Beyond averages and medians, general counsel should understand statistical ways to describe the variability and distribution of data. One uses the statistical term called variance.
Let’s say you use ten law firms to prepare and prosecute patents. You know how many patent applications they each handled during 2010. It’s easy to calculate the average production for the firms – divide the total by ten. Next, subtract each firm’s production from the average (a difference also called the deviation) and square that number – multiply the deviation by itself. Thus if the average is 17 and Firm Rees handled 20, the deviation of three squared equals nine The sum of all ten squared deviations is then divided by 10 (the number of data points). That number is the variance.
The larger the variance the wider the range of applications. Note that variance gives more weight to extreme scores because you are squaring them. The problem with the variance figure is that the units, in this example patents prosecuted, is squared. General counsel and chief IP lawyers don’t readily think in terms of a law firm working on 21 applications multiplied by 21 applications! For that reason, another statistical tool stands ready to turn variance into a more easily understood statistical descriptor of variability: standard deviation.