

• Pattern Recognition: Data analytics leverages historical transaction data, customer profiles, and industry benchmarks to identify patterns that may indicate fraudulent activity or credit risk. Machine learning algorithms can continuously interpret and learn from large datasets of customer profiles, making their identification of risky customers more accurate over time.  For example, data analytics models are able to alert fraud management teams if transactions are emanating from pin codes where there have been large instances of fraud. Such customers can go through a more detailed review by the Risk Management Team. 
• Real-time Monitoring: Advanced analytics tools enable real-time monitoring of transactions and credit applications. They can spot unusual patterns, such as unexpected spikes in transaction volumes, irregular timings, or changes in transaction size, that may indicate fraudulent activities. This means that any suspicious or anomalous behaviour can be flagged immediately, for a more detailed check.

• Credit Scoring Models: Credit scoring models, powered by data analytics, rapidly assign credit scores to B2B clients based on structured and unstructured data elements gathered from hundreds of sources. Some of the data used in credit scoring include identity data, financial statement data, transaction data, payment history, statutory compliance data, court data, management and ownership data, employee and customer feedback data, etc. These credit scores help banks, financial institutions, fintechs, and companies make informed decisions about whether or not to extend credit, set credit limits, and adjust payment terms.

In addition, fraud analytics tools can analyse vast amounts of Know Your Customer (KYC) and payment transaction data, helping banks, NBFCs, fintechs, and insurance companies in Anti-Money Laundering (AML) activities and prevent transactions with sanctioned entities in India and overseas.  

Alternative data: A game changer

While traditional sources of data, such as financial statements and credit reports, are valuable, they may not provide a complete picture, especially for emerging or smaller businesses. This is where alternative data plays an important role. 

• Social media and online presence: Alternative data sources include social media activity, online reviews, and a company’s digital footprint. Analysing these sources can provide insights into a company’s reputation, customer and employee sentiment, and overall health. For example, if a financial institution integrates alternative data sources, including social media activity and news mentions into its credit scoring model, it will be able to extend credit to small businesses with limited financial histories while still accurately assessing credit risk. This will help it expand its customer base and grow its lending portfolio. Similarly, many fintechs are using gamification to collect and analyse psychometric data about potential borrowers before lending to them.
• Sentiment analysis tools: These are applied in human languages to identify, score, and analyse different levels of emotion. For example, when applied to company earning statements or CFO calls with analysts, these tools help assess whether the company’s underlying sentiment is optimistic, pessimistic, or neutral. 
• Geospatial data: Location-based data can be used to assess the physical presence and stability of a business. Unusual or frequent changes in the location or address can be red flags. 
• Night-time-light data: Various satellite companies now offer Night-time-light data, which can be used to estimate the extent of the socio-economic activity in a region. Lenders have begun analysing this unstructured data to evaluate whether it makes sense to expand their presence in a particular region or area.

The road ahead

The landscape of fraud analytics and alternative data is continually evolving. As businesses accumulate and analyse vast datasets, the potential for innovation in fraud prevention and credit risk management remains limitless. By embracing these technologies, banks, fintechs, and companies can make more informed decisions, safeguard their assets, and thrive in an increasingly complex and interconnected world. However, it is essential to use data ethically and in compliance with data privacy regulations.

Data analytics and alternative data have emerged as indispensable tools for combatting fraud and managing credit risk in B2B settings. The future of B2B risk management belongs to those who embrace the data revolution, paving the way for enhanced financial security and growth.