Finding the important X's and sampling data

Our next task is to reduce all the inputs that we’ve identified to just those few that impact variability the most. We can start to strip down the list of requirements by utilizing observation & recording the quantitative findings. For example, the actual variability in time to complete, the differences between Ben and Sue at processing orders and the variance in product dimensions at each stage of the process.

We can take sample data from a process by:

  • Selecting a continuous data measure (dimensions, dollars, time).
  • Analysis of historical values & historical variance.
  • Definition of the unit that we’re going to monitor. This must be something we can monitor through different parts of the process.
  • We should then take 2 to 5 measurements for each unit along the different stages of the process. We should do this for 3-5 consecutive units.
  • After time has passed, we should repeat the previous step.
  • We should then create a multi-vari chart.

The multi-vari chart below was drawn by hand and is purely for illustration. Apologies for the mess! Each vertical block shows the time in which we tested. So we may have completed a test in the morning; another at lunch time and another in the afternoon.

The plots follow a particular product along the process. So, the red line shows the 6 check points for that product. From this, we may be able to derive a particular area in the process that is building in a large amount of the variance.  We can enhance this chart further by taking the mean for all observations in each time period and use a dotted line to map the average between each time period.

With mean and average line added:


Content based on study of the Six Sigma Black Belt course and Six Sigma for Dummies