“Organizations that design systems are constrained to produce systems which are copies of the communication structures of these organizations.” Most of us have come across Conway’s Law at some point. And we have a hunch that the structure of organizations probably impacts software quality.
This weekend’s first of two – admittedly scholarly – articles by Microsoft Research provides evidence in that “…organizational metrics are better predictors of failure-proneness [of software] than the traditional metrics used so far.” More details in the full article: https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/tr-2008-11.pdf
The same holds true for test effectiveness: “The relative number of bugs reported by a test suite per execution … is negatively correlated with organizational path and higher level manager metrics … . This suggests that test suites with owners stemming from the same organizational subgroups are more effective with respect to finding code issues.” More details here: https://www.microsoft.com/en-us/research/wp-content/uploads/2014/09/The-Impact-of-Test-Ownership-and-Team-Structure-on-the-Reliability-and-Effectiveness-of-Quality-Test-Runs.pdf
(I owe Celonis’ CIO, Martin Klenk, for pointing me towards the first article.)
While the above might point to small and autonomous teams as a solution, Matt Le May thoughtfully suggests that “the quest for small, fast, and autonomous teams is creating fragmented, confusing, and disjointed experiences.” Judge for yourself: https://medium.com/on-human-centric-systems/agile-is-ruining-your-product-3674c8afea76
Enjoy your long Pentecoast weekend,