First Thoughts on Control

This series of notes is about control. It will begin with the basic rules for simple systems and move to applications in more complex circumstances. I was an undergraduate when I first started watching crews work using timelapse photography. What I saw through the lens didn't line up with what I was being taught or was being told by managers in corner offices. It wasn't obvious to me why work happened the way it did. My interest what caused work to be the way it was, what controlled it grew. The disconnect between observed practice and explanations appeared to be a real opportunity. Following intuition, I became more and more convinced we were on to something before the now obvious underlying concepts were in place.

Early enough we were thinking about the systems, teams and the processes involved in getting work done. Goldratt taught us about dependence and variation and we could see that work planning was unreliable. Glenn Ballard coined the term, Last Planner, giving us a different perspective, a new unit of analysis. And then I read "Out of Control" - a powerful book with a clever title, by Kevin Kelly. He argues that we need to reconsider how nature controls complex biologic systems. The sharp line between things that are born and things that are made is becoming very blurry. We are now applying techniques used in making things to that which is born — genetic engineering, and making things almost as complex as that which are born — the internet. Projects, particularly complex fast moving ones, seem to be near this line and pose real challenges for control.

Kevin traces the development of modern control to Norbert Weiner. Young Norbert was raised to be a genius. His father wanted a Nobel Prize in the family. His work provided the foundation for cybernetics. It started when he joined the effort to make uniformly thick sheet metal. The stakes were high because uniform feedstock would allow machines to make car parts — fenders, auto bodies etc. Three variables were obviously involved - the heat of the metal, the pull tension and the roller setting. But no one had found the magic relationship that could assure a long uniformly thick roll. Norbert cut through the problem. He showed with a simple mechanism that controlled the last variable produced the desired result as long as the other variables were within in a range. Being "right" didn’t matter as much as the ability to adjust quickly. His device detected the thickness of the metal just past the rollers adjusted them in response. Controlling the last variable with rapid feedback was all that was required as long as the other variables were within some range.

I was struck by how well the Last Planner System™ conformed to this rule. Work is made ready as it moves forward to action and only released to certain criteria. Control of the planning system is applied just as in Norbert's sheet metal manufacturing process. Detectors just downstream provide feedback to make performance more reliable. Feedback in both systems is very quick. The sensing and adjustment mechanisms are tightly linked. Adjustment is so quick the product stays within tolerance.

There is more to say about control of more complex systems but the first rules are clear. Control the last variable while keeping the others within some range, and tightly couple the detection of variance to the control gate. A system that follows these rules will produce uniform sheet metal and predictable release of work from one crew to the next.

Related Posts

• Why Are So Many Projects Tragedies?
Laurent Bossavit answers that question in his posting The Trouble with Projects. Laurent starts with the question, "W...
 
• Explore a PM’s Brain
Raven shares her curiosity and interests at Raven's Brain as she explores the world of project management. Explore with...
 
• Set It and Forget It? Hardly!
In Koskela's and Howell's (K&H) paper The Underlying Theory of Project Management is Obsolete they describe thermo...
 
• Comparing Lean and Agile
Scott Reynolds writing at Tales from the SharpSide, "Lean forces you into a culture of quality. Agile is about interacti...
 
• Preview of Linguistic Action Foundations of Lean Construction
I've written about linguistic action as a theory that does a better job at explaining why projects succeed and why they ...
 

This entry was posted on Wednesday, October 2nd, 2002 at 10:00 am and is tagged: PM practice. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed. Convert this post to a PDF document.