The Value of Loss Measurement Systems
By Sam McNair, PE, CMRP
I have found that the most valuable weapon in any production facility’s business improvement arsenal is an effective loss measurement system. Every lost opportunity to make a pound of product or a box of widgets should be captured. And if the loss event is above a threshold value (which should decline annually), must have an origination point (an asset or a processing step), and some searchable categorization (such as equipment downtime, lack of materials, etc.) attributed to it. Note that these are not root causes, just the readily observable origin and effect. Root cause comes later.
If you collect this data diligently and honestly (both very hard to do due to human factors issues), and analyze it, this will open the door to a goldmine of opportunities. Obviously losses that are attributed to a specific asset just add justification to your existing maintenance cost-based failure elimination program.
Often the first reaction to a loss capture system proposal is to belabor the difficulty of the technical aspects of capturing this data. This is the wrong problem to fear! There is little excuse for technical limitations in today’s world of plantware, data historians, and process control monitoring systems. The technical part of data gathering is really the easy part. Just don’t waste years developing ultra-sophisticated, complicated, fully automated systems that people don’t use or just misuse. Keep it simple and functional, and involve your people in the system.
The real heartache lies first in getting people to accurately assign an origin and initiating cause, and secondly in getting leadership to not react emotionally (“heads will roll!”) to the truth when it is presented. If leadership reacts inapropriately even once, you can expect to lose up to a year of progress on data collection.
Instead, when you see an unpleasant truth, try this leadership approach I was introduced to as a child. I grew up on a farm. You get dirty working on a farm. My mother once told me that there is no shame in getting dirty, just in staying dirty. Likewise with the data; there is no shame in discovering the truth about where you are, just in not doing anything about it after discovery. This is a prudent approach for management to take as well. Where you are today is not nearly as important as where you want to be tomorrow and having a constructive plan to get there. The act of discovering where you truly are is, in itself, a major step forward.
As a practicing reliability engineer, as an operations person, and now as a consultant, I assure you that there is no higher-return activity available to you than collecting and analyzing loss data. But you must be committed to following up and executing the results of a well-done RCFA based on the data. Unfortunately, good RCFA (and, most especially, the follow up) is the other very hard thing for most companies to do. Unless you know you are excellent at this, get some help in this quarter.
We have all been at places that collect asset failure data and select bad actors based on high cost or low MTBF. And I don’t recommend that you stop doing so. It is an essential part of a balanced program. But the really high return on investment and rapid payback comes more from the production loss data than from the asset repair cost data. Do both and win big.
The reason for this is that a lot of big production loss occurs in many small "bites" that individually are just part of "business as usual". The pain is chronic but ignored. Common sources are procedural issues in the human factors arena or minor stoppages (in discrete manufacturing) and instrument and control issues (in process industries). Yes we all remember that big production line crash a few years ago. But taken together the picture changes and those everyday losses that are common mode, or common cause, often accumulate to huge losses; much bigger than the big incidents we all remember. And the "'little ones” that are actually really big are often the easiest to fix.
If you want to measure payback in months or years, look to your asset maintenance cost data. But even if you can afford the associated maintenance costs, you can seldom ignore the downtime because of its high production cost impact. If, as a reliability engineer, you want to measure payback in days or weeks, look to your production loss data.
© Life Cycle Engineering
