The discipline of distributed algorithms aims at distributing a computational task among multiple, concurrent processors to achieve higher throughput, efficiency, performance and other advantages. Although parallel execution time, concurrency, scalability, isoefficiency function, and speed-up have been proposed in the literature as performance metrics, the use of speed-up and its variation, scaled speed-up, dominate the literature. Speed-up is defined as the ratio of the execution time of the best known serial algorithm to the execution time of the parallel algorithm on a given number of concurrent processors. This paper analyzes critically the role of speed-up from the perspective of an emerging class of distributed algorithms, termed asynchronous, distributed, decision-making (ADDM) algorithms. ADDM algorithms constitute the underlying control of many real-world systems, wherein the constituent sub-components are geographically dispersed, they interact between themselves asynchronously, and are permitted autonomy in local decision-making. The term real-world implies that such systems are subject to computer control and that they relate to everyday life and are beneficial to the society in the large. Analysis reveals several key limitations of speed-up from the perspective of describing the performance of ADDM algorithms. First, by definition, speed-up utilizes the execution time of a centralized system as the basis for comparison and is, therefore, only a relative measure. The centralized implementation is obviously neither optimal nor the absolute best. Second, for many systems that are amenable to ADDM algorithms, speed-up may not constitute a relevant performance metric. This paper adopts Ferrari's definition of performance and presents a frame of reference for the performance evaluation of ADDM algorithms. While novel decision criterion may need to be developed for each ADDM algorithm, given the diversity inherent in systems, the frame of reference implies that the ideal or absolute standard, relative to the criterion, must either be (i) derived automatically as the absolute best measure, utilizing the intrinsic definition of the problem or (ii) may consist of perfect decisions, determined by transcending the physical limitations of time. The absolute ideal is independent of the underlying computing infrastructure and, although unrealizable in the real world, it may serve as the absolute standard for comparing the effectiveness of the state-of-the-art distributed systems. Finally, this paper illustrates the determination of the absolute standards, corresponding to the designer's choice of the performance criteria, for a select few ADDM algorithms.