Worker Mistakes – How to End Them
- May 27th, 2010
- Posted in Systems Thinking and Measures
- Write comment
*Reader warning: as this blog post will require some thinking and possibly learning.
When a system is stable, telling the worker about mistakes is only tampering. – W. Edwards Deming
Not long ago, I observed a manager informing workers about their mistakes. Explaining in a calm tome about the types of mistakes that were coming out later in the process. Workers intently listened and there were promises to do better. Later, I asked the manager how often these mistakes occurred. She replied, “all the time.”
- Image via Wikipedia
Although I have seen managers try to end mistakes through training, software edits, mistake-proofing, etc. I rarely find they understand the source of these mistake. A manager and anyone else wanting to reduce mistakes must understand variation. If you do not understand variation you may want to read this post (Service Metrics: What You Need to Understand) to give you some clue before reading on.
Too many managers treat mistakes as if each one is a special cause, when they are being produced predictably by the system (structure, work design, measures, technology, etc.) that they work in. The use of control charts is the ONLY way to know whether the system is stable or not. In service industry I rarely find many special causes with regards to mistakes, when I do I may find a new worker in training or an unusual circumstance that that worker already knows about.
However, when mistakes are predictable (between the limits) and the result of a stable system, more training and/or more communications about them will do no good. In a stable system, the focus on the worker is misguided. Systemic changes are needed to eliminate the mistakes. Wishing, begging, inspecting or imploring workers will do no good.
Systemic issues have to be corrected in the design and management of the work. Too many managers see the system as something they can’t control (too hard to do for department-separated service organizations) and instead turn to the one thing they can control . . . the worker. The worker becomes frustrated, morale falls and business improvement is not achieved.
Systems that workers work in need management . . . workers, not so much. Understanding the nature of these systems and aligning them with customer purpose almost always leads to breakthrough performance and innovation leadership. The only time it doesn’t is when organizations give up.
To improve systems requires experimentation with method (innovation). This requires that we derive measures associated with customer purpose (what matters to customers). All of this requires a different approach to how we handle mistakes.
Leave me a comment. . . share your opinion! Click on comments below.
Make the new decade a profitable and rewarding one, start a new path here. Download free from www.newsystemsthinking.com “Understanding Your Organization as a System” and gain knowledge of systems thinking or contact us about how to get started at [email protected]. Reach him on Twitter at www.twitter.com/TriBabbitt or LinkedIn at www.linkedin.com/in/trippbabbitt.
Hi Tripp
I have a practical example and would like to hear your perspective on our planned course of action: I have analyzed the calls in a telecoms hotline and found that about 30% of calls were due to simple handling mistakes in a previous call. Simple things such as summarizing what has been agreed with the customer, checking combox settings when changing a price plan, checking all address types upon an address change, etc. So I thought it would be appropriate to have a feedback loop, so that agents know about their own handling mistakes which would support learning and continuous improvement.
I guess these handling mistakes are predictable and the result of a stable system. What would you consider ‘working on the system’ in that context? You seem to disagree on the feedback loop for agents. On the other hand, that feedback mechanism could also be achieved by changing the system, not by managers tellling agents about their mistakes.
How would you specifically apply ‘experimentation with method’ to that situation? Would welcome your comments.
Regards, Frank
Frank-
I think you may be missing the point. You should be involved but unless you engage the agents in performaing check you are no better than the command and control crowd. This should be an effort not just by you, this is very command and control thinking. You have little chance of sustainability unless managers and workers are engaged in learning. Otherwise, break out the lean six sigma project plan.
You have told me little about customer purpose, as you will find other measures too . . . if you perform check. The 30% failure demand if predictable should have the question asked to the workers. If they were involved from the beginning this wouldn’t be an issue. Many agents won’t understand predictable vs. unpredictable demand, but they should (and so should managers to that point).
Ultimately, you will need to track down the system conditions (why the system behaves the way it does) that are causing the 30%, this will require the knowledge and context you gain through check.
May I recommend you grab Eliyahu Goldratt’s Theory of Constraints book. He talks about this very situation. BTW, I love Deming!
He talks about involving the people who do the work in the problem solving, how to incite innovation and how to recognize what needs to be changed and what it needs to be changed to. It’s a great guidebook on how to approach a situation like the one you face. The process of defining, analyzing, implementing, measuring and controlling are blended with helpful and thought provoking insights by Goldratt and his enriched experiences.
Use that and Mr Babbitt’s insights on engaging the employee in solving the problem and you may have something!
To Mr. Babbitt-feedback does not always have to be about the mistakes an agent makes. Contrary, it should be based on historical data, collected from multiple and reliable sources, and analyzed for trends. Then you discuss what they are doing well so they keep on doing it! Having discussions that are focused on customer defined performance attributes helps the agent engage and become an owmer of the solution. Through this same discussion you and your agent correct the performance that is inconsistent with the goals.
Coaching coaches to ask the questions you refered to: “is this systemic or is this an individual problem?” and getting them to use their data to make those discriminations is fundamental to the team’s overall success!
Thank you for this discussion.
JoAngela-
I agree that there is much to like about Eli (as I stare at The Goal). Just that he has a very manufacturing bend. Just like any theory we need to see what applies to service . . . something many lean folks struggle with.
I agree feedback doesn’t need to be about the mistakes, but that seems to be a major focus. The data tells me that the system is at fault and the work design is even worse. Until we redesign the work and think differently about causes of costs we will be stuck. Management must learn that costs are in flow and not scale and managing costs increases them.
This is a great post and I agree with; “A manager and anyone else wanting to reduce mistakes must understand variation.”
Without the profound knowledge part of -knowledge about variation- one would have trouble with the method. Especially when the managers involve the workers in helping with innovation to come up with theories to test on how to fix or improve the system and let’s not forget,,, gain knowledge together.
Thanks
Keith