top of page

Design of Experiments (DOE)

Process and design optimisation

DOE yield surface model example w empty border
Why do you need DOE?

The first question you probably have is what is design of experiments and why do I need this?

Let's answer the why do I need this first!

If you are looking to optimise your process yields (often the most important item as it impacts your company costs) or whether you want to understand your sensitivity to your manufacturing process (can be pretty much any kind of manufacturing) or physical design parameters then you may have followed the typical practice and simplest option of of changing one variable or factor at a time, right?

So why change this tried and tested method? Well, often there are interactions between processes and design factors and these are not take into account when you change only one factor at a time (OFAT). So, how can we take these interactions into account. Well, we can create experimental designs where more than one variable or factor is varied at a time. This is where design of experiments (DOE) comes into play.

What does DOE do for you?

DOE allows you to quickly identify your key process or design conditions by reducing the number of learning cycles and experimental conditions to get you to your optimised process or design. 


It allows you to create experimental designs where you can add many factors in one or more experimental runs (taking into consideration the size of your process batch size or design variables).

There are many techniques (other than the simple full factorial) to help reduce the number of runs while minimising the loss of information such as fractional factorial, Taguchi, Box-Behnken, D-optimal and others.

This allows you to efficiently identify those key factors that have a significant impact on your process or design allowing you to identify which parameters need greater control or optimisation to improve process stability and improved yield.


This can be done easily using commercial software such as JMP or Minitab. Alternatively, you can do this manually using knowledge of the various methods and some model fitting in Excel (not straightforward but there are tools available online). See the links below for information on the major commercial software options


Minitab DOE planning

DOE significant factors and yield improvement example w empty border v3
DOE process steps 3D w large empty border
How can we help you?

We can help to guide and train you on:

  1. The fundamentals of DOE experimental design.

  2. How to identify and prioritise factor choices.

  3. How to set sensible factor limits to ensure you get useful data from your experiments.

  4. Show you practical examples of how to optimise your experimental design and factor settings.

  5. How to use existing data to understand the impact and assessment of variation on your DOE.

  6. How to set the correct response (output) settings for analysis.

  7. How to analyse the output data, using multi-level variability analysis, model screening, model fitting.

  8. Determine the optimum factor settings and expected variation and if needed, definition of the next set of experiments (e.g. RSM)

Get in Touch

+44 (0)7780667106

Man on the Phone
bottom of page