The historical development of techniques for prediction of transient response sensitivity of SISO feedback control systems to plant changes is reviewed starting with the analogue approach of Tomovic and leading up to the current state of the art. It is shown that for a given transient response performance specification, significant and quantifiable improvement in sensitivity reduction can be achieved in conventional systems by including feedback compensation.Extension of the methodology to the prediction of the transient response sensitivity of high-order systems via loworder models is included. The problem of parameter tolerance setting to meet a performance envelope specification is used to illustrate the method. Finally, transient response sensitivity functions are used to assess the relative merits of systems stabilized by sensor measurements compared to observer compensation. The intuitive assumption that the observerbased design (which is readily implemented via microprocessor technology) is less robust than an instrumented system is confirmed