On Conway's law
Conway’s law is a well known “law” of systems. It is an observation really:
Organizations which design systems (in the broad sense used here) are constrained to produce designs which are copies of the communication structures of these organizations.
https://en.wikipedia.org/wiki/Conway's_law
Martin Fowler also wrote similarly:
The key thing to remember about Conways Law is that the modular decomposition of a system and the decomposition of the development organization must be done together. This isn’t just at the beginning, evolution of the architecture and reorganizing the human organization must go hand-in-hand throughout the life of an enterprise.
https://martinfowler.com/bliki/ConwaysLaw.html
I would distill it as “codified systems follow, at best, the structure of the system that produces it”. This matches my personal experience.
Companies that do understand Conway’s law try to modify both the existing code structure and the organisational structure, slowly nudging both towards the desired structure. Or allow the teams to be flexible to nudge the system structure towards the company goals.
Companies that do not understand Conway’s law produce a structure that does not reflect the existing codebase structure, and instead focus on the expected business structure.
This creates a divergence between the code structure, its ownership and s responsibilities that translate into debt and constant wasted resources of dealing with the ever growing divergence
The mirroring hypothesis
Recently I came across “Exploring the Duality between Product and Organizational Architectures: A Test of the “Mirroring” Hypothesis” Alan MacCormack, John Rusnak, Carliss Baldwin explore the vailidity of the mirroring hypothesis
- https://thinkinglabs.io/articles/2026/06/20/beyond-shades-of-conways-law-validation.html
- https://www.hbs.edu/ris/Publication%20Files/08-039_1861e507-1dc1-4602-85b8-90d71559d85b.pdf
The mirroring hypothesis I interpret as a weaker version of conway’s law:
The mirroring hypothesis predicts that these different organizational forms will produce products with distinctly different architectures. Specifically, loosely-coupled organizations will develop more modular designs than tightly-coupled organizations.
It refers to the “modularity” or “coupling” aspect of conway’s law rather than the particular
They found strong evidence to support the mirroring hypothesis. I also believe this is a strong arguments for self organised units, similar to what agile proposes. Fred Emery coined the two design principles.
DP1 and DP2
DP1 and DP2 are that two design principles that Fred Emery attributed to all forms of organisation.
DP1 is the redundancy of parts. So many parts do the same thing to preserve the organisation. Think of managers referring to the same work and autocracy.
DP2 is the redundancy of functions. Each component of the organisation has the ability to do the same thing and they collaborate. It is democratic.
There is also the Laissez-faire which refers to organisations that fail to follow any of these arrangements.
I believe that Conway’s law, the agile formulation of acting as an independent cell with redundant functions and DP2 structures refer to the same thing. They are all about keeping an eye on the structure to fix things structurally, and enable the modularity that resilient structures exhibit.
They are politics so they are hard to optimise for. Without team or business structure considerations, rigid authoritarian or chaotic outcomes dominate the company which are counterproductive in different ways.