Gas path analysis (GPA) is an effective method for determination of turbofan component condition from measured performance parameters. GPA is widely applied on engine test rig data to isolate components responsible for performance problems, thereby offering substantial cost saving potential. Additional benefits can be obtained from the application of GPA to on-wing engine data. This paper describes the experience with model-based GPA on large volumes of on-wing measured performance data. Critical is the minimization of the GPA results uncertainty in order to maintain reliable diagnostics and condition monitoring information. This is especially challenging given the variable in-flight operating conditions and limited on-wing sensor accuracy. The uncertainty effects can be mitigated by statistical analysis and filtering and postprocessing of the large datasets. By analyzing correlations between measured performance data trends and estimated component condition trends errors can be isolated from the GPA results. The various methods assessed are described and results are demonstrated in a number of case studies on a large turbofan engine fleet.