Design of Experiments

Optimal performance of your process is locked inside current performance, just waiting to be discovered. The optimal process can emerge once all the variables are adjusted appropriately. The Design of Experiments (DOE) tool helps align process variables and arrange them to ensure optimal performance.

What is Design of Experiments?

Design of Experiments (DOE) is a Six Sigma tool that helps project teams determine the effects that the inputs of a process have on the final product. DOE helps uncover the critical relationships between variables in a process that are often hidden under all of the data and identifies the most critical inputs that must be modified to ensure optimal process performance. Once Design of Experiments identifies the critical inputs of the process, it helps project teams understand the impact that modifying the variables will have on process performance.

Design of Experiments Terminology

Six Sigma Design of Experiments is a systematic process that breaks down the variables of production and analyzes each one. This process has its own set of terms that we must understand to become conversant with how the technique works.

• Factor – This is an independent variable, or a variable you have control over. In DOE, factors are deliberately modified to determine the point of optimal performance.
• Level – This is a measurement of how much a factor has been modified. Levels can be discrete or numeric.
• Run – An experiment typically done at two or three levels for every factor; each separate level constitutes an experimental run.
• Response – The outcome of the run.
• Replication – Refers to multiple sets of experimental runs. Replication provides even more data and greater confidence in evaluating the results.

How to Apply Design of Experiments

Design of Experiments terminology is more clearly understood when applied to a practical example. Suppose a project team used DOE to optimize the process for baking a cake.

• Factors – The factors in the process, the variables that the team controls, consist of the ingredients of sugar, flower, eggs, water and oil. The oven is also a factor. These are inputs into the process.
• Levels – The levels in the cake baking process are the temperature of the oven, the cooking time and the amount of each ingredient used. These are the potential settings of each factor.
• Response – The cake is the output or the response of the run. The characteristics of the output, cake, are then evaluated  to determine if the levels in this particular run lead to optimal performance, or a cake that ranks well in taste, color, and consistency.

Design of Experiments provides many ways to bake a better cake:

• Comparing alternatives – The team can test the results of using two different types of the same ingredient by keeping other factors the same.
• Reducing variability – Determining how the levels can be changed so that the cake is always the same quality.
• Targeting an output – Deciding what changes to make in the ingredients and amounts of ingredients used that make the cake just right.
• Evaluating tradeoffs – Discovering how to produce the best response (cake) possible by using the simplest factors (the smallest amount of ingredients).

Selecting the Factors

Many inputs determine the output of a process. The factors that are most relevant to the end result are the ones most important to DOE. These factors can be identified by the project team in a brainstorming session. In ordinary circumstances, where time and budget are finite, the team should limit the experiment to six or seven key factors. These factors are controlled by setting them at different levels for each run.

Setting the Levels

Once the factors have been identified, the team must determine the settings at which these factors will be run for the experiment. The example of baking a cake demonstrates that some factors are measured in numbers, such as oven temperature and cooking time. Some factors are also qualitative, such as how much icing to use. These are measured in categories and are converted into coded units for linear regression analysis.

The more levels that are identified for each factor, the more trials will be required to test these levels. To ensure that an optimal number of levels are selected, focus on a range of interest. This range includes settings used in the normal course of operations and may also include settings of more extreme scenarios. The greater the difference in factor levels, the easier it becomes to measure variance.

Evaluating the Response

The response is the outcome of the experiment. Outcomes are helpful in improving the process when they can be measured in quantitative terms, rather than in qualitative attributes. A response that is quantifiable makes the experiment well suited to the additional scrutiny of statistical regression techniques.

Design of experiments allows inputs to be changed to determine how they affect responses. Instead of testing one factor at a time while holding others constant, DOE reveals how interconnected factors respond over a wide range of values, without requiring the testing of all possible values directly. This helps reveal secrets hidden behind the different factors and levels in a process and allows the project team to understand the process much more rapidly.

Once completed, Design of Experiments helps the Six Sigma project team better identify the combination of inputs that lead to the highest-quality product or service.