This paper describes an approach taken to optimise the timetable of the mobile library service operating on the Isle of Wight. The mobile library visits over 90 communities on the island, offering books, DVDs, videos and CDs and operates on a periodic timetable. The optimisation problem is formulated as a multiple travelling salesmen model with additional time-balancing constraints on route durations. The paper also shows ways in which data required for the model, in particular travel times, was gathered and discusses practical issues arising in preprocessing the data to fit the purposes of the case study. The model is used to produce an improved timetable over the current one that implies driving time reductions of up to 25% and yields routes that are better balanced in terms of time spent on the visits made each day. The model is also used to test various scenarios differing with respect to the number of locations visited and days over which the service operates.