No Single Metric Captures Productivity
This paper argues that there is no metric that adequately captures the full space of developer productivity, and that attempting to find one is counterproductive. The authors encourage the design of a set of metrics tailored for answering a specific goal.
The urge to measure the productivity of developers is not new. Since it is often the case at organizations that more code needs to be written, many attempts have been made to measure productivity based on lines of code (LOC). For example, in early 1982, the engineering management of developers working on software for the Apple Lisa computer decided to start tracking LOC added by each developer. One week, the main user interface designer, Bill Atkinson, optimized Quickdraw’s region calculation machinery and contributed -2000 LOC. The management stopped asking for his LOC. Although measuring engineer productivity by LOC is clearly fraught, anecdotes like this abound on the internet. Organizations have continued to search for better easy ways to measure developer productivity. We argue that there is no metric that adequately captures the full space of developer productivity, and that attempting to find one is counter-productive. Instead, we encourage the design of a set of metrics tailored for answering a specific goal.
- Ciera Jaspan
- Caitlin Sadowski
