This article looks within the Six Sigma DMAIC (Define-Measure-Analyze-Improve-Control) process to examine how it approaches problem prevention. There is only one step in the Classic Six Sigma (without the Lean components) DMAIC process that directly addresses this issue: the Control step. One could argue that the Improve step, once completed, establishes new or improved process elements that prevent future problems. When a team has defined improvements for a broken process, applied the corrective action and proven it to be effective, then the project goes into its final stage, where a host of actions can control the process so that it does not revert to a state of unacceptable performance. In that context, at the tail end of a Six Sigma project, there is an emphasis on preventing the original problem from returning. Additional problems, or the potential for new problems, may be dealt with as well. This is typically due to discoveries made from Failure Modes and Effect Analysis, Fault Tree Analysis, Cause and Effect Analysis, Root Cause Analysis and Ishikawa (Fishbone) Diagramming methods.
This article will explore what top Lean Six Sigma (LSS) experts have to say about the Control step from the DMAIC process. We need to understand these methods and actions and their context before we move on to define and justify a new problem prevention model to include at the end of the DMAIC steps.
The first is from the book, "What is Lean Six Sigma" by Michael George et al. of the George Group.
"The purpose of Control is to make sure that any gains your team makes will last. That means creating procedures and work aids that will help people do their jobs differently from now on. In Control you will...
-
Document the new, improved procedures
-
Train everyone
-
Set up procedures for tracking key "vital signs"
-
Hand off ongoing management to the process owner
-
Complete the project documentation
These actions will help you...
Probably the most common tool in Control is the Control Chart..."
The second is from the book, "General Electric's Six Sigma Revolution: How GE and Others Turned Process into Profits" by George Eckes.
"Control, develop, document and implement a plan to ensure that performance improvement remains at the desired level."
"There are two major types of control at the project level. One is qualitative and the other quantitative. Process standardization refers to the stability of the process steps once the project team has completed its improvement work. ...once solutions have been implemented, the team should have created the "should be" map - the way the process should run once the solutions are implemented".
Mr. Eckes recommends a set of qualitative and quantitative methods for control:
|
Non-statistical controls:
- Checklists
- Schedules
|
Statistical controls:
- X bar & R charts
- Individual & Moving Range charts
- X bar & S charts
- Moving X bar & R charts
- Other types of statistical charts
|
|
Non-statistical controls:
- Periodic status reviews
|
Other types of statistical controls:
- Bar charts
- Pie charts
- Pareto charts
|
"Process throughput - standardization matrix tools."
Mr. Eckes does a nice job of classifying process environments where each of these four control methods applies:
a) Low standardization / Low Throughput Processes
- Periodic status review
b) High Standardization / Low Throughput Processes
- Check lists
- Schedule (such as a Gantt Chart)
c) Low Standardization / High Throughput Processes
- Bar charts
- Pie charts
- Pareto charts
d) High Standardization / High Throughput Processes
- Statistical control charts
Another useful control method Mr. Eckes recommends is the Response Plan.
"A good Response Plan creates an ongoing action plan for the process participants to follow so that that the only change in sigma is positive."
Mr. Eckes recommends a Response Plan table with 6 columns, each headed as follows:
-
Process Map "Should Be"
-
Measures
-
Specifications & Targets
-
Data Collection Methods
-
Control Methods
-
Process Improvement
Key learning from Eckes on the Control step:
-
Control must occur at both the tactical level (project team) and the strategic level (business leadership)
-
The type of tactical control chosen is based on the extent of standardization of the new process and the throughput rate of the new process.
-
The vast majority of processes will have high throughput and high standardization that leads to statistical control charting as the preferred method of tactical control (typically a mfg. process).
-
Statistical control charts can alert the members of the process when the process has changed irrespective of whether defects have been made or not.
-
Becoming aware of assignable causes to the process can act as prevention to defects occurring and also assist the project team to improve sigma performance after a project team has officially disbanded.
The third book we reference to explore the Control step is "Six Sigma Beyond the Factory Floor" by Snee & Hoerl. Dr. Ron Snee, formerly of Dupont, and Dr. Roger Hoerl, formerly of GE, have been prolific educators, consultants and authors in the area of Six Sigma for many years. You would be very hard pressed to find two other people who know more about deploying Six Sigma than these guys. We have selected key points from their view of what comprises the Control phase because they come from real case studies at the end of the DMAIC process when applied to non-manufacturing processes. Their observations of what comprises the Control Phase are summarized as follows:
-
Implementation of a control plan
-
Ensure gains are maintained from the improved process
-
Prevention of new problems
-
Track, estimate & monitoring costs
-
Visibility of process information
-
Communication of new roles & responsibilities
-
Provide training where needed on the new process
-
Initiate a continuous improvement process
-
Produce data on process performance & mechanisms for review & use in the Continuous improvement process
-
Regular management reviews
-
ID of additional opportunities for further improvement
These experts' insights into the Control Phase show us how Six Sigma has characterized problem prevention at the end of a DMAIC problem solving project. This helps us design a proactive process and a set of best practices to apply at the end of the Control step so practitioners have a detailed strategy to prevent problems in the first place. After all, should we not be proactively slowing down the creation of new DMAIC projects?
