Operations & Management
Lean Insights: Lean and performance management
November 18, 2009 By Dr. Timothy Hill
Toyota doesn’t use performance appraisal, but they do performance management. How? They use the before-after information from simple kaizens and A3s to set the processes they want to improve.
A kaizen (it’s a simple fix, just go ahead and do it) or an A3 (the sheet of paper used to plan out larger kaizens) are both continuous improvement. There will be a gap to close and measurements taken to assess its initial size. The continuous improvement will be aimed at closing that gap. Ideally, it will be closed enough so that the problem goes away — you eliminate, not manage, your problems.
You track those measurements and put into place visual displays (charts, graphs, andons, etc.) to display your progress. As progress is made and the continuous improvement efforts take hold, you set a new standard for that task. You deliver continuous improvement so that you can set a new standard.
It’s the process of creating new standard work and displaying progress that form performance management.
It’s not using performance management to:
• assess progress toward under-specified goals;
• address behaviour issues that should have been resolved by many trips to the gemba (the actual place where work is done); and
• assess performance towards under-defined work standards.
In fact, William Deming (one of the fathers of the quality movement) said of this form of performance management that it forced half of the people to be sub-standard. He said to throw out traditional performance management.
He was right. More than 75 percent of all North American firms don’t do performance management (PM). Those 25 percent that do tend to use PM data for at least four different reasons. There are problems with how people rate themselves and how others rate the target person. Forced distributions, grading on a curve, central tendency errors and more all combine to render traditional PM literally useless.
How people rate themselves: The most popular PM measure is a narrative form wherein the rater describes their own work. This is subject to over inflation when rating one’s own work.
How others rate the target person: Given that the rater has not been going to gemba on a frequent basis, they will be forced to try to remember how they think the rate did. There are many rating biases that come into play here.
Forced distribution: Based on the incorrect belief that performance should be normally distributed, a forced distribution has so many As, Bs, Cs and the like. If the number of As has already been met, but you have an A for your performance, you will receive a B rating. This is similar to rating on a curve, or “belling” the ratings.
Central tendency errors: Not wanting to give too high (or low) a score can lead to a rater giving a middle rating. Sometimes the rater wants to avoid giving too high a rating because there is no more room in a pay-for-performance scale; the rater won’t know where to place them next year; and so on. Since raters really dislike giving negative ratings, they bump them up to a middle-most rating. Central tendency errors confound the rating system, with repercussions for advancement, career counselling, termination or remuneration.
By concentrating on improving standard work through continuous improvement, Toyota and others can use the before-after values, make those values into visual displays (charts, tables, etc.) and connect accountability to those who brought the idea forward.
The before-after kaizen improvements are the criterion measure for performance management. As each kaizen is implemented, the before-after data is gathered and tracked. This tracking is displayed visually and serves as the basis for shifting from kaizen to standard work. As the new standard is implemented, checks on the performance are conducted.
A person can pull the andon cord, a poka-yoke can be implemented to ensure consistent performance or a piece of autonomation (“intelligent automation” or “automation with a human touch”) can be used to test the conditions of the new standard work. Regardless of the method, the checks on performance are conducted. In this way, performance management is realized.
The cycle of kaizen improvement leads to buy-in from employees. The accountability loop is closed by having the criterion measure from the kaizen event being tracked. As the transition to the new standard takes place, autonomation makes sure that the standard works within tolerance. Note that an emphasis on standard work and a buy-in to lean area required.
Voila — performance management without a performance appraisal!
Dr. Timothy Hill is an Industrial and Organizational Psychologist and Certified Lean Six Sigma Black Belt with global expertise in Human Resources/Human Capital. He can be reached at email@example.com.
The Birth of Lean is a must for people who want the story behind the story of the figures who shaped Toyota management. The standard histories of Lean and the Toyota Production System (TPS) have overlooked the trial and error that went into TPS. Toyota has unquestionably displayed a tenacious consistency over 70 years — often giving the impression that progress was linear. There was a range of confusion and confrontation and the occasional crisis.
Reading the translated first person accounts drives home the notion that the individuals who shaped TPS from the 1940s to ’70s brings home the fact that the path was not always smooth. In the developing days of TPS, Toyota was struggling to survive.
The interviews also tell the story of survival during that struggle — a lesson that should be learned as the North American manufacturing sector struggles with the current environment. The challenges to get leaders to concentrate on the gemba, the shift to pull production, the elimination of waste and more are told here.
Question: How many improvement items should I have in my A3?
Answer: Ideally one. The Sobek and Smalley book (Understanding A3 Thinking) presents an A3 with several improvement items and lists them with the names of people accountable for doing them, the due dates, etc. The book is great, I recommend it along with Shook’s Managing to Learn, another good A3 book.
However, I prefer changing one item at a time so as not to interfere with the impact of other interventions. Not only is this a good experimental design to follow (i.e., you change one thing, look for the change in the outcome) but it prevents you from becoming confounded from having too many variables change at once.
By addressing one item at a time, you can follow the effect that comes from changing it. By addressing a few items and changing them in a given sequence, you’re still closer to good experimental design. If you’re going to add more than one item to change, be sure to:
• Track the people who are assigned to implementing each task,
• Make sure to address the potential root causes with the action items
• Identify who will be implementing the countermeasures
• Make it clear about what exactly will be done
• Make clear the due date on which the actions will be completed
• Make the implementation order and location clear