Research · Book in progress

Leading Software Architecture Revolutions

I am writing a book on leading large software architecture changes with Joseph (Joe) Yoder, founder of The Refactory and co-author of the well-known "Big Ball of Mud" pattern. The book is a pattern language for senior technology leaders, the people who must decide whether to undertake a major architecture change and then carry the organisation through it.

The idea

Successful software products last a decade or more, and over that time their architecture falls behind the business. New markets, business models and team structures arrive, and the system cannot follow. When the gap is too large for ordinary, gradual improvement, the organisation needs a deliberate, visible effort to transform the architecture. We call this an architecture revolution.

A revolution does not mean rewriting the system from scratch. The book argues against the big-bang rewrite and recommends changing the architecture in small, safe steps.

Why it is about leadership

The hard part of an architecture revolution is rarely the technology. It is making the organisation aware of the need for change, securing commitment, keeping momentum and budget for several years, handling resistance and reorganising teams. Most writing on software architecture covers the technical side and leaves these responsibilities out. Our patterns cover the leader's role and draw on the change-management literature as well as software practice.

The patterns fall into four groups that follow the course of a transformation:

  1. Group 01

    Creating awareness

    Showing the organisation that the current architecture is a real problem.

  2. Group 02

    Preparing and measuring

    Setting goals and measuring whether the effort is moving in the right direction.

  3. Group 03

    Prioritising activities

    Choosing what to change first, from the overall vision and roadmap down to the next small step.

  4. Group 04

    Organising teams

    Aligning teams with product domains so that they own their part of the change.

The work draws on my experience leading PagSeguro's architecture transformation as CTO, and on Joe's decades of consulting with companies in the same position.

Peer-reviewed

Published papers

All are peer-reviewed through the writers' workshops of the Pattern Languages of Programs (PLoP) conferences. Authors: Marden Neubert and Joseph W. Yoder. The first four papers form the Leading a Software Architecture Revolution series. The last is a separate paper on how architecture affects a company's agility.

Talks on this work

  • "Architecting Agility: How Software Architecture Influences Agile Practices", OOP 2025, Munich, and Agile Brazil 2024.
  • "Domain Modeling Meets GenAI: Turbo-Charging Agile from Domain to Deployment", XP 2026, São Paulo, and Agile Brazil 2025.

Links on Speaking.

Earlier publications

Before the architecture work, I published on agile methods at scale and, earlier still, on search technology and e-commerce data science.

  • "Scaling Scrum Step by Step: The Mega Framework". AGILE 2012.
  • "Moving back to Scrum and scaling to Scrum of Scrums in less than one year". Experience report, SPLASH/OOPSLA 2011. Conference page (opens in a new tab)
  • Papers on marketplace fraud detection, product categorisation and web form filling with university partners at UOL (WWW 2009, LA-WEB 2009, ACM SAC 2010, JASIST 2011).
  • "Adding compression to block addressing inverted indexes". Information Retrieval, 2000.
  • "Efficient distributed algorithms to build inverted files". ACM SIGIR 1999.

Full list on Google Scholar (opens in a new tab).