Sunday, 16 December 2012
Why Do We Benchmark Against the Average?
People ask us why we benchmark high performers against the average rather than against low performers? Surely we want as big a contrast as possible?
No, absolutely not. Here's why.
Here is a distribution graph representing performance levels for a given role in a population:
Can you see why we don't cross reference against low performers?
In any given population, we could profile competence against a particular task. What we would always find is a normal distribution, as in the above diagram, because that's a quirk of the statistical method. So low performers aren't necessarily a bunch of losers who need to be fired forthwith; you might say that their performance levels are acceptable, that the average performers are good and the highest performers are outstanding, and you don't know why. The high performers are defined only by their performance, not by whether you like them or think they're really good.
Typically, the midpoint of the curve marks a difference in a person's goal or outcome.
In a customer service example, to the left, the outcome might be to work to the customer service standard procedure. To the right, it might be to delight customers. The goal of the procedure isn't to delight customers, it's to manage all customer interactions equally.
Where high performers always excel is in their ability to adapt.
To the right of the centre line, we can say that all of the staff share the same outcomes, but the average performers haven't adapted their behaviour to the culture to the extent that they achieve maximum performance for minimum effort. We often find that the average performers do everything 'by the book', but following procedures leads to consistency, not excellence. What we need to do is refine excellence and turn it into a procedure, and you'll then see performance levels rise across the organisation.