NYU Researchers explore learning from success and failure

In a paper published in draft form in January 2011 ("Reaching and Falling: Why Failure in Exploration differs from Failure in Exploitation"), NYU Stern School professors JP Eggers and Jung-Hyun Suh studied learning from success and failure in a new-product development context. They compared learning in exploration (longer-term, more speculative) vs. exploitation (shorter-term, iterative improvement-based development) among mutual-fund firms over a 40-year period.

By using this frame of reference, Eggers and Suh discovered some fascinating results. Firstly, failure in exploratory new-product development had negative implications--but only for what the authors termed "near-miss" failures. Little or nothing was learned from more profound failures.

The authors generated a different but related conclusion from studying product development via exploitation. In this case, near-miss failures correlated with significant positive learning. More significant failures had negative learning. This result makes sense, as exploitation changes are more small-scale, a step forward that, if wrong, can easily be redone. This "probe-sense-respond" model is frequently used to move tech products from version 1.0 to 3.0, such as Microsoft Windows. And, of course, disastrous exploitative changes (Windows Vista, anyone?) tend to send those responsible scurrying into holes to hide.

Here's how Eggers and Suh summarize some of what they learned about learning from failure:

We suggest that positive organizational learning will result from both failures in exploitative activities (which are more pressing for the organization and must be dealt with) and successes in exploratory activities (which bring important new knowledge to the firm that the organization is more likely to encode and utilize). Meanwhile, significantly less learning will result from successes in exploitative activities (which may be viewed as validating the efficacy of existing processes and thus making additional learning unnecessary) and failures in exploratory activities (which may possess little information that can be assimilated and may be met with retreat from risky opportunities).

Here is James March's seminal paper, "Exploration and Exploitation in Organizational Learning," which provided the framework Eggers and Suh utilized.