Many of you have probably heard about Conway’s Law. Introduced in 1967 it says that “any organization that designs a system will produce a design whose structure is a copy of the organization’s communication structure”. It basically means that the design of any complex computer system will reflect the communication boundaries of the organization that produced it (in other words, the computer system’s design will mirror the organization’s internal communication structure).
In the past, when teams were usually grouped by their function/expertise, Conway’s Law resulted in building siloed, multi-layer application architectures (UI, middleware, database, etc.). Therefore an understanding of Conway’s Law led to the creation of cross-functional teams and support for DevOps culture. If a team includes the full range of skills required for the development (like user-experience, backend, UI, database, etc.) it has fewer dependencies which, in turn, allows for having loosely coupled, microservice-based system architectures.
But even these cross-functional teams don’t operate in a vacuum and incorrectly designed org structures will eventually influence the overall architecture of the system. As one said – it seems like HR designs software architecture.
Recently an idea of “Inverse Conway Maneuver” has been introduced. The idea is to use Conway’s Law to our advantage and consciously structure an organization in a way that results in the desired system architecture. It means that technology specialists (i.e. software architects, technology leaders, etc.) should be involved in working out an org structure and it should not be left only for HR. As Bloomberg says: “Technology change is driving […] organizational change across increasingly software-driven enterprises. […] The causality question behind Conway’s Law, therefore, is less about how changing software organizations can lead to better software, but rather how companies can best leverage changing technology to transform their organizations.”.
Unfortunately, from my experience, I can say that many leaders and organizations ignore Conway’s Law altogether (similarly to ignoring an impact of Dunbar’s Number). They consider org structure and software architecture as disconnected things and are always surprised to see the results that are not as they hoped for. This encourages these leaders to implement (yet) another re-org designed by HR in Powerpoint that is meant to address all the problems. And another cycle begins…