Jump to main content

Please note that on our website we use cookies necessary for the functioning of our website, cookies that optimize the performance. To learn more about our cookies, how we use them and their benefits, please read our Cookie Policy.

Can you pay? Will you pay? – How credit decisions are made

Every 1.3 seconds, round-the-clock, Intrum Justitia helps to make a decision whether or not an individual or a company will be granted credit. This process is crucial to efficient credit management and is also highly complex and diverse.

A credit decision is made up of two major components; first, data and, secondly, a process which evaluates and combines this information. Intrum Justitia collects information on previous credit decisions, how the credit developed and, if permitted in a particular market, data on payment demands and enforcement history. This is then complemented with external data, such as information from local chambers of commerce.Roland Gruneus

“Former payment behavior is a critical factor and persistently late or defaulting payers are more likely to be declined credit,” says Roland Gruneus, Chief Technology Officer at Intrum Justitia.

Every process begins with a set of binary conditions to screen out applicants that are not eligible under any circumstances due to age limits and similar factors. A process often involving complicated algorithms is then carried out to combine and assess the available information. This process can be fully automated as in the case of e-commerce, where decisions have to be made within seconds, or partially manual which is more common when large transactions or when companies are involved.

The precision in credit decisions is typically very good but depends of course to a large extent on the quality of the data available. An important factor is to follow up decisions to see how they match with what actually happened. When deviations between the outcome and the data used to construct the decision model occur, the precision is negatively impacted and the model then needs to be adapted. This can happen for instance when a merchant starts addressing new customer segments and the new group does not match the data on which the model is based.