This talk gives an overview of my first-year probation work on using timed event structures as a semantic model for concurrent processes. I will first introduce the basic idea of event structures as a model of true concurrency. I will then present my current definition of timed labelled event structures, where events are equipped with durations and finishing times. The aim is to use this semantics to extract timing information from concurrent processes, including worst-case completion time, while the treatment of average completion time remains work in progress. I will give some intuition about how this analysis might work.