Variation

The top performers have less variation in performance too. Variation can be understood by calculating the average productivity and the standard deviation of productivity. Generally, when I use the term standard deviation a persons eyes gloss over. Standard deviation just measures variation. It is common to calculate average productivity rates, and few organizations take the next step and calculate the standard deviation of productivity as well. The higher the standard deviation the more variation and the less consistent the organization. Those organizations with the smallest variation in productivity have the highest productivity rates, and those organizations with the highest variation have the lowest productivity rates¹. If an organization is inconsistent, then it is going to have low productivity rates.

The standard deviation of productivity should be less than the mean productivity rate. The best in class companies have standard deviations of productivity that is less than 20% of its average productivity rate. If the organizational productivity rate is 10 hours per function point, the standard deviation needs to be less than 2 hours per function point. The range of productivity will be from 8 hours per function point up to 12 hours per function point.

The idea of consistency is not unique for software development. Anyone who plays golf understands that consistency is the key to reducing the number of golf strokes. There have been numerous books written that consistency is the key to success in life. Jim Loehr of the Human Performance Institute has written “consistency is the key to success and world class performance.”

¹I measure variation in performance using a 95% confidence interval. The width of the confidence interval measures variation. The confidence interval is based upon the standard deviation, confidence coefficient, and the sample size.