In 2017, online lending platform Creditstar Group AS sought to improve their scorecards in Poland in order to increase loan approval rates and decrease defaults. The lender's biggest challenge was understanding the spending behaviours of their so-called 'thin-file' customers or clients with incomplete or non-existing credit histories.
Creditstar's Data Science team undertook the task of testing Nordigen's transaction categorization and behavioural scoring engines to determine whether Nordigen's solution improved the overall accuracy of Creditstar's scorecards.
After comparing credit scoring model performance with variables Creditstar used before and after integrating with Nordigen, Creditstar reported that Nordigen's transaction categorization helped them improve their scorecard accuracy and increase their GINI by 8.4 percentage points.
"Nordigen variables related to loans and financial customer services are consistently found to be top predictors according to the feature importance analysis," says Creditstar's Principal Data Scientist Mykola Herasymovych.
"Progress requires innovation and we're incredibly proud to support Creditstar on their mission to improve people’s lives by providing seamless financial services by adopting the latest financial technology," adds Rolands Mesters, co-founder and CEO of Nordigen.
Creditstar is one of Europe's leading non-bank lenders with over 500,000 registered customer accounts across 8 countries, including Poland, Sweden, Spain, Finland. and the United Kingdom. Creditstar was founded in 2006 in Estonia, where it still has its headquarters. Creditstar uses automated processes, algorithms and data analysis to make financial products easily available to a population of more than 175 million people in their target markets.