FMetrics Scorecard To Be Data-Independent

Fractal Analytics is into preparing scorecards of various models primarily for financial services companies and banks to focus on customer behaviour analysis. The scorecard enables companies to predict consumer behaviour helping them to be more responsive to introduce targeted products.

Recently the company introduced FMetrics Scorecard to predict the eligibility of a loan applicant. Srikanth Velamakkani, CEO, Fractal Analytics speaks with Biztech2 about the new scorecard and how scorecards can radically change the business dynamics of companies.

Can you overview the basic advantages of scorecards?

We started building scorecards around the risk side for a host of customer domains. With Algorithms as platform, we figure out the customer’s past history of credit card transactions, Internet banking, whether he has availed of any loans in the past. This is enough information to select the right candidate. Also scorecards are an advantage to chalk out the prime candidates for the various telecalling loan campaigns run by banks and even the customised loan offers among the final list i.e to offer a lower interest rate to a specific lot of customers or to give a free gift.

How is the newly introduced FMetrics Scorecard different from the conventional scorecards?

The statistical scorecards operate on the past information of the customers to decide about the customer behaviour or eligibility. FMetrics is data-independent. It is basically targeted towards new banks, banks with new products and banks that do not have enough data to get started. Banks can deploy FMetrics for home loans, personal loans, auto loans and credit cards. It even works in absence of credit bureau information.

So practically, how does FMetrics works after a customer walks in a bank and submits the loan application?

Firstly, information filled in by the customer goes through the bank’s policy filter and is only worked upon if it passes the same. Post that, all the information is compared with the various parameters to end up on a score decided by the bank.

For e.g. If the bank’s cut-off score is 600, then anything above that is accepted and the loan disbursal is executed in two days. There is a lower cut off score say 425. If the customer’s cut-off score is between 425 and 600 then it will be subject to the credit officer and contact point verification. Any score below that will be rejected right away. This process ends up in about fifteen seconds.

How will scorecards help in fulfilling the mandatory requirements of BASEL-II?

BASEL-II norms say that the banks need a PD (Probability of Default) model for an internal ratings approach for the set of portfolios and customers. The customer is approved or rejected but what is the probability of default for a bank. The custom scorecards can immediately give the PD, which makes it BASEL-II compliant. For banks introducing new products, banks with new branches, the scorecard does not give the PD right away but the banks can impute the PD from the past records in congression with the statistical scorecard.