Design Rules for Highly-Evolvable Enterprise Software

In large and mid-size organizations, inferior enterprise IT systems result in massive waste of time and resources, as well as missed opportunities.  To cite a single dramatic example, the IRS has spent several billion dollars on failed attempts to modernize their systems.

This collaborative research project with Shinsei Bank developed design rules to reduce the cost of developing, deploying, and modifying enterprise IT systems by an order of magnitude or more.  The starting point for the research is the software development methodology pioneered by Jay Dvivedi over his forty-year career, including two decades of work at Citibank and Shinsei Bank.  The research also draws on my doctoral research on computer-assisted organizing, Herbert Simon’s theories of complex systems developed in The Sciences of the Artificial, and prior work on Shinsei’s IT systems by my colleagues Dave Upton and Brad Staats.  I am very grateful to Jay for his patient mentoring, to Shinsei Bank for supporting this research, and to the team at Shinsei’s Banking Infrastructure Group.

The project has four primary goals:

  • Formulate a set of design rules for producing highly-evolvable enterprise software architectures
  • Develop a toolkit, in the form of an open source software framework, to support the creation of systems that follow these design rules
  • Study how computer system designs that mirror the underlying logic of the business can reduce costs of auditing and compliance, specifically in the case of J-SOX
  • Investigate the use of computer-assisted organizing techniques to dynamically optimize the allocation of work to available workers

The findings of the project are reported in blog format here.

Computer-Assisted Organizing

My Ph.D. dissertation, entitled Computer-Assisted Organizing (2009), explores how networked computer systems influence the structure and performance of human organizations.  I argue that computer-assisted organizations can more easily decompose information processing tasks into differentiated subtasks and delegate these subtasks to specialists, often outside the formal boundaries of the organization.  Computers propel this division of information processing by lowering the communication and agency costs arising from specialization, while simultaneously facilitating the creation and application of specialized knowledge.  As a result, valuable capabilities increasingly reside in organizational networks rather than within single organizations.

My dissertation includes an empirical investigation of how computer-assisted work (CAW) contributes to the ability of organizations to grow efficiently. Using survey data from over two hundred small wealth management firms, I analyze the relationships between CAW, growth aspirations, product variety, and product customization. The findings suggest that CAW enables business to grow more rapidly, especially when they have complex and customized product offerings. An additional study finds evidence of a link between CAW and revenue growth in over two hundred mid-size firms drawn from seven industries.

The January 2009 issue of Research Magazine features the wealth management study in an article titled “Mastering Advisor Technology: The optimal use of tech tools is essential for staying competitive and building business”.

Here are a few particularly thought-provoking books and articles for further reading.

Argyres, 1999, “The Impact of Information Technology on Coordination: Evidence from the B-2 ‘Stealth’ Bomber

Autor, Levy & Murnane, 2001, “Upstairs, Downstairs: Computers And Skills On Two Floors Of A Large Bank

S. Zuboff, 1989, In the Age of the Smart Machine

Sustaining Innovation in Mature Organizations

This research investigates how mature organizations can achieve both efficiency and adaptability.  Often, organizations become rigid and inflexible as they pursue greater efficiency.  “Standard operating procedures” and tightly controlled processes impede exploration and kill new ideas.  Scholars describe this as the “productivity dilemma“: measures that increase productivity in the short term tend to hinder learning and adaptation in the long term.  In essence, successful organizations dig their own graves.

We develop a theory of how perturbations disrupt established organizational routines, creating opportunities for innovation.  Perturbations occur naturally in young and growing organizations, but mature organizations must deliberately induce perturbations in order to sustain innovation.  Such deliberate perturbation runs counter to deeply ingrained tendencies of mature organizations, where managers normally struggle to keep processes under control and prevent disruptions.  Some organizations, such as the Toyota Motor Company, seem to use perturbation effectively, and we use examples from Toyota to illustrate our theory and develop hypotheses about how organizations can leverage deliberate perturbation.

This is collaborative work with Bradley R. Staats, Michael L. Tushman, and David M. Upton.  The paper is available for download from SSRN.

Share on Google+Share on LinkedInTweet about this on TwitterShare on FacebookPin on Pinterest