This is an example of how an L8 matrix might be used. Here are the typical steps:
 

• Choose the variables you wish to examine. In the following case 6 variables and one interaction.
  
• Choose two levels for each main variable.
  
• Assign variables to the columns of a Taguchi matrix. Appropriate columns are left blank for particular interactions, in this case column C for interaction AB.
  
• Carry out the work according to the +s and -s of each row. Generally, + is the higher level and - is the lower level.
  
• Put the results in the results column.
  
• Work out the effects according to the +s and -s of each column (see below).
  

The results are worked out as follows:

Result of increasing A (Temperature of Mix) from 60 to 70 is (-51-54-53-57+58+62+54+59)/4 = 4.5

Result of increasing B (Coating Temperature) from 45  to 50 is (-51-54+53+57-58-62+54+59)/4 = -0.5

Effect of the interaction AB is (-51-54+53+57+58+62-54-59)/4 = 3.0
The other effects are worked out in a similar manner according to the +s and -s of the table. The overall results table is as follows:

Effects Table

Conclusions
From the above table you would conclude:
· The most important main factors are A and D.
· For best results A and D should be set high (effect of both is positive).
· There is a significant interaction between A and B (shown in column C).
· Careful examination of the results shows that it is best to have one of these variables set high and the other low. Since we have already decided to set A high, it is best to have B set low.
· Since the effects of the other variables are less important, their levels can be decided according to other criteria, such as cost or ease of control.
· You should carry out a run with the following settings to confirm the findings: A+, B-, D+
· Remember that these matrices can sometimes give misleading results. Always check the results in separate experiments.