Redefining Failure

by Stephen Shapiro

Redefining FailureI view life as a series of experiments. When you look at it through this lens, failure means something completely different.

One definition of an experiment is: “A test or investigation, especially one planned to provide evidence for or against a hypothesis.”

The only way an experiment can fail is if you don’t get the evidence.

Even if the evidence proves your hypotheses was wrong, the experiment itself was a huge success.

When you view innovation through the lens of experimentation, it redefines failure.

When developing new ideas, the best approach (especially when there is “market” uncertainty) is to create small experiments that can be scaled over time.

The experiment can give you one of four outcomes:

  1. Our hypothesis was validated by the experiment. Let’s make a larger investment in a larger experiment.
  2. Our original hypothesis was wrong, but we found a different direction that looks promising. Let’s create a new experiment with the new hypothesis.
  3. Our original hypothesis was wrong and we should kill the idea.
  4. Our experiment did not give us enough data to determine whether or not the hypothesis was correct.

Of these four outcomes, only the last one is a failure. With the other three, the experiment was successful. It either confirmed that we are on the right path or it stopped us from making further investments.

The problem with some innovation efforts is that insufficient data is gathered throughout the process. Experimentation is not the mantra.

When you view innovation as a series of experiments, you must make sure that the experiments don’t fail. It is totally fine if your hypotheses are invalid, as long as you determine that early in the experimentation process.

P.S. One book I really like is “The Science of Success” by Charles Koch. The entire book is about how experimentation made Koch Industries one of the most innovative – and successful – companies in history.

Don’t miss an article – Subscribe to our RSS feed and join our Continuous Innovation group!


Stephen ShapiroStephen Shapiro is the author of three books, a popular innovation speaker, and is the Chief Innovation Evangelist for Innocentive, the leader in Open Innovation.

No comments

  1. Thanks Subbu. I’m glad you liked the article.

  2. I agree that business transactions are ALL experiments, even if people don’t realize it. Part of the challenge, however, is that unlike scientific experiments, where all but a few variables can be controlled, business experiments are testing quite a few variables simultaneously: technology, feature set, marketing delivery, individual execution, market segment idiosyncracies, and on and on. Therefore, the data gathered from the results is not as validating as one might like, and certainly not as validating as typical scientific experiments. So, in this light, “failure” in experiments is more common than one would like. I think this is a rich area for future research – how to design business experiments that minimize cost and maximize usefulness of results given the inevitability of multi-variable testing.

  3. RSS Feeds are really very helpful and you could get site and news updates from it.’-,

Leave a Reply