The scorecards were developed from a generic sample of Netherlands data extracted from the Creditsafe data pool. To increase the discriminative power of the scorecard solution, segmentation was conducted. The aim of the segmentation was to define a set of sub-populations that, when modelled individually and combined, rank risk more effectively that a single model on the overall population. Stepwise Logistic Regression has been used to develop the scorecards. The output from the logistic model has been transformed to a score which fits the required score to ln(odds) relationship, which is then transformed into a 1 – 100 Creditsafe score, which is shown on our website.
A suggested limit is applied to all companies using a combination of company information, financial information and the Creditsafe score. The credit limit is Creditsafe NL’s recommendation of the total amount of credit that should be outstanding with a particular NL business.