Digital IIR filter implementations are important building blocks of most communication systems. Conventionally, the filters are specified via amplitude and phase in the frequency domain as given by the matched filter theory. Digital filter implementations, nonlinear analog components and channel characteristics introduce a multitude of additional effects, though. These are not taken into account by the matched filter theory. Which, in turn, leads to results providing a rough estimate, at best. Our work reforms the design process, defines the system's bit error rate as the main objective and searches the huge -- yet finite -- filter design space for suitable coefficients.