After a couple of months of being there, my boss told me that the most significant thing I had done was bring the word “Failure” out into daily discussion.  I refused to accept a blip in scrap numbers or equipment failure without investigation, explanation and a response plan.  I was not very popular for not believing that these things just happen.

Failure is a seen as a harsh and negative judgement on people or their efforts.  But until each failure is acknowledged, the process of root causing it, preventing it through a control and fixing the system that allowed it to go unidentified until it occurred, will not start.  That list of activities, which “failure” intiates, sounds like Continuous Improvement.  Instead of just have a CI initiative, let the failures that are costing money direct the CI efforts.

1)  Failure is everywhere.  This means that there are endless opportunities for improvement.

2)  In describing success, quantifying the description makes judging the outcome so easy.  Qualifying success often leaves failure unrecognized and therefore, likely to be repeated.

Most people identify the effects of the failure, instead of the failure itself.  This distinction is one of the most important (and difficult) related to this topic.  The post on the BP Oil Spill provides a good example of the distinction.

This effect due to a rupture in the pipe leading from the well and a control was in place.  The blow out valve (or preventer) is a common, well known piece of equipment to stop a well at its source.  It is simply a gate valve that can be actuated remotely.  It is powerful enough to cut through a steel pipe.

This valve did not actuate when the rig sank and broke the pipe just above the blow out valve.  It was known that only 30% of these valves operated reliably at the depth of this well.  There will be a root cause for this failure which can be controlled in the design modification that will come.

However, given the known effectiveness of the blow out valve (control), a program of testing the valves was absolutely required.  The frequency of the testing would be driven by the results.  If the testing determines that a valve cannot be counted upon, then the well must be capped until an effective is in place.

If there is any question about the ability to actuate the valve remotely, then a manual actuation method must be available.  Then, the robots which have been used to try an cap, cut, stuff and film the well, could have simply closed it with the manual actuator.

It is apparent that like all failures. this one was predictable and preventable.