Big Waves and Black Swans: Project Management Lessons From TEPCO, Part 2
This article is part 2 of a four-part series on project management lessons from the tsunami that damaged the TEPCO Fukushima Dai-ichi nuclear complex.
Yesterday, I noted three big project-management errors in their risk analysis. They missed a big risk, they didn’t account for single-point-of-failure issues, and they didn’t review the risk plan regularly. Let’s look at the missed risk.
According to the AP report of the tsunami plan, they tied the risk analysis to a Japanese earthquake no larger than magnitude 8.6 and a resulting wave that reached no higher than 19 feet.
The actual earthquake measured magnitude 9.0, which is roughly four times the power of an 8.6 quake.1
The water rose not 19 feet but 46 feet.
Nassim Nicholas Taleb wrote a book called The Black Swan about these types of mistakes, using the US housing crisis as a key example. Too many people assumed housing prices would never decline significantly… because they hadn’t done so in recent memory. Likewise, there hadn’t been a monstrous tsunami in recent memory in Japan… but indeed there had been one about 1150 years ago. Obviously that’s not recent history by anything other than geologic time… but it also indicates that such an event wasn’t unthinkable.
However, it’s a fact of project management life that people on a project miss black swan risks all the time.
And not all black swan risks are historical problems. In the Zubulake case, for example, UBS Warburg didn’t have a lot of history to draw on to suggest that e-discovery was about to become a breakout industry in the legal world.
So what’s a project manager to do? How do you solve this problem?
You can do three things to minimize the likelihood of being blindsided by black swans.
1. Team-Driven Risk Analysis
First, multiple members of the project team, from a wide variety of backgrounds and disciplines, should take part in risk identification and analysis. It must be a team exercise. On most legal projects, that should include the whole team. These meetings become unwieldy when you get above, say, ten participants, but few legal projects have that many people attached to them in any real depth. When you do pick and choose, make sure to get a broad cross-section; avoid the dangers of groupthink. Read James Surowiecki’s The Wisdom of Crowds for another aspect of group contributions to problem solving.
When someone suggests a risk, put it on the risk sheet. It’s like brainstorming; no ideas are rejected out of hand.
2. Team Risk Reviews
Second, review the risk sheet regularly with the team. For a legal project, intervals of one or two weeks are appropriate; don’t let more time than that pass without a team risk review — which can be a very short segment of a meeting for some other purpose. The team should be familiar with the risk sheet, of course, not seeing it fresh at the meeting. Ask if there’s anything new, anything changed, or anything on which they want to reflect.
In the TEPCO case, there was new information discovered that should have caused a rethinking of the tsunami risk — clear evidence of an older massive quake and wave, plus information that two of the tectonic plates off the coast of Japan were “stuck,” meaning that pressure was building up that would likely be relieved in a dramatic rupture.2
3. Use Economics
Every risk should have attached to it a cost and a probability, with the probability calculated as “T-shirt sizes”: small, medium, large, extra large. We rarely know the true likelihood of a particular risk.
Black swans fall into a particular category of low-probability/high-cost risks. Multiplying a very small number (probability) by a very high one (cost) yields an almost random answer. Small variations in the small number cause large fluctuations in the total, yet we can rarely predict the small number with any accuracy at all. Indeed, the large number may be equally unpredictable; how do you really calculate the full cost of what happened in Japan, for example?
In these cases, you need to consider also the cost of mitigating the risk. (Mitigation, in project management rather than legal speak, means to eliminate the risk, minimize its likelihood, and/or minimize its impact or cost.)
Consider together the exposure, or probability times cost; the mitigation cost; and the risk tolerance of the client (in this case, the citizens near the plant in addition to TEPCO itself). What would it have cost to mitigate the tsunami risk?
Obviously, building a 50-foot-tall seawall would likely have been an overwhelming cost (and would have had a rather large impact itself on the citizenry). However, the project manager’s job wasn’t to stop a tsunami; it was to keep supplying power to the reactors even should some event, such as a tsunami, take down the nuclear plant’s main source of power. Here’s where the missed risk played out. What would it have cost to put the generators atop a 75-foot structure? They even had a number of 75-foot-high structures already on the grounds….
Next up, the single point of failure.
1It’s a logarithmic scale, which means that each number is ten times the power of of the previous number. In other words, a magnitude-9.0 quake would be ten times as powerful as a magnitude-8.0 quake.
2I live in an earthquake zone (Seattle) where we’re going through the same rethinking. In the past 20 years, we’ve learned some “interesting” new facts. 1) There was a megaquake 300 years ago roughly the same magnitude as the recent Japan temblor that dropped an area of Seattle 25 feet — an area about three miles from where I’m sitting writing this, by the way. 2) Our own fault line may be “stuck.” 3) Tsunamis aren’t the big issue, because Seattle is not on the ocean but on a narrow body of water that empties into the Pacific about 80 miles away; however, few of our buildings were designed with the potential of a 9.0 megaquake in mind. The point is that if you live in an earthquake zone and are at all interested in the world around you, you’re aware of new findings. If your job involves structures — or nuclear power plant risks — I cannot imagine how you could shrug off these issues without at least raising them.