Selection of customers for operational and usage profiling.

Operational and usage profiles collected from customers provide developers and testers with valuable quantitative information on usage patterns of software being developed. Unfortunately, gathering such profiles from a large set of customers can be challenging due to time and resource constraints. In this paper we propose to use information about customer defects to narrow down a list of candidate customers to profile. We present a technique for selection and prioritization of minimal set of customers for operational and usage profiling to cover certain set of defects. The technique optimally selects a minimal set of customers for profiling and, once the set is identified, prioritizes the customers within the minimal set. We describe a validation case study confirming that this approach is scalable for a large customer base. Analysis results can then be used to close gaps in testing coverage and to improve the maintenance process.