Understanding Conway's Law in Agile

BONUS: From Waterfall to Flow—Rethinking Mental Models in Software Delivery With Henrik Mårtensson

In this BONUS episode, we explore the origins and persistence of waterfall methodology in software development with management consultant Henrik Mårtensson. Based on an article where he details the history of Waterfall, Henrik explains the historical context of waterfall, challenges the mental models that keep it alive in modern organizations, and offers insights into how systems thinking can transform our approach to software delivery. This conversation is essential for anyone looking to understand why outdated methodologies persist and how to move toward more effective approaches to software development.

The True Origins of Waterfall

"Waterfall came from the SAGE project, the first large software project in history, where they came up with a methodology based on an economic analysis."

Henrik takes us on a fascinating historical journey to uncover the true origins of waterfall methodology. Contrary to popular belief, the waterfall approach wasn't invented by Winston Royce but emerged from the SAGE project in the 1950s. Bennington published the original paper outlining this approach, while it was Bell and Tayer who later named it "waterfall" when referencing Royce's work. Henrik explains how gated process models eventually led to the formalized waterfall methodology and points out that an entire generation of methods existed between waterfall and modern Agile approaches that are often overlooked in the conversation.

In this segment we refer to: 

• The paper titled “Production of Large Computer Programs” by Herbert D. Benington (direct PDF link) Updated and re-published in 1983 in Annals of the History of Computing ( Volume: 5, Issue: 4, Oct.-Dec. 1983)

• Winston Royce’s paper from 1970 that erroneously is given the source of the waterfall term. Direct PDF Link.

• Bell and Thayer’s paper “Software Requirements: Are They Really A Problem?”, that finally “baptized” the waterfall process. Direct PDF link.

•  

Mental Models That Keep Us Stuck

"Fredrik Taylor's model of work missed the concept of a system, leading us to equate busyness with productivity."

The persistence of waterfall thinking stems from outdated mental models about work and productivity. Henrik highlights how Frederick Taylor's scientific management principles continue to influence software development despite missing the crucial concept of systems thinking. This leads organizations to equate busyness with productivity, as illustrated by Henrik's anecdote about 50 projects assigned to just 70 people. We explore how project management practices often enforce waterfall thinking, and why organizations tend to follow what others do rather than questioning established practices. Henrik emphasizes several critical concepts that are often overlooked:

• Systems thinking

• Deming's principles

• Understanding variation and statistics

• Psychology of work

• Epistemology (how we know what we know)

In this segment, we refer to: 

• Frederik Taylor’s book “The Principles of Scientific Management”

• The video explaining why Project Management leads to Coordination Chaos

• James C. Scott’s book, “Seeing Like a State”

• Queueing theory

• Little’s Law

The Estimation Trap

"The system architecture was overcomplicated, and the organizational structure followed it, creating a three-minute door unlock that required major architectural changes."

Henrik shares a compelling story about a seemingly simple feature—unlocking a door—that was estimated to take three minutes but actually required significant architectural changes due to Conway's Law. This illustrates how organizational structures often mirror system architecture, creating unnecessary complexity that impacts delivery timelines. The anecdote serves as a powerful reminder of how estimation in software development is frequently disconnected from reality when we don't account for systemic constraints and architectural dependencies.

In this segment, we refer to Conway’s Law, the observation that explicitly called out how system architecture is so often linked to organizational structures.

Moving Beyond Waterfall

"Understanding queueing theory and Little's Law gives us the tools to rethink flow in software delivery."

To move beyond waterfall thinking, Henrik recommends several resources and concepts that can help transform our approach to software development. By understanding queueing theory and Little's Law, teams can better manage workflow and improve delivery predictability. Henrik's article on coordination chaos highlights the importance of addressing organizational complexity, while James C. Scott's book "Seeing Like a State" provides insights into how central planning often fails in complex environments.

About Henrik Mårtensson

Henrik Mårtensson is a management consultant specializing in strategy, organizational development, and process improvement. He blends Theory of Constraints, Lean, Agile, and Six Sigma to solve complex challenges. A published author and licensed ScrumMaster, Henrik brings sharp systems thinking—and a love of storytelling—to help teams grow and thrive.

You can link with Henrik Mårtensson on LinkedIn and connect with Henrik Mårtensson on Twitter.