How Your Customer Data Can Build Alternative Credit Models

Published by Pngme on

Information derived from non-traditional data sources reveals dynamic and meaningful insights that can be used in assessing risk and ultimately improving loan approval rates.

Non-traditional data, alternative data, and machine learning have paved the path for stronger credit review metrics and have allowed banks and FIs to considerably increase their credit approval rates. Not only is this a win-win for banks and customers, it’s also a clear step towards financial inclusion. 

Although the terms are sometimes used interchangeably in this rapidly evolving space, for the purposes of our work, we find it useful to differentiate between non-traditional and alternative data. Both forms of data are helping banks and FIs approach credit scores in new ways.

Non-traditional data is financial data that banks have historically not been able to use in credit models, such as USSD transactions. 

Alternative data can encompass any form of data not traditionally used in credit models, such as behavioral data, (geographic location, insurance, gambling etc.)

Pngme uses non-traditional data in two different ways in order to build new credit models for customers:

  1. Create model on historical SMS data using backtest of Pngme services: If you have historical SMS data and an existing lending program, Pngme can build a preliminary model, which is then revised over one month with our data science team and tested across randomized control groups.
  1. Run a randomized control trial with Pngme (or pseudo randomized trial, if a randomized trial is not possible): Without historical data, Pngme can help you start a lending program with our models. Pngme can initially provide a general purpose model, and once lending data is available, we can use SMS data from loans to build a credit model tailored to your user base.

Pngme provides the embedded data layer that is powering new credit models. Gain a deeper understanding of your customers’ financial identities, uncover hidden customer segments and unlock new opportunities for your product road map with our technology. Pngme can provide the technology to collect user-permissioned data quickly and provide the tools to make that data understandable and help you have data behind every action and lending decision.


Follow us on Twitter, Facebook & Linkedin