If a task is performed different ways, the process – and therefore the result – is unpredictable.
![Getting Lean Getting Lean](http://circuitsassembly.com/cms_bu/images/stories/column_icons/lean.gif)
Why should people follow a process? So they can deliver a predictable outcome. And if the outcome is not adequate, we improve it. The right process will deliver the right results. And the right result is a predicable outcome.
But can you create a standard work procedure (for more on this, see “Standard Work Procedures,” May 2008) without first examining the current process? In other words, can you start with invention? If it’s conceptual, you don’t really know until you try it. And once you try, you are creating a process by which the outcome can be examined.
A principle of Lean is anything can be improved. The standard is a tool we use to make problems visible. To understand how to improve, something must be measured. This is done by exploring the weaknesses of the current process to understand the data and the facts. The current process is baselined to understand where it is, where the weaknesses are and if the changes lead to better results.
Precisely what to do depends on the scope of the change. If an operator or office worker can examine their current work and see one piece that can be improved, they can change. That’s a process known as self-evaluation, and it underscores that anyone can participate in continuous improvement. An operator, for example, could make a change to their standard work by repositioning material so it’s easier to reach. It’s a new standard and it’s something they can do themselves. It’s a small, incremental change. They would evaluate how they do things today, and maybe conclude that moving a bin reduces their reach and shaves off a second of time.
For a process with many pieces, however, such as many individuals or an entire value stream, change begins with a cross-functional team that examines the current state. As opposed to a transactional process, for example, the approval process may go through many levels of approvals, and the team may decide there are too many.
Then there’s the issue of whether to use a Design of Experiments. The purpose of a DoE is to confirm the understanding of the problem, to pinpoint the problem that must be resolved. A DoE allows understanding of the part of the process or variable that is having the greatest impact on the outcome. Once this is known, those attempting to solve the problem could focus on that particular variable. In short, the DoE helps focus on where the biggest opportunity is, but it doesn’t play a role in making the actual change. The latter is the Kaizen, the “change for the better.”
Diffusion of standards. In an organization, the greatest challenge is to sustain the standard; the second is to diffuse the standard, so the same lesson isn’t learned over and over again. The larger the company, the more difficult it is to diffuse standards. Even within Toyota, from which many Lean principles are derived, there’s a constant puzzle of how to take the standard work they diffuse in Japan and translate that process to other factories without having to relearn the standard. The difficulty is, How do different workers learn the same lessons? Classically, as we age, we learn by failure. One way to overcome this is to raise awareness of the best ways to do things, to help others on the work level to learn the best ways of doing things and take that knowledge back to their teams, where they could possibly improve the standard.
These “people systems” require maintenance. Taking a message and translating it down to the floor – to ensure it was adopted per its intent – requires checking. That process is called the management standard work. Determining whether the standards are being implemented and upheld is a component of the management standard work. All other tools we use in Lean – making the process visual so you can see the problems; making the process flow in a straight line with known customer-supplier relationships – make it possible to see when there is a problem, so the management system can see the problems at a glance and determine whether a standard exists, if it is being followed, and whether it is adequate. Management needs to be able to do that as quickly as possible because they have a lot of processes to maintain.
Similarly, this process is followed when diffusing a new strategy. One must look at the process by which the strategy is diffused throughout the organization. As changes are made to the strategy, there needs to be a check that the action is delivering the desired results. This all ties into the “Plan, Do, Check, Act” methodology of quality. It’s all interwoven to help management get results.
At its core, Lean is a management system to help problem-solve in a sustainable way. It is very management – or leadership – intensive. It’s actually respectful to check. Standards aren’t used to be punitive, as in “you must do it this way.” In fact, they are team-based: Follow the standard as it is today because we want to know if it works. And if you have ideas to improve it, diffuse those ideas because the goal is to always do it better.
Robert Hemmant is global Lean architect at Celestica (celestica.com). His column appears bimonthly.