Allen Ward (d. 2004) was a mechanical engineering professor at the University of Michigan whose research on Toyota's product development system provided one of the most distinctive elements of the Poppendiecks' lean software framework. His discovery that Toyota practiced set-based concurrent engineering — exploring multiple design alternatives simultaneously rather than iterating on a single solution — became the foundation for the set-based-design concept and the decide-as-late-as-possible principle.
The Set-Based Concurrent Engineering Discovery
Ward's research, conducted through the University of Michigan's Japan Technology Management Program, involved direct study of Toyota's product development processes. Western engineering practice typically followed a point-based approach: select a promising design early and iterate on it, adding constraints as the design develops. Toyota's approach was fundamentally different: maintain multiple design alternatives through most of the development process, allowing each alternative to generate information about the design space, and converge to a single solution only when the decision must be made.
This approach appears counterintuitive — maintaining multiple designs seems more expensive than focusing on one. Ward's analysis showed the opposite: the information generated by parallel exploration allowed Toyota to make far better decisions at the convergence point, avoiding the expensive late-stage rework that afflicted point-based approaches when a committed design turned out to have fundamental problems.
Ward collaborated with jeffrey-liker and Durward Sobek on this research, publishing "Toyota's Principles of Set-Based Concurrent Engineering" in the MIT Sloan Management Review in 1999. This paper is the primary academic source for the concept the Poppendiecks would translate into software.
Influence on the Poppendiecks
mary-poppendieck's 2015 essay lean-software-development-backstory-2015 explicitly credits the University of Michigan program — and Ward's research specifically — as a primary source for lean software development's product-development-oriented principles. The set-based-design concept in lean-software-development-agile-toolkit-2003 is a direct translation of Ward's finding: rather than committing to a single architectural or design approach early, software teams should explore multiple options in parallel and defer the commitment decision until the last-responsible-moment.
This connection also shapes the decide-as-late-as-possible principle: late decision-making is not procrastination but a disciplined strategy for preserving optionality until the information needed for a good decision is available. This framing — decision timing as an information problem — comes directly from Ward's analysis of Toyota's concurrent engineering practice.
Ward's Death and Legacy
Ward died unexpectedly in 2004, the year after lean-software-development-agile-toolkit-2003 brought his ideas to the software community. His posthumous book "Lean Product and Process Development" (2007), completed by Durward Sobek, extended his research into a comprehensive framework for lean product development. The timing is significant: Ward's ideas were reaching their widest audience in the software community precisely when he died, a fact that mary-poppendieck notes in the Backstory essay with evident regret. The Poppendiecks are among the primary carriers of Ward's intellectual legacy outside the product development engineering community where his work originated.