Telescope performance can be characterized by a number of metrics e.g. mirror reflectivity, seeing, readout noise, observing overheads. In deciding where to invest limited operational resources to improve performance, one needs to predict the impact of given enhancements on scientific productivity. E.g. for the same cost, is it more important to reduce CCD readout noise by a factor of 2, or to improve instrument throughput by 30%? Knowing the mix of programs at a given telescope, and the dependence of signal-to-noise on the various parameters, the % gain in scientific productivity can be predicted for a given % improvement in any parameter, allowing optimal investment of the operational budget. We describe operational metrics used to monitor the performance of the 4.2-m William Herschel Telescope on La Palma, and give examples of current and planned enhancements which have been prioritized by comparing predicted gains and costs. These enhancements should deliver a total gain approximately 30% in productivity, equivalent to approximately 100 extra observing nights per year.