Computational Modeling of Project Organizations Under Stress
Can we design organizations the way we design products? Can we change their designs to make them more efficient/effective? These are questions my research partner and I endeavored to answer last year using Stanford’s Virtual Design Team software. Watch for our article in the March 2007 Project Management Journal. Building upon prior research on enterprise centralization and knowledge dynamics, our effort used computational methods to assess the behavior and project performance of different organizational designs in varying environments. The results reinforce contingency theory and suggest particular characteristics of different project environments that make one form relatively more or less appropriate than another. Practically, the answers to the research questions have direct and immediate application to project/portfolio managers and senior executives. Theoretically, broad classes of organizations are generalized and prescribe a novel set of organizational design guides. You can read the entire report from which our article was derived at: http://www.nps.navy.mil/gsbpp/ACQN/publications/FY05/PM-05-007.pdf
Posted by:
jdillard at 10 Jan 2007 under Implementing project management
3 Responses to “ Computational Modeling of Project Organizations Under Stress ”
Comments:
Leave a Reply
You must be logged in to post a comment.
January 11th, 2007 at 7:12 am
The structure vs. the people.
Perhaps this is our own version of the nature/nurture debate. I wonder if we’ll one day understand how much of the results of a team come from the structure, and how much depend on the level of functioning (interpersonal responsibility and maturity) in:
a) the leader
and
b) the team
I’ve seen large corporate efforts hamstrung by terrible structure.
And it occurs to me–a well functioning team would be more likely to:
1) notice that there’s problems
2) talk about them
3) make changes and see how they go.
And good structure can support such conversations.
I can see a lot of synergy between structure and leadership functioning.
January 12th, 2007 at 12:20 am
It strikes me ironic how modeling is applied to organizational structures when, in reality, there is no perfect organization–they are all man-made artifacts.
When I was at HP, we brought in Executive VP Rick Belluzzo (now he’s the CEO and Chairman at Quantum) to address project managers about key organizational issues facing project work. He said:
“I’ve spent a lot of time thinking about organizations and how to define the perfect organization, and I can’t do it; I can’t figure it out. I can’t draw and develop a field model that solves one hundred percent of our problems.
What I’ve learned is you have to optimize your organization based on the problems you want to solve today, and determine what things you want to manage, what areas you’re going to have to invest your time in and make work, because the organizational model doesn’t do it naturally.
And that’s where project management comes in. We need project managers to manage those areas you don’t have direct authority over and that need to be integrated, and you make the decision to move things along. We’re going to have a lot of that and that has to be a skill for us because that’s what’s going to make us better than other organizations.”
In essence, it’s up to project managers to do their jobs, regardless of what organizational structure they work in. A lesson for upper managers is not to expend excessive efforts on reorgs or designing the perfect structure, but to ensure that the organization does not get in the way of doing project work.
Randy Englund,
January 12th, 2007 at 8:11 pm
Thanks to Randy and Paul for their insightful comments.
Though a long-time believer in computational modeling for product development, I was quite the skeptic when first exposed to computational organizational theory. Until my colleagues, Dr’s Ray Leavitt and Mark Nissen exposed me to Jay Galbraith’s premise that all orgs are information processors, with different levels of formalization, centralization, noise, team maturity, etc. My feeling was that we couldn’t model the human behavioral things, and indeed, there is no perfect model. But what has occurred is the further and further validation of SimVision/VDT as a modeling tool for project orgs. We just applied it in the Navy to a projectized F-18 engine overhaul at Lemoore NAS. Though more of a “current operations” outfit than a PMO, theeffort to construct the model made some things apparent, and then the org design changes helped to further streamline an already tightly run operation. From my view, the bottom line is, every org must “task organize” for the project at hand, and we all know that is a dynamic roll-on/roll-off thing, but org modeling can add some particular new insights re: project (decision-making) control versus project risk (of all tasks not being completed). I’m convince it’s a valuable tool.
John