Skip to Main Content
The paper presents an in depth study of the time between retried calls by analyzing the distribution properties in the 0-5 minutes, 5-60 minutes and 1-180 hours intervals. For this study a large quantity of real life data was used (5.5 million calls from 2007-05 to 2008-09) coming from a prepaid long distance voice over internet protocol (VoIP) provider. The distribution and identified spikes were fitted against common continuous distributions. An analysis of variations across different destinations and time intervals was performed. Using the time between retried calls it was shown that customer complaints are very frequent (90% of the cases) after days with high retry rate. The results obtained lead to the assumption that a VoIP telecommunication network can be characterized using this metric. Time between retried calls and well known human habits are shown to be connected, fact that can be used to build better customer oriented services.