Head of Risk Management Division, Erste Bank Croatia
Interview with Dejan Donev, Head of Risk Management Division, Erste Bank Croatia
Pre-approval loan management system to amaze and engage mid-tier customers
The ambition and the success of a project whose ultimate goals are to extend the number of pre-approved loans and enrich the loan portfolio throughout extensive use of analytics
Quickness, or the ability to give the customers what they need and want in nearly real-time, is a key factor to success. In order to promptly satisfy customers’ requests while reducing risk and optimizing the individual offers, Erste Bank Croatia has developed an automatic loan pre-approval system. Dejan Donev, Head of Risk Management Division, explains the latest about the project.
What’s the evolution of the Loan Approval process and Limit Management System in Erste Bank Croatia?
It took several years of data preparation, modelling and analysis to develop and roll-out the current automatic loan approval system. The following step was the development of the limit management system for pre-approved offers, including data, statistical analysis and model building. The loan pre-approval process lets the bank evaluate and define in advance customers optimum risk profile thus accelerating the response time when the bank’s client asks for a loan. Currently our customers can execute a pre-approved loan on our Internet banking platform within minutes and in 3 clicks only and see the money on their account straight away.
Can you explain your customer acquisition strategy?
When we introduced the limit management system for the pre-approved offers, the process involved only 30% of our 600.000 customers (the oldest and most trusted top clients) and one product of our portfolio (the cash loans). Starting in February 2016, the latest ongoing project (CLO implementation) aims at extending the pre-approved loan execution to three additional products (credit cards, overdrafts, revolving products) and reaching 60 to 70% of our customer base, all on our internet banking platform (the mobile part will be completed in the first quarter 2017). The goal is to be more attractive for the mid-tier customers that have some kind of relationship with us but not very strong. Throughout efficiency of the service and optimization of individual offers, we want to get them to consider us as their primary bank.
Analytics to target non primary solvent customers with ad hoc offers?
To estimate the risk and the size of a loan, it’s critical to know client’s incomes, but non primary customers don’t credit their salary on their Erste Bank account. So we needed to analyze all the transactions (more than 1 billion rows of data) the bank clients have done in the past three years and a half and match such data with information from the loan requests provided to us when applying for a loan. And build models for income prediction.The model has a very high level of accuracy. What it’s crucial is not to overestimate income. We estimated salary, risk, indebtedness in other banks, behavior and what product is more likely to interest the “customer” in order to customize the product and propose it throughout the channel the customer is more confident with. We have worked with SAS and our partner Assisto Consulting to develop these models. We learned how to implement and replicate it, so now we are independent: we can add or change parameters according to the bank’s goal.
What are the results and business benefits you expect from this project?
The main reason of this project was the return of investment calculated at the beginning. We gave a set of data to SAS and Assisto, and they calculated exactly the number of pre-approved loans expected by the end of the project (December 2016) and how much money we could earn from them. We had already 110.000 pre-approved loas and the business case promised to double the number, to have 4 products instead of one and return on investment in two years. Previously, three banks in Slovakia implemented the same project showing positive results. We presented a very concrete business case and the board gave the approval for the project in less than one month. After about 11 months, we overcome the expectations by around 30% achieving a total number of 230.000-240.000 pre-approved offers.
The evolution of a successful project: the 3 main phases
Phase 1: automatic loan approval process for on-demand loan with feedback in 10 minutes and time to cash in 20 minutes vs 2-3 days
Phase 2: loan pre-approval process to allow customers applying for loan to have answer in advance or in 1 minute plus 10 minutes for time to cash, online or in branch.
Phase 3: move from 30% of customers base pre-approved to 60%-70% and extend the bank offer from
one product of our portfolio (the cash loans) to three additional products (credit cards, overdrafts, revolving credits) on mobile channel as well. Replacing maximum amount offered with optimum amounts tailored to each customer individually.
Extend the number of loans and enrich the loan portflio.
Increased profitability with much less associated risk by:
- Decreased time for laon apporval proces feedback to 10 minutes and and time to cash in 20 minutes
- Replaced maximum amount offered with opitmum amounts tailored to each customer.
- Activation of sleeping clients base.