Credit scoring in the age of big data
• Providers are using alternate and trended data to help move the sector forward, writes Pedro van Gaalen
Technically speaking, the credit bureau business model was built on big data, long before it became a technology buzz word. What has changed since then is how these huge data sets are analysed to make sense of the reams of structured and unstructured data that credit bureaus have access to today.
In this regard, advanced machine learning algorithms and predictive analytics capabilities are being applied to big data sets to deliver a consolidated yet granular view of credit-active consumers.
“Given the fragmentation of financial services in SA and the ever-increasing number of potential data points, credit bureaus play a vital role in enabling the infrastructure and operation of financial intermediaries,” says Frank Lenisa, director at Compuscan.
“After collecting, ‘sanitising’ and vetting consumer data, bureaus deliver unique insights that can predict credit risk and determine the creditworthiness of individuals by overlaying this data with verification services and decision sciences that analyse thousands of predictive variables and credit scores.”
However, credit bureaus have moved beyond traditional credit scoring models, as these only provide a limited view of a consumer at a specific point in time, says Lee Naik, CEO of TransUnion Africa.
“By utilising alternate and trended data to gain a more holistic view of a consumer and identify trends in their credit behaviour, lenders can tailor their products to consumer performance, which builds better loan relationships and delivers better long-term value.”
Alternate and trended data can include information from social media, digital footprints, psychometric data, deeds and titles, qualifications, rent repayments and information collected via surveys and questionnaires, among others.
And by expanding the breadth and depth of the data being analysed, Naik suggests that the application of this technology extends beyond merely predicting credit risk among credit-active consumers. “New credit scoring modelling that utilises alternative and trended data can help SA’s financially marginalised population gain access to credit.”
A recent TransUnion study identified 3-million South African consumers who previously could not gain access to credit using traditional scoring models.
“Our colleagues in Colombia identified an additional 4.7million consumers (13.1% of the credit-active population) by using a recently developed TransUnion credit scoring model. We now have the opportunity to unlock this type of potential locally, and drive financial inclusion and growth in the South African financial services sector,” says Naik.
Says Lenisa: “Having distilled this information, bureaus then package it and make it commercially available to banks, specialist lenders, credit providers and retailers via digital solutions that offer integration via fit-for-purpose front-end interface. This enables these credit providers to make real-time decisions on credit extension.”
This functionality has helped to drive market innovation and disruption in the financial services market, particularly within the same-day or payday micro-loans sector.
Simon Russell, MD of Experian South Africa, adds that companies can also benefit greatly from accessing this type of data and insights via digitalised solutions, as they can also quickly negotiate informed credit terms, set appropriate credit extensions to their best customers, and renegotiate credit terms with riskier clients, if required.
“Data is also important for making informed decisions, and is the starting point for developing a business’ modernisation strategy. Proper usage helps to identify risks, potentially problematic customers and to deliver appropriate credit decisions. Furthermore, combined data sets, predictive analytics and best-of-breed processing technology enables faster decision making, allowing businesses to meet customer demand while still ensuring due diligence and the effective allocation of resources.”
He adds that gaining access to a complete view of its customers through the adoption of efficient digital credit rating solutions, a business can also create a long-term credit strategy based on anticipated future behaviour, and develop long-term credit strategies to provide a better customer experience and offer customers more favourable terms.
“These digital solutions can speed up the credit approval process, allowing credit managers to make immediate and accurate credit decisions to expedite application processes. This results in more applications being handled, which allows businesses to grow quickly, while creating a sustainable and profitable digitalised business for the future.”