What Can Project Managers Do?

How Better Leadership Can Drive Smarter Organizations

Read the original article here.

As a professor I see a constant stream of systemic failures of the university. My partner sees it every day in their job as an engineer. Nor are my friends who have even the cushiest of corporate jobs safe from systemic mismanagement, describing to me one situation after another that could be right out of an old Dilbert comic. As Taylor Dotson points out in his review, Bent Flyvbjerg and Dan Gardner offer an understandable and attainable solution to a problem that many of us may have  simply assumed to be inherent to large organizations.

(Mis)Aligned Interests

As a professor, it is my main goal, and the goal of most of my colleagues, that students truly learn in our classes. Of course almost none of us have been taught how to be good teachers. And those of us lucky enough to be on the tenure track are judged far more by our research than our teaching. Teaching must therefore be cast aside, at least to some degree, to make way for our own professional development and success.

Meanwhile our students see their degree as more of a certification to get a better job. Good grades serve the same purpose. Almost every student I’ve had has expressed at least a desire to learn, to truly take the time to internalize the material in their classes. But when the shit hits the fan, and the student has to choose between making the easy choice to hit the rubric requirements for an A, or really grappling with the material but getting a B, they go for the A every time. Oftentimes even if that means cheating. Nor do I blame them.

Meanwhile, administrators are trying to think of the university like a business. To them, the goal of the university is to satisfy their customers, which is to say get the most students the most good paying jobs. And they’ll cut corners to do it. I’ve had students miss class because the roof of their dorm slid off. The waitlists for my classes are often as large as the class itself, yet administration only lets me teach them once a year rather than once a semester. I’ve had classes scheduled to be in two separate rooms at the same time for the same class. As long as it doesn’t obviously affect jobs prospects for students, cost cutting knows no bounds!

The problem, as we see from Flyvbjerg, as well as the “high reliability” scholars who came before him, is that goals of the different parts of the organization that is the university aren’t aligned. The professors want one thing, the students want another, and the administration wants yet another. So we all end up working against each other and the whole university is worse off for it. An institution of higher learning where learning ends up secondary to everyone.

Bite Sized Pieces

The Paris subway system is one of the largest transit systems in the world. It has over 300 stations and over 220km of track and is the second busiest system in Europe, 10th in the world. When they decided to start automating the trains in the early 2000s, they thus had a pretty massive undertaking.

But the thing is, they didn’t do it all at once. They started when building line 14 in 1989. Taking advantage of the from scratch construction, and what the city of Leon learned by automating their Line D in 1992, Paris decided they would try to make line 14 fully automated and driverless. It was completed in 1998, thus providing a basis for learning in the inevitable build up to full metro automation.

The conversion of existing lines to full automation started with line 1, both the busiest and most congested line at the time. Their hope was to be able to improve frequencies to one train per 80 seconds and thus add an additional 70k passengers a day. But a fully automated train without a conductor has to operate perfectly. Any problems or mistakes without a human’s ability to solve unexpected errors creatively and on the fly could mean safety hazards for passengers, or massive delays. They had already learned this lesson from the failures of line 14.

So the Paris metro did the work in bite sized pieces. They converted one station at a time, sometimes as slow as adding two new platform doors per night. The trains were, in fact, equipped with everything they would need and the stations fully converted well before the trains went driverless. In order to convince the unions involved in metro operation to not protest the automation efforts, the city guaranteed that no one would lose their job. Operators were either moved to different positions, or their positions simply not filled when they retired. An additional benefit of this strategy, on top of less political opposition, was that drivers could still be present even when trains began to operate autonomously. Conversion began in 2007 and didn’t finish until 2012. They took it very slow and completed automation incrementally. The Paris metro has since also fully automated line 4.

We can compare that to the DC metro, which tried to semi-automate the entire system at around the same time. However after two Red line trains collided in 2009, the DC metro switched back to fully driver operated systems and hasn’t re-automated since. Meanwhile the MTA, New York City’s metro, automated line 7 and it immediately started overshooting stations.

Clearly, then, Flyvbjerg got it right to counsel that we aim for modularity, just like Collingridge counseled to keep unit sizes small and scale up gradually.

Ultra Mega Projects

I’ve been invited to give presentations and participate in open meetings at organizations wondering how to better take on complex projects with tight safety or operational margins. What do you do when it just has to work? I’ve cautioned about the deviations spiral, how to increase the reliability or organizational structures, and what makes technologies more flexible–all strategies to help organizations work smarter. But the questions I’ve always gotten in response tend to sound like: “what should our managers do differently?” and “how do we train our project managers to make this happen?”

To be frank, I’ve often been stumped. I’ve thought, “look, I just told you what the organization needs to look like, what it needs to do, and what sorts of policies tend to make that happen. I’m not even sure that retraining managers is the best way to go about this. I don’t know how to answer your questions!” And instead I just respond with “I don’t know,” because I don’t.

But Flyvbjerg does. His book doesn’t really tell us how to address the biggest problems. Nor does it really add much insight compared to scholars looking at the same problems decades before he did. But what he does do, is answer the questions I get asked all the time by organizations looking to improve: what can project managers do?