Design Rules
The modularity theory that makes contract manufacturing possible at all.
Summary
Baldwin and Clark, both Harvard Business School professors, developed a formal theory of modularity and applied it to the history of the computer industry.
The theoretical framework starts with the design structure matrix, a tool for representing which elements of a design depend on which others. A design is modular when it can be partitioned so that dependencies within each module are dense and dependencies between modules run only through defined interfaces. Achieving this requires splitting the design into visible information, meaning the architecture, interfaces and tests that all modules must respect, and hidden information, meaning the internal decisions each module can make independently.
The economic argument is that modularity creates options. Once modules are independent, each can be redesigned without disturbing the others, which means each becomes a separate experiment whose value is the best outcome among the attempts made. Baldwin and Clark formalise this using option pricing, showing that the value of a modular design increases with the number of modules and with the uncertainty in each, because more independent attempts produce a better maximum. This is why modularity is valuable precisely where outcomes are uncertain.
They identify six modular operators: splitting a design into modules, substituting one module for another, augmenting by adding a module, excluding one, inverting to move a common function to a higher level, and porting a module to a different system.
The historical application is IBM's System/360, which imposed a common architecture across a product line and made components interchangeable. Baldwin and Clark argue this created the conditions for an entire industry of independent module suppliers, which eventually displaced the integrated firm that had created the architecture.
The book is formal and demanding. A promised second volume on the industry's later evolution appeared much later. The framework has been influential in strategy, in software engineering and in organisation theory.
Key takeaways 7
- A design is modular when dependencies within modules are dense and dependencies between them run only through defined interfaces.
- Modularity requires separating visible information, meaning architecture and interfaces, from hidden information internal to each module.
- Modularity creates options because each module can be redesigned independently, making each a separate experiment.
- The value of a modular design rises with the number of modules and with the uncertainty in each, since more independent attempts produce a better maximum.
- Six operators describe what can be done to a modular design: splitting, substituting, augmenting, excluding, inverting and porting.
- IBM's System/360 imposed a common architecture that created the conditions for an industry of independent module suppliers.
- An architecture that enables independent module suppliers eventually transfers value from the integrator to those suppliers.
Who should read it
Technologists and strategists making architecture and outsourcing decisions who want the formal theory of why modularity creates and transfers value.
Where it sits in the reference 2
Similar books 6
Chosen by shared topics, the entries that hold them, and the people who recommended both.
