TU Delft Supervisor: Dr. G. Gousios Magnet.me Supervisor: Ir. A. Nederlof

A lot of interesting information can be extracted from the way users interact with a website. The regular way of extracting this information is through Google Analytics. Unfortunately, Google Analytics has two limitations. First, because Google Analytics uses a pageview as unit of measurement, it is unable to describe all user interactions. For example, tracking users on different devices is hard. Second, the reports generated by Google Analytics are not tailored for a specific business. For example, Google Analytics ’ support for key performance indicators is very limited. We investigated how to build a system that is capable of providing these deeper insights. An event is used as the unit of measurement to provide the flexibility that is required to track all user interactions. We designed the system to cope with the data volume required by Magnet.me. This is done by using MongoDB as storage engine. We have shown that our system satisfies our requirements by performing performance tests