Monday, January 30, 2023

Rookie mistakes in Risk Management

First comes the planning
You've followed all the standard protocols for setting up your risk management program.
You've put slack in your cost estimates, and you've put slack-buffers in your schedule plan.
All good.
Risks have been listed, prioritized, and the minor ones to be 'unmanaged' set aside (minimize distractions)
Otherwise, mitigation planning has been done.
All good.

Now comes execution
There are a lot of ways to screw up risk management. No news there, but ,,,,,
Rookies sometimes do these things, but, of course, you won't:
  • Rookies ask for, or accept, single-point estimates from team leaders, work package managers, or analysts. This is a big mistake!

    Estimates should be given as a range of possibilities. No one works with single-point precision, and no one works without control limits, even in tried-and-true production regimes.

    And you should recognize that 'far-future' estimates are almost always biased optimistically, whereas near-term estimates tend to be neutral or pessimistic.

    Why so? First, "the future will take care of itself; there is always time to get out of trouble". And second, near-term, "we have all the information and "this" is what is going to happen; there is little time to correct matters".

  • Rookies sometimes consume the slack before it's time. What happens is that rookies fall into the trap of "latest start execution" when it comes to schedule; and, in cost management, rookies often put tight controls on last rather than first, or early on. Then, when they need slack, it's already been consumed. Oh crap

    Experience and wisdom always argue for using slack last, hopefully not worse than 'just in time'
  • Rookies fall for the "1%" doctrine. In the so-called "1% doctrine", a very remote but very high impact event or outcome has to be considered as a 'near certainty' because of this risk matrix math: "very very small X very very large = approximately "l" (*).  Or, said another way: "zero X infinity = unity (or 1 or 100%)". 

    Accepting that doctrine leads rookies to spend enormously to prevent the apocalypse event. But actually, 'nearly 0' is a better approximation than 'nearly unity' (in arithmetic, 0 times any finite number is 0)

    What about the 'infinity' argument? Well, actually 'zero x infinity' is at best matter of 'limit theory' for one thing. And that's not easy stuff. But actually, anything times infinity is 'indeterminate' and thus not a workable result for the PMO. (**)

    Put the math aside. Isn't this about risk-managing a 'black swan' event you can actually imagine? Perhaps, but that doesn't change the conclusion that 'nearly 0' is the best value approximation.

(*) In probability statements, "1" is understood to be all possibilities, or a confidence of 100%

(**) But more specifically, general laws of mathematics are not applicable to equations with infinity. It's commonly understood that if you multiply any number with 0, you get 0, but if you multiply "infinity" with 0, you get an indeterminate form, because infinity itself is not determined yet. Our science currently has 7 indeterminate forms; infinity is one of them. 

Of course, the good news is that we've advanced beyond the ancient Romans who have no Roman numeral for zero. It was not considered a number by them.

Like this blog? You'll like my books also! Buy them at any online book retailer!