Title: Failing to Learn? The Effects of Failure and Success on Organizational Learning in the Global Orbital Launch Vehicle Industry
Authors: Peter M. Madsen (Brigham Young University) and Vinit M. Desai (University of Colorado at Denver)
Publisher: Academy of Management Journal, vol. 53, no. 3
Date Published: June 2010
When the space shuttle Columbia disaster occurred in January 2003, NASA’s response to the horrific accident that killed seven crew members was immediate and exhaustive. Within months, a special investigative commission had sifted through 30,000 documents, conducted more than 200 interviews, and produced a 4,000-page report on its findings. The panel identified the cause of the explosion and recommended 29 specific changes to NASA operations, communications, and management. This stands in contrast to the minimal investigation NASA conducted the previous year after the space shuttle Atlantis lost a piece of foam insulation during its launch, the same event that triggered the Columbia explosion. Why such different organizational responses? This paper offers a simple reason: The Atlantis mission was a success, and Columbia was a catastrophe.
To determine whether organizations learn more from success or failure, the authors focused on the orbital launch industry — which includes both commercial satellite operators and government space programs — because the field is relatively young, the differences between success and failure are so clear, and failed launch attempts are not uncommon. In all, they reviewed more than 4,600 launches across nine countries from 1957 to 2004 and found that organizations that had experienced failures (of which there were 426 during the period studied) were subsequently less likely to suffer failures than companies that had no failures on record. The results are based on a logistic regression model used by the authors to measure the likelihood of future success and failure. Organizations also retained knowledge gleaned from failure at a far better rate: They held on to 89 percent of the knowledge they gained from failures one year later, while retaining only 34 percent of the knowledge they gained from successful events. The researchers evaluated retained knowledge by adding what is called a depreciation parameter to their model, a complex formulation used in academia to measure knowledge loss. The desire to fix errors, document the reasons for failure, and assign accountability helped organizations codify those lessons and use them to inform future choices. The results suggest that rather than learning from their successes, organizations may be more likely to become complacent and prone to future mishaps.
The authors argue that organizations that stigmatize failure, either by turning those responsible into scapegoats or by trying to shift the blame elsewhere, may severely hamper the potential for future organizational improvement. Their research also shows that organizations do a poor job of unearthing new knowledge from near misses or small failures.
Bottom Line: Organizations that learn from their failures are more likely to improve future performance, but organizational success does not produce similar gains.
- Matt Palmquist was a founding staff writer and is currently a contributing editor at Miller-McCune magazine. Formerly, he was an award-winning feature writer for the San Francisco–based SF Weekly.