Design of Experiments (DOE)

Introduction: DOE

Design of Experiments (DOE) is a statistical method used to study the relationship between input factors and output responses in a process or system. It allows organisations to test multiple variables simultaneously, identify key influences, and optimise settings to improve performance and quality.

Background

DOE originated in agricultural research in the early 20th century and was formalised by statistician Sir Ronald A. Fisher. Since then, it has become a core tool in engineering, manufacturing, and scientific research. In Lean Six Sigma, DOE is used to support data-driven decision-making by revealing how factors interact and how variation can be reduced.

Key Elements/Features

• Factorial Experiments: Studying two or more factors simultaneously to capture both individual and interactive effects.
• Randomisation: Assigning conditions randomly to reduce bias from uncontrolled variables.
• Replication: Repeating experiments to measure variability and increase reliability.
• Blocking: Grouping similar conditions to isolate effects and improve accuracy.

Applications/Examples

• Manufacturing: Determining optimal machine settings to reduce defects.
• Pharmaceuticals: Testing drug formulations to balance effectiveness and safety.
• Agriculture: Studying fertiliser levels and crop yields.
• Marketing: Testing combinations of price, packaging, and promotion strategies.

Relevance/Impact

DOE offers a structured approach to experimentation that saves time and resources compared to one-variable-at-a-time testing. It helps organisations:

• Identify key factors and their interactions.
• Optimise processes for quality and efficiency.
• Support innovation through systematic research.

By embedding DOE in process improvement and product development, organisations gain robust insights that drive higher performance and competitiveness.

See also

• Hypothesis Testing
• Full Factorial design
• Lean Six Sigma
• Analysis of Variance (ANOVA)