The implementation of a queuing management system is a critical task in every single store attendance service. This is very difficult to accomplish because we have to deal with customers, and in particular with its impatience. The goal must be always to have short queues and a very fast attendance service. However, rarely this is achieved. For many reasons, attendance service deteriorates as time pass or as the number of people waiting increases, reducing the quality of service of the store, and causing discomfort in the people that are in the queue. High waiting times are very risky to the store, because customers get impatient of waiting and leave the queues, provoking frequently customer dissatisfaction and leading to loss customers. In this paper, we analyzed and implemented several approaches for predicting the expected waiting time given an attendance time history of a store, and how it could evolve in the future. To do this, we used a real-world data set collected in a specific telecommunications company store, having important service management issues, which makes it a very good case study, and a strong application case.