Huge marketplaces with millions of customers
Great experience in Risk Scoring Models and Profit Optimization
Improving processes and client experience
Very sensitive to statistical modelling
Reasons to contact us
Implementation of the model for BNPL Finance Product on the online marketplace Kasta for purchasing in installments.
Main Challenges:
- 80%-90% of customers must receive approval, as a low approval rate will negatively impact the user experience.
- The delinquency rate should be close to 0, as the financial product has a 0% interest rate.
Efforts were concentrated on finding non-linear complex patterns that helped achieve the desired results. Modern machine learning algorithms provided the necessary resolution at a level of 60% Gini.
To implement the model, the data architecture was significantly modified to ensure fast processing of large amounts of data.
The company CreditUp offers short-term loans provided online on 24/7 basis. The core of the service is a 100% automated APS (Application Processing System). Customers fill up the application form by themselves from the web-page interface and receive the decision and money on their bank cards within a few minutes. Revenue is generated through interest charges, fees and commissions on loan balances. From the beginning and until now CreditUp needs customers who carries loans but don’t default.
In the U.S., a company like CreditUp could get information about customers from credit bureaus and know exactly what risk it takes. But in Ukraine not so many people have got records in the credit bureaus yet as it may satisfy CreditUp. So, the only way is to rely on internal data, and that is the main risk challenge appeared at the inception phase.
In another hand, due to the high and unclear risk of new customer acquisitions, revenue of the company is extremely dependent from customers retention. In most financial institutions in the U.S. analytics functions would exist within risk, sales, marketing, IT departments. But for CreditUp such organization is not suitable due to human capital limitations and also because of a need for speed in order to get the growth goals achieved.
In the late February ITIS-analytics became a single independent source of analytics support for CreditUp executives and gave a broad view on all aspects of the business, as opposed to a one-dimensional narrow focused analytics.
The ITIS-analytics system was integrated with CreditUp data sources. The first major achievement was implementation of the statistical scorecard built by the analytics system that led to the increase of new customer acquisition by more than 100%. The statistical model uses the data entered by a customer, as well as his/her behavior on the web-page. For instance, customers who read the agreement longer tend to pay in time. CreditUp’s business strategy, based on a customer centric approach, requires deep customer behavior research. Using ITIS-analytics system the customers were split into business segments. This segmentation gives clear understanding of who are the most profitable customers. It ensures that the loyalty program targets right people. In order to support customer retention campaigns CreditUp employs statistical attrition model.
As CreditUp is a finance company 100% based on Hi-End IT solutions, ITIS-analytics provides online Business Intelligence tool available for executives at any place in the world at any time via a tablet or smartphone. The freshest sales, revenue or losses figures accessible in a few clicks.
CreditUp has solid growth of the customer base and revenue.
