How Analytics is helping Credit Managers become Finance Leaders
Credit managers have always played a critical role in making sure that the right balance is maintained between risk aversion and profit maximization. As the digitization of financial services continues, they can leverage that experience to take the next step change – from day-to-day, tactical support to long term, strategic impact. Combining the latest technologies with their deep business expertise – they can examine their businesses to answer questions such as “where can we grow”, “how can we reduce lending risk” and “how can we tap into market segments which we have previously avoided”.
The proliferation of analytics provides tremendous opportunities to grow business. As per Accenture, more than a third of the world’s adult population make little or no use of formal financial service, and serving this huge market of unbanked and underbanked could generate about $380 billion in new revenues. To seize this opportunity, credit managers need to use analytics extensively. Analytics can help where traditional scoring models do not work. By collecting data from borrowers’ smartphones on their utility payment patterns, contacts quality, phone usage patterns, messages on banking and non-banking transactions, etc., along with the psychometric data collected through questionnaires to gauge borrower’s characteristics, like honesty, ethics, etc., predictive algorithms can determine the intent to pay, facilitating them to offer loans to even those individuals who do not have a credit history. To begin with, these predictive models may be less effective, but eventually, as machine learning algorithms improve, the decisioning will get more and more accurate. This would enable them to confidently automate the processes resulting into lower costs, extended reach, and enhanced customer experience.
Even in situations where the loan applicants have been assigned credit scores by the credit bureau, there is often the question of accuracy involved. In fact, in the US, it was found that more than one in five consumers have a ‘potentially material error’ in their credit file that makes them look riskier than they are. Predictive analytics can enable credit managers to reduce the lending risk by making data driven decisions. Using statistics and machine learning techniques, they can analyze the data available from various sources to create credit scoring models on their own, specific to their business. These models incorporate financial and non-financial data such as demographic and profile information to do forward-looking analysis of the probability of default for a borrower over various timeframes, and calculate the potential expected loss in case of default. In-house analytics can empower them to generate insights that are transparent (for inputs, methodologies, and assumptions) and flexible (in changing assumptions and input parameters, which can be updated frequently).
In fact, Predictive analytics can be utilized to improve the customer experience throughout the loan lifecycle. Marketing departments can benefit through improved targeting in their campaigns, and credit risk departments can create scorecards to make more informed decisions on whether or not to accept an application. Opportunities for cross-sell and upsell can also be identified by analyzing the behavior of existing customers and by assessing the risk of default, proactive actions can be taken to mitigate the risk early. Collection analytics can predict the likelihood of delinquent customers paying back the debt and the right channel to reach out to these customers. This would not only help in increasing the interest revenue, but also in reducing the collection costs.
Although analytics based credit scoring systems are rapidly being implemented and used by many FIs nowadays, these systems have their own challenges. They are either not comprehensive in their offerings or not user friendly thus not enabling business users to use the system on their own without much dependency on others. A predictive analytics scoring system should be able to build models using different types of algorithms using a variety of data coming from various sources including social media, utility and mobile data. Using techniques like decision trees and strategy maps, it should provide answers to questions such as the potential segments and the optimum strategy of managing customers. The real benefit of a scorecard is only evident on implementation, so the system should provide hassle-free deployment, and also APIs to enable the handling of remote requests for model scoring in real-time. Last but not the least, as every model degrades over time due to reasons like economic drift, automated model monitoring capabilities are a must in order to guide credit managers to tweak the models whenever required and neutralize the impact of these changes on business.
Forward-looking credit mangers are evolving themselves today, to set the expectations of tomorrow where they want to be looked upon as finance leaders. With the help of predictive analytics, these managers have started getting involved into organization’s strategic endeavors and are taking broader set of responsibilities which could include defining the organization’s overall vision.