The 10% Precedent

Ten percent is a number that shows up frequently when top management is setting and evaluating performance results for a part or all of an organization.

Here are some examples that can be found in many organizations today:

    * Increase sales by 10%

    * Reduce the accident rate by 10%

    * In crease market share by 10%

    * Remove the bottom 10% performers

    * Address and explain all budget variances in excessive of 10%

    * Improve profits by 10%

    * Reduce labor costs by 10%

Why is this? What is so magical about the number – 10 percent? 

When examined in this fashion, questions surface like: “Why not 11%”?  Or “Why not 15%? Or 25%” If 10% is arbitrary, in that it has no basis in facts about the patterns in performance, why not go for a bigger number? Wouldn’t that be better?

The problem with using numbers like this in setting goals and evaluating performance is that there is no credible support for their use—remember we’ve established they’re based on wishes.  Moreover, it is an indication that the management group is out of touch with the dynamics underlying performance of the organization.  What it is reflective of is those in management not looking at performance from a systems point of view; reflecting a limited perspective that often leads, not to improvement, but to making things worse.

The problem with an arbitrary number like 10% is just that, it’s arbitrary and not grounded in an understanding of the capability of the system.  A closer examination will often reveal that it is either too high or too low, leading some to have to reach down to achieve it and others to sub-optimize the system to attain it—neither of these produces a positive contribution to the organization.

There is No Substitute for Knowledge

As with any measure or outcome, there will be variation in what is observed—outcomes are not always the same.  Hence getting to a more fundamentally sound goal requires the use of both systems thinking and statistical thinking. It requires learning from the pattern of variation in performance measures and relating this (pattern) to the very processes that produced them.  The information presented in the pattern can be transformed into knowledge of the producing process—like reading the gradients in an X-ray—that will inform appropriate action.   

Statistical thinking informs us that variation can be common to the process or special to the process.  The former situation means the performance of the process can be counted on to fall within a six-sigma wide band centered on the average; thus any goal beyond this band is very unlikely without changes to the process itself.  So setting a goal beyond what is likely without a corresponding strategy for its attainment is unavoidably setting others, and the organization, up for failure.  Not only is the attainment unlikely, the time spent trying just adds unnecessary cost.

If the pattern in the variation indicates a special variation situation, then no amount of goal setting—no amount of effort—will improve performance until the process itself is brought to a common pattern of variation situation.  That is, the goal must be to work on the process so that its performance predictably falls within its six-sigma wide band.  Once this is accomplished then the work of reducing the width of the band and/or causing a favorable shift in the average should be the goals.   

There is nothing arbitrary or whimsical about this approach to managing performance. Challenging individuals and teams for improvement under this type of management process typically stimulates productive value-added problem solving.  It is an approach that leads to significant performance gains—for both individuals and the organization.  Thus, it is far better to demonstrate your acuity in goal-setting than to be arbitrary.