“Two Case Studies Motivating Efficiency as a “Spendable” Quantity“, by Alistair Cockburn.
A bottleneck limits the rate of production of a system. Clearly we want to improve its performance. But what should we do about non-bottlenecks? Alistair discusses several strategies, including the non-obvious one that the non-bottleneck may be better off deliberately spend extra time exploring alternatives before feeding into a bottleneck. He points out that there’s no single strategy that always works; it’s context-dependent.