10 per cent statistics and 90 per cent management action - Are we in control? Are we capable? Will we continue to be in control? Have we done the job right? Could we do the job better?

Statistical Process Control (SPC) has been used mainly in manufacturing industries, although it does have some application in service industries. 

Principles

SPC recognises that not all items produced will be the same.  There is always going to be some variation around the output of a process with a minimum, maximum and average.  If we plot the dimensions on the item against the number obtained it is likely that we will see the following 'normal' distribution curve:

The main principle is therefore, 'have we made it ok?'  This involves asking:

• Are we in control?  Do we have consistency?
• Are we capable? Are we able to produce conforming product all the time?
• Can we continue to produce conforming products? Do we continue to be in control?
• Have we done the job right?
• Could we do the job better?

Approach

Establish control charts for the output being measured.  Take five parts from production at regular intervals.  Measure the parts and calculate the average and the 'range', which is the largest minus the smallest reading.

Plot the results on the control chart.  The central area in the example control chart below represents results where the process is in control.  The amber areas provide a warning that the process may be going 'out of control' and the red areas show where the process is 'out of control'.  This indicates that products are likely to be out of specification.

Under normal circumstances, the results will vary due to small process changes, differences in materials and errors in measurement.  These are said to be due to 'Common causes of variation'.

Sometimes there will be a major change in the process, likely to result in the process going out of control and register in the outer area of the control chart.  Such changes are said to be due to 'Special causes'.