Sometimes it is nice to be reminded of stuff you've long forgotten or thought unimportant. This happened with a client a while back.
First off, for 20 years I've been working with companies to help them drive performance with their employees, channels and consumers. I design programs that influence behavior in order to hit business objectives. One of the standard questions asked during the discovery process is, "What is the average (insert performance metric here) of your target audience?" The idea is that my recommendation will move the average (mean) performance of the target audience in the appropriate direction. Sounds good. But it is flawed.
I was reminded of this when the client said after I asked the question - "I don't care about averages."
Averages can be misleading.
Case in point... if you have 100 people in the room and they all make $50,000 per year - the average is $50,000. But if one leaves and Bill Gates takes his/her place, the average income in the room jumps to $700,000,000 (as of today 10:43 a.m. EDT - per Bill Gates Wealth Clock)
If my client had this kind of situation - and I heard the average sales were $700,000,000 and success would be a 10% increase in sales ($70,000,000) I'd be in deep trouble. Most of the folks in the room would need to increase sales by 14,000,000 percent to hit the 10% overall increase!
As the client reminded me - what I should be asking is what is the distribution of sales (or other performance metric.) In other words - I'm more interested in the median of their audience performance. How many people are performing at different intervals. This allows me to see the real issue. What most clients want is a movement in the median - not the mean.
My solution would be to make sure I take care of the top performers and structure the program so that I increase the number of people in the higher metric intervals. I need to be concerned with the number of people who perform at higher levels - not the average of the total group.
