Tuesday, September 29, 2015

Standard Deviation Revisited

It is interesting to note how mysterious Standard Deviation is to most people.  Not everyone is a math geek, of course, but with a little effort, understanding Standard Deviation and its possible applications can be a true boon to a great many business people.  It is such a useful measure that I feel it is worth examining as we did a little over 2 years ago.

So, what is Standard Deviation?  In basic terms, it is a measure of how widely values are dispersed from the average value.  Depending on your business, it is potentially a very important statistic.

My favorite example as to how this measure can be applied is in the case of Call Centers.  Managers and analysts of call centers routinely track the amount of time it takes for their customer service people to handle the incoming calls.  Other measures of quality are, of course, also very important, but the Average amount of time taken per call is obviously very central to the effective operation of nearly any center.

So let’s suppose that you are such a manager or analyst, and you track weekly results of your department.  Let’s also say your records show that the representatives can provide your customers good service in a Mean Average of 6-8 minutes.  So, what would better to see, a range of actual talk times of 1-65 minutes or a range of 4-10 minutes?  (This is an example of the relative dispersal that Standard Deviation measures.)

Quite clearly, you would prefer to see Smaller values versus Larger values (tighter grouping around the average times), as it would indicate a more uniform, trustworthy approach by your reps.

So how can you use Standard Deviation in Excel? If you are still using Excel 2007 (or an earlier version), using the old STDEVP may be your preferred function. This assumes that you are using the Entire Population of data (therefore, the “P”). If you are using just a Sample of your data, then you will want to use STDEV.

This was made more intuitive in Excel 2010 and Excel 2013, where the function for finding standard deviation for the Population is STDEV.P and the Sample is STDEV.S. (A worthy clarification on the part of Microsoft...)

To make this data more visually accessible, this type of information can be more Effectively Illustrated with a Line Chart or a Bar Chart. In the blink of an eye, you can then see whether your data (talk times in our example) are under control in this regard.

The call center example is, of course, merely one of the virtually countless ways of using Standard Deviation in your Excel reports.  It is just another way that Excel can improve our business lives.  Give it a try some time and find out how simple it is to use this metric to further understand your data.

No comments: