This paper describes a case study in which the software architecture of a business application was modified to improve runtime performance. Such modifications should be performed whenever application users encounter known areas of sluggish response, long periods of maintenance, or a change in processing volume requirements. For this particular study, a framework for source code instrumentation was designed to provide convenience, data granularity, and improved control for profiling of elapsed time, operating system events, and CPU counters. The study confirms that proper selection of algorithms and data structures is essential for peak performance. Furthermore, known optimization methods, when summarized, can be used as a roadmap for tuning once hotspots are identified. Upon completion, this optimization project resulted in a speed-up factor of 18 for a typical data set.