Six Sigma Tools

Variety of tools can be used to drive quality improvements within the DMAIC model. Many of these tools have been incorporated into Six Sigma software so that the computer carries out the underlying calculations.

Most can be classified into two categories:
a) Process optimization tools - which enable teams to design more efficient workflows,
b) Statistical analysis tools - which enable teams to analyze data more effectively.

Some of the most important tools:

1) Quality Function Deployment (QFD): The QFD is used to understand customer requirements. The "deployment" part comes from the fact that quality engineers used to be deployed to customer locations to fully understand a customer's needs. Today, a physical deployment might not take place, but the idea behind the tool is still valid. Basically, the QFD identifies customer requirements and rates them on a numerical scale, with higher numbers corresponding to pressing "must-haves" and lower numbers to "nice-to-haves." Then, various design options are listed and rated on their ability to address the customer's needs. Each design option earns a score, and those with high scores become the preferred solutions to pursue.

2) Fishbone Diagrams: In Six Sigma, all outcomes are the result of specific inputs. This cause-and-effect relationship can be clarified using either a fishbone diagram or a cause-and-effect matrix (see below). The fishbone diagram helps identify which input variables should be studied further. The finished diagram looks like a fish skeleton, which is how it earned its name. To create a fishbone diagram, you start with the problem of interest -- the head of the fish. Then you draw in the spine and, coming off the spine, six bones on which to list input variables that affect the problem. Each bone is reserved for a specific category of input variable, as shown below. After listing all input variables in their respective categories, a team of experts analyzes the diagram and identifies two or three input variables that are likely to be the source of the problem.


3) Cause-and-Effect (C&E) Matrix: The C&E matrix is an extension of the fishbone diagram. It helps Six Sigma teams identify, explore and graphically display all the possible causes related to a problem and search for the root cause. The C&E Matrix is typically used in the Measure phase of the DMAIC methodology.
­Failure Modes and Effects Analysis (FMEA): FMEA combats Murphy's Law by identifying ways a new product, process or service might fail. FMEA isn't worried just about issues with the Six Sigma project itself, but with other activities and processes that are related to the project. It's similar to the QFD in how it is set up. First, a list of possible failure scenarios is listed and rated by importance. Then a list of solutions is presented and ranked by how well they address the concerns. This generates scores that enable the team to prioritize things that could go wrong and develop preventative measures targeted at the failure scenarios.

4) T-Test: In Six­ Sigma, you need to be able to establish a confidence level about your measurements. Generally, a larger sample size is desirable when running any test, but sometimes it's not possible. The T-Test helps Six Sigma teams validate test results using sample sizes that range from two to 30 data points.
Control Charts: Statistical process control, or SPC, relies on statistical techniques to monitor and control the variation in processes. The control chart is the primary tool of SPC. Six Sigma teams use control charts to plot the performance of a process on one axis versus time on the other axis. The result is a visual representation of the process with three key components: a center line, an upper control limit and a lower control limit. Control charts are used to monitor variation in a process and determine if the variation falls within normal limits or is variation resulting from a problem or fundamental change in the process.

5) Design of Experiments: When a process is optimized, all inputs are set to deliver the best and most stable output. The trick, of course, is determining what those input settings should be. A design of experiments, or DOE, can help identify the optimum input settings. Performing a DOE can be time-consuming, but the payoffs can be significant. The biggest reward is the insight gained into the process.