Yes, this post will be primarily a copy of a Harvard Business Review (HBR) article. The article is the point of this blog.
Over my decades in consulting, one of the major requirements in our niche was to insure that our clients received the savings we had forecast that they would achieve in a mutually agreeable manner. In the very beginning, when I started back in 1974, computers were not easily available or accessible as they had to be huge room-size machines and many clients did not even have them. So we used adding machines with paper tape to the shortly thereafter personal hand held-calculators. Personal computers starting with the very first Apple or Radio Shack model proved far superior but had minimal calculating capacity. None were perfect and because of this limitation, a straight line average over a years period was typically used to compare history to current and prove savings had been achieved.
I found well into my career that straight-line bases had fallacies in them, equipment-wise as well as, and not always being reasonably comparable between the two periods we were looking at. Over time, I developed and learned many other techniques from seasonal or earned hour versions, etc. On one project I learned about linear regressions and found this to be very reasonable to use if there was a variety of peaks and valleys in the client's data.
Later on, I learned from a client about multiple regressions and this also worked if there were not only peaks and valleys but product mix issues. So my learning continued as well and I went so far as to learn to use the Monte Carlo analysis.
Recently I came across this article in HBR called "The Flaw of Averages" by Sam Savage. It pretty much explains and illustrates what I am saying above but from a broader perspective. I hope you learn from this article as it is all very true.
This article on Averages is what over the years I eventually learned and helped me to become quite successful in measuring productivity and performance improvement. Enjoy and learn!
Bob Jacobson October 27, 2022