Is the temperature in the room where you sit with your computer always
exactly the same? Does it always take the same amount of time to drive
to work? Does the sun rise and set at the same time each day? Hardly.
All those things vary every day, right? That is the way the world is;
the world is variable. Since the 19th century, statistical methods and
measures have been the preferred and most accurate way to describe
process variability. However, right down to this day, many business
leaders summarize their businesses performance using non-statistical
measures, failing to take variability into account. The belief that any
single number (for example, last quarter's profits) actually represents
business reality is imprecise at best and at worse can lead to
disastrously wrong predictions. Too many leaders try to manage in the
21st century with 18th century thinking. New leaders must not make this
In many ways, leaders who disregard variability act like the
psychologist B.F. Skinner's experimental pigeons. In a famous
experiment, Skinner introduced pellets of food at random into a box
filled with pigeons. Inside the box, pigeons were moving around,
somewhat at random too. The pigeons foolishly associated whatever
movements they happened to be making at the time they got fed with
receiving the food pellets; The pigeons started 'thinking' that they
could influence whether they got fed or not by how they acted, and
began to make certain movements over and over in an attempt to get more
food. If the pigeons' heads were rocking when the food pellet first
appeared, they rocked their heads. If they were turning around when
food came they turned around. Of course, these movements had nothing to
do with whether or not the pigeons got fed (the food was introduced at
random), but the pigeons didn't know that.
Like the pigeons described above, leaders who fail to recognize the
importance of variation often leap to conclusions with too little data.
Such leaders may may other mistakes as well:
- They may see trends when no trends are present.
- They may miss real trends.
- They may blame individuals or groups who have no real influence.
- They may give credit to individuals or groups who were simply lucky.
- They may launch disruptive programs for improvement when no improvement is necessary.
- They may fail to identify areas where improvement is needed.
In order to make predictions about the future, it is necessary to know
something about the past. Could you predict the temperature of your
room three hours and twenty minutes from now, based on one measure,
such as the temperature right now? Can you specify to the second, or
even the minute, how long your drive to work will take tomorrow based
on the time it took this morning? Getting either prediction right would
be unlikely unless you were able to bring knowledge of temperature and
drive-time variations (based on your observation of how these things
had occurred in the past) to bear on your prediction. Can you predict
what next quarter's profit will look like based on last quarter's
results? Not without knowing a lot about how profits have looked in the
past, and then not even very well, as entirely new things may happen
that the past could not help you to anticipate.
You cannot understand or predict your business numbers without first
understanding the variation that influences those numbers. New leaders
must pay as much attention to the statistics of variation as to the
profit-and-loss statement. In any business, key processes determine
whether customers will pay for products or services. The new leader
must be a wise student of these processes rather than a superstitious
pigeon. If, for example, a important business process requires a
particular response, and the interval between a customer's request and
receipt of the response determines customer satisfaction, then the new
leader must statistically study variation in response time, and in
customer satisfaction over a number of contacts well enough to learn
whether the business process is working (resulting in an average
reasonable customer satisfaction) or requires change. The new leader
cannot know when changes need to be made without first studying the
business process variation. Superstition costs money!