Wednesday, January 11, 2012

Planning: Maximum Acceptable Rate Of Failure.

Share this ARTICLE with your colleagues on LinkedIn .

I am a fan of Mark Joyner and his organization (which is actually a school of thought), Simpleology. He teaches by using the examples and quotations of persons of great accomplishment. Project planning and management have a great deal to do with the notions of probability and efficiency -- the picture below, culled from Mark's collection, is a great way in which to begin our discussion of navigating the Laws Of Probability - with two different approaches.

One approach focuses on precision, refinement and utter dependability (a qualitative approach), while the second focuses on redundancy and back-up systems (a quantitative approach):

For more information on Simpleology, visit Mark's site at

In planning any business, project, operation (military, civilian or surgical), we, as analysts and managers are trying to achieve the most efficient acceptable result. The standards for "acceptable result" are specific, however to each and every individual undertaking. And it not a typical Project Manager or Senior Systems Manager's techno-choice; it is a critical management decision which will determine how we will deal with probability, risk mitigation, resource constraints and other variables at the very outset of the project.

We have to ask ourselves some question's about the very objectives of any mission:

1) What is the maximum acceptable rate of failure or loss given our working resources to solve the problem? The issue of "acceptable losses" is a horrifying one, and otherwise mature businesspeople cower like children when they are forced to address this with transparency.

2) What will be the cost of that level of failure?

3) What is the probability of occurrence of this maximum acceptable rate of failure, given our processes and plans?

4) What can we do to mitigate the risk of exceeding the maximum acceptable rate of failure? Intuitively, the answer comes back to redundancy, manual overrides, multiple hulls or levels, alarms, vigilant monitoring systems, SWAT team on standby, backup 'failsafe' systems and the like. The only alternative is to invest in developing a far better system in the first place.

5) Once we have implemented the above-described mitigation mechanisms, will we have adjusted the subject system or project's "net failure rate" (the rate of unacceptable types of failure involving lost assets and too much lost time) to fall within maximum acceptable failure rate parameters.

In sum, it comes down to quality versus quantity. A wonderful example is choosing a miserably-built but inexpensive car with a ten-year complete repair and maintenance warranty, versus choosing a very well-engineered automobile which is a great deal more expensive.  In the latter situation, it is interesting to note that the manufacturer invariably invested much more in designing, troubleshooting and pre-emptively solving potential problems than the designer of the cheaper one which sells based upon its extended, all-inclusive warranty rather than on its reliability. 

The Maximum Acceptable Rate Of Failure can be arrived at by either a qualitative-focused approach or a "probabalistically-adjusted" or compensatory approach. The question which is begging to be asked (if questions can actually "beg"), but which is incredibly difficult to answer is this: "What combination of quality in design and quantitative 'safety-netting' is optimal to achieve our desired result?

"Engineer for reliability, but be  prepared for failure."  -- Douglas E. Castle

Share this page

No comments:

Post a Comment

View DOUGLAS E. CASTLE's profile on LinkedIn
Douglas E. Castle
Bookmark and Share