Churchman's Advice: Experiments for Everyone

Experiments are just conscious, formal ways to gain experience.  

C. West Churchman (1913-2004) observed that “each action [by a manager of a corporation] could be construed as an ‘experiment’ …[and] is an attempt to not only improve the financial status of the corporation, but also to increase the understanding of the way in which the corporation behaves and ought to behave.” (Design of Inquiring Systems, 1971, New York: Basic Books, p. 271.)

Churchman’s advice holds generally for any organization, not just for-profit corporations. Every executive and manager should be conducting experiments to increase understanding.

But by what method?

The basics of experimental design are covered by The Model for Improvement, our friendly universal guide for learning.

Here’s the framework—a good experiment needs a clear aim, explicit measures to assess the results of the experiment, and changes (for example, factor level settings in a two-level factorial design).

You then should outline how you will do the experiment and what you expect to happen—who will do what, how and when; and don’t forget to state your predictions. That’s the Plan.

Next:

Do the experiment, noting surprises along the way;

Study the results and how well you followed the plan;

Act to repeat the experiment, perhaps with specific changes or refinements or abandon one or more of the factors as not useful and run a modified follow-up experiment.

Simple to state, hard to do well. But following the Model for Improvement increases the odds that you and I can design effective experiments and learn to do a better job as experimenters, improving as we go.