A customer self-service tuning system assists a mobile device user in adjusting his network settings and device configuration. Device, network and performance metrics are stored for a population. Over time all of these suggest changes in the device or network that change measured performance. The potential improvement(s) are presented to the user for selection. Performance measurements are taken and the user can choose to go forward with the new configuration or to revert. Each time a user requests self-care service, a new cohort is extracted from the then current overall population and analyzed for similarity to the user at that point in time. Each time an improvement is selected, the history of device states and network states is augmented. Identified variances are transformed into an action plan specific to a user and implemented on the device or network upon concurrence.