Despite widespread recognition that challenging times place unpredictable demands on people and businesses, I still run across many managers who would prefer to avoid the logical conclusion that stems from this: failure is a lot more common in highly uncertain environments than it is in better-understood situations.

Instead of learning from failures, many executives seek to keep them hidden or to pretend that they were all part of a master plan and no big deal. To those executives, let me argue that an extraordinarily valuable corporate resource is being wasted if learning from failures is inhibited.

Naturally, to an executive raised on the concept of "management by exception," any failure at all seems intolerable. This world view is reinforced by the widespread adoption of various quality techniques, for instance, six sigma, in which the goal is to stamp out variations (by definition, failures) in the pursuit of quality.

Managers are supposed to be right, aren't they? And having the right answer is just as valuable in management as it was in grade school, right?

Perhaps not.

For many years, scholars such as my esteemed colleague, Sim Sitkin of Duke, (see his article, "Learning Through Failiure: The Strategy of Small Losses") have been studying how organizations learn, and they have come to the conclusion that intelligent failures are crucial to the process of organizational learning and sense-making.

Failures show you where your assumptions are wrong. Failures demonstrate where future investment would be wasted. And failures can help you identify those among your team with the mettle to persevere and creatively change direction as opposed to pig-headedly charging blindly ahead. Further, failures are about the only way in which an organization can re-set its expectations for the future in any meaningful way.

Not all failures, of course, are going to be useful from a learning point of view. The concept of intelligent failure makes a difference here. Sitkin's criteria for intelligent failures are:

  • They are carefully planned, so that when things go wrong you know why
  • They are genuinely uncertain, so the outcome cannot be known ahead of time
  • They are modest in scale, so that a catastrophe does not result
  • They are managed quickly, so that not too much time elapses between outcome and interpretation
  • Something about what is learned is familiar enough to inform other parts of the business.

?I would add a couple of other criteria:

  • Underlying assumptions are explicitly declared
  • These can be tested at specific checkpoints, identified in advance, since planned results may not be equivalent to outcomes.

If your organization can approach uncertain decisions as experiments and adopt the idea of intelligently failing, so much more can be learned (so much more quickly) than if failures or disappointments are covered up.

So ask yourself: are we genuinely reaping the benefit of the investments we've made in learning under uncertain conditions? Do we have mechanisms in place to benefit from our intelligent failures? And, if not, who might be taking advantage of the knowledge we are depriving ourselves of?

Columbia Business School professor Rita McGrath studies innovation, corporate venturing, and entrepreneurship. Her latest book is "Discovery Driven Growth" (2009).