mardi 24 juin 2008

The Six Sigma Methodology – Five-Step Problem Solving Model


One of the more common models used in process improvement is the DMAIC model (see Figure ). The DMAIC model stands for “define, measure, analyze, improve and control.” Within every step, there is a variety of tools used to ensure you are working toward the best improvement recommendation(s). Generally, after the completion of each step, the information is reviewed with the sponsor of the project.
Define
Specifically, what are you trying to improve? In the Define step, you must identify your key performance indicators. Every team should have one or two key measures. These are measures which are important as seen through the customer’s eyes. Customers generally want things done correctly and on time. If something is broken, they want it fixed quickly. Key measures are all about what is important to the customer. Supporting process measures should also be identified. The path to improving the key measures is through the process measures. Process measures can be viewed as “sub-measures. For example, if a key measure is overall cycle time, process measures might include the various interval times within the overall process. In addition, the team should spend time on operational definitions so everyone knows exactly what is to be measured.
Measure
The Measure step requires spending time improving the accuracy of the measurement system. Verifying the data is a true depiction of what is actually happening in the process is essential. Unfortunately, this is a step that is often skipped or taken for granted. Remember, if the data is questioned or is unreliable, the credibility of the team’s recommendations is at stake. Data is then gathered on each of the measures identified in the Define step. Trend charts should be created for each metric using either daily or weekly data. These charts show how the process is performing over time.
Analyze
In the Analyze step, the team uses various statistical tools to help discover clues for improvement, thus reducing potential distracting emotional bias. In this step, root cause analysis is conducted on outlier points. Data is sliced, diced and segmented to pinpoint areas of concern.
Analysis is performed to determine what drives results. Essentially, the team is looking for clues in this step, clues which will help the team focus on the areas most important to improving results. Also at this stage, a determination is made on the capability of the current process. In other words, are results good enough? Is it worth the time and resources for the team to continue?
Improve
The purpose of the Improve step is to verify what changes will actually improve results before you implement those changes. The key here is that you have sound statistical evidence of improvement. This can be achieved through the use of various statistical tools and experiments. During the Improve stage, the team will brainstorm ideas for improvement based on the clues learned from previous steps and then test those ideas to make sure they indeed help.
Improvements are thus believable and can easily be conveyed to your stakeholders. A common mistake teams make is implementing ideas they “think... will improve results without supporting data to verify those ideas truly will make things better. Implementing improvement ideas based on “assumptions ... can best be categorized as random acts of hope. This is a hard concept for professionals and managers to grasp. But, most processes are quite complex and are influenced by a multitude of variables. For example, if a team identifies just ten process changes, this would result in 1,024 different possible combinations of those new ideas. Twenty ideas results in 1,048,576 combinations. Again, even if people think they know the optimum combination of ideas, the odds are against them. It goes without saying that it is generally hard to improve results. Obviously, the use of sound statistics to verify what indeed will improve results (i.e. move the key measures) is imperative in the Improve step.
Control
The Control step ensures improvements are sustained once the team disengages. Maintaining the gains, by building monitor and control plans, ensure long term success. The Control step is vital. There is a tendency for people to go back to doing things they way they’ve always done them so a proper control plan is essential. Training, tracking, and effective communication plans are all aspects of the Control step. A memo from a VP declaring that everyone will “follow the new process.. is not good enough. The team must plan for monitoring and tracking of adherence to the new procedures on a going forward basis. Often times not enough emphasis is placed on the control aspects of process improvement and this is a main reason why gains are not maintained. Notice there was no emphasis on “improving.. things in the first three steps. It is extremely important to move through each step, in order, before changing the process or implementing new ideas. A classic mistake teams often make is trying to “improve.. results right from the start. If it was so easy to improve, there would be little need for teams.
Finally, a good idea for any team to consider is to utilize the available resources of an internal process improvement consultant. A “third party.. process improvement consultant, trained in the DMAIC methodology, can offer an objective point of view and can help a team move through the Six Sigma steps quickly and correctly, thus greatly improving the odds of team success. Most teams have identified roles and team members including: Team Sponsor, Team Leader, Project Manager, Team Members, Systems Support Manager, etc. Adding a trained Six Sigma expert to the essential list of team members is something every project manager should consider going forward.
The proper application of Six Sigma tools, in conjunction with proven Project Management techniques, can be a powerful combination for any Project Manager.