Their "mirroring" terminology sounds much vaguer and less useful than the original "Conway's Law" framing: "Organizations build architectures that mimic the co

Their "mirroring" terminology sounds much vaguer and less useful than the original "Conway's Law" framing: "Organizations build architectures that mimic the communications structures of the organization" aka "if you have three teams you'll get a three-pass compiler"[0] Their "mirroring" concept blurs two fundamental questions into one jumbled mess: (a) is the organizational structure well adapted to the technical requirements of the domain and (b) are the technical requirements of the domain actually the most important requirements to adapt to for the domain. If you don't disentangle those questions, you don't know if the orgs that don't seem to be mirroring are failing to structure correctly or if you're failing to understand why they structured correctly the way they did. By way of example, lots of highly successful sales-driven organizations and/or lean startups might well argue sales organizational needs trump technical organizational needs, and they could well be right. Focusing on the paper's definition of mirroring prevents you from drawing any of the important conclusions around whether orgs are failing to structure correctly or are structuring correctly around a different and potentially testably more valuable optimization metric. And yes, I'm a bit frustrated as the same interview time and analysis time could have allowed the authors to disentangle those forces at least somewhat and give a simply better answer to both the original question and more, simply at the cost of a better framed research question. [0] https://en.m.wikipedia.org/wiki/Conway%27s_law https://en.m.wikipedia.org/wiki/Conway%27s_law