"Agile Management for Software Engineering" (2003) is Anderson's first book and the intellectual precursor to the kanban-method. Published by Prentice Hall, it applies eliyahu-goldratt's Theory of Constraints to software engineering project management, making the economic case for agile methods in the context of large corporations rather than small teams or startups.
The TOC application
The book's central move is treating a software development organization as a system with constraints — bottlenecks that limit throughput — and applying Goldratt's five focusing steps to identify and exploit them. Anderson argued that the standard project management pathologies in large software organizations (death marches, late delivery, feature bloat) were predictable consequences of managing software development as a cost center rather than a throughput system. The corrective was not to add more discipline (more process, more documentation, more oversight) but to identify and remove the constraints that limited the flow of working software to customers.
This framing was analytically distinct from the then-dominant Agile arguments. Where the Agile Manifesto emphasized values and principles, and where Scrum and XP emphasized process ceremonies and engineering practices, Anderson's TOC framing emphasized economics — throughput, inventory, operating expense — borrowed directly from Goldratt's "The Goal" and "Critical Chain."
The business case argument
The book's subtitle — "for Business Results" — signals its target audience: managers and executives in large corporations who needed an economic justification for agile adoption, not just practitioners who were persuaded by software craftsmanship arguments. This orientation toward the organizational economics of software development became a persistent thread in Anderson's work, culminating in the kanban-book's explicit framing of the Kanban Method as an evolutionary change management approach rather than a software development process.
Pre-Kanban significance
The book predates the corbis-kanban-experiment (2004-2007) and the development of what became the Kanban Method. It does not use "kanban" in the sense Anderson would later develop. But it establishes the intellectual vocabulary — constraint, throughput, flow, WIP — that Anderson would carry into his Corbis work and eventually synthesize with don-reinertsen's queueing theory, taiichi-ohno's pull systems, and w-edwards-deming's quality management.
The book belongs to the pre-kanban-era of Anderson's career, when he was working at microsoft on Feature-Driven Development and early agile adoption in enterprise contexts. The msf-for-cmmi-process-improvement paper from 2005 represents the last work of this era before Anderson's thinking shifted at Corbis.
Relationship to later work
"Agile Management for Software Engineering" and the kanban-book form a coherent intellectual arc: the first book establishes the theoretical framework (TOC, systems thinking, economic framing), and the second shows how that framework was empirically tested, refined with additional theoretical inputs from Reinertsen, and formalized into the Kanban Method. Anderson's lessons-in-agile-management blog compilation documents the thinking that bridges the two.