FACTORING AND DEBT COLLECTION OPTIMIZATION
Managing overdue receivables is a particularly important task for companies to remain competitive. If we want to support the collection of a large number of small and unsecured receivables then it is worth modelling the probability of collection for each account receivable and each phase of the collection process. The outcome of this modelling is then used for process differentiation and to determine the right steps of the collection process at the level of individual transactions. There are various statistical methods for the assessment of the above processes as well as of the relevant players during debt collection. We compare the probability of different outcomes, resource requirement and usefulness of each step of the collection process to improve the company’s debt collecting process which will lead to well-grounded planning and higher return.
ATM RELATED SERVICES
Banks have two conflicting sets of considerations regarding their ATM services. The first approach is to increase customer satisfaction, which requires ATMs at adequate locations, carrying an optimal level of cash with the right denominations. The second approach is to minimize running costs and the costs of the alternative to the cash stored in the machines. Using mathematical-statistical support for processes, demand for cash is forecasted in the surroundings of a certain location (ATM) based on regional databases and historic transaction data. Clustering algorithms help us create regions with similar cash demand and determine the points where the clients of a bank need an ATM. An optimal filling algorithm can be developed based on the turnover forecast, whereby the cash levels in each ATM can be minimized while the probability of running out of money will remain at an acceptable level.
Considering the high costs of acquisition it is especially important to retain customers. Once we have an accurate churn model, we optimize retention campaigns in relation to the results of customer value calculation. The essence of the above-mentioned models is that we estimate the probability of attrition event within a certain period of time, with regard to certain customers.
A customer value calculation project is meant to quantify the value the company represents for its customer, in cash. On the one hand, this includes the discounted amount of incomes realised by the customers while they were with the company, and on the other hand it also includes other additional values available for the company (e.g. network effects). Results may form the basis of efforts to retain customers and of loyalty programs, so the company can focus on the most profitable clients.