Measuring Software Productivity
It is curious to me how most of the software engineering advice and best practices are based on mostly anecdotes. Software organization generates so much artefacts and other data that could be measured but I rarely see it turned into actual numbers. For instance I have never seen monetary value (or some other value) assigned to (say) a CI/CD-pipeline or something obvious like having a revision control system in use
I once sold this at the customer by calculating the cost of making manual updates to a production system, but I didn’t go further than that. One could probably assign probabilities of mistakes and their cost but even just the cost of work was enough that time. I’d like if I had framework for this and I wonder, if frameworks like DORA and SPACE could help me?
I recently read a book called Build: Elements of an Effective Software Organization | Swarmia and I’ve read before Your Code as a Crime Scene, Second Edition: Use Forensic Techniques to Arrest Defects, Bottlenecks, and Bad Design in Your Programs by Adam Tornhill and while they give ideas (and latter is more tuned towards improving software) I am still looking for more ”business-like” book on the topic with hard numbers. Search continues on…