Rapid developments in technology and changes in societal needs have an enormous effect on many segments of the financial sector, such as the payment system, which in simple terms is a system used to settle financial transactions. Recent decades have seen this system evolve into an intricate and complex IT-system, which also applies to ING Bank Netherlands. Transaction processing in ING is subject to a central European clearance system, which does not operate continuously. Although the customer experience for transactions is handled in real time, the actual booking of the transfers has to wait for the clearance system to be processed. This has an impact on the load that needs to be handled by the IT system, where transactions accumulate during weekends and bank holidays. This results in peak days, where the bank needs to book considerably more transactions. Timely processing of data is one of the most critical factors in banking for customer satisfaction; therefore, a reliable prediction of the number of transactions to be processed is extremely valuable. Within ING, we developed a model that predicts the number of daily payment transactions, which we consider as a time series analysis problem. The solution is a multiple linear regression model that combines calendar features along with features based on domain knowledge about business and ING's IT systems. This study resulted in a competent model used by the IT department to take effective actions and do capacity planning further ahead. Our predictions brought business impact in the bank, with 1) considerable reduction in cost 2) IT incident prevention and 3) finally, customer satisfaction improvement. Moreover, our resulting model is able to score R-squared of 0.97.