Executive Interviews: Interview with Daniel Levinthal on Learning Organizations
July 2008 - By Dr. Nagendra V Chowdary

Daniel Levinthal
Reginald H Jones Professor of Corporate Strategy at the Wharton School,
University of Pennsylvania, Current Chair of the Management Department at Wharton.

the perturbations in our behavior that underlie the variation in outcomes? They run experiments with an openness to the fact that many, if not most, experiments will fail. If there is an organizational presumption that experiments should succeed and that a failure to access success implies some deficiency in the management team responsible for the effort, there will be little learning. Outcomes will be manipulated to appear more successful than they actually are. Experiments that, ex ante, were sensible initiatives will persist longer than they should and tie up excessive organizational resources.

• All the successful learning organizations seem to be Fortune 500 companies (Southwest Airlines, Intel, GE, WalMart, Cisco, P&G, Charles Schwab, Disney, Dell Computer, Goldman Sachs, Microsoft, Johnson & Johnson, Toyota, Shell, etc) and therefore it seems prohibitively expensive and highly unlikely for SMEs and startups to be learning organizations. Do you agree with this?
  All organizations need to be learning organizations. It is foolhardy to presume that one has it all figured upfront and does not need to take advantage of feedback from experience. The kernel of truth that may underline the contrast you are making between larger and more modest scale enterprises is how organizations with different resource pools need to think about the issue of experimentation, both the risks and extents of experimentation. A larger, more well endowed organization can afford to place multiple "bets". In the limit, a small startup effectively consists of a single "bet". However, that doesn't mean that the underfinanced startup isn't engaged in learning. Its initial conception of its product or market is likely to prove partially right and partially wrong. How such an organization makes sense of these partial failures and successes will prove critical to its ultimate survival and success. So, the degree of parallelism in the process of experimentation may be a function of a company's scale and scope but not the need and inevitability of experimentation.

• Definitely every company would like to be a learning organization. Success would not have smiled at them unfortunately. Why do you think many companies fail in being learning organizations? What are the impediments to being a successful learning organization? What can other companies learn from these failed attempts? Do you agree with the observation that productive failure is better than an unproductive success?
  Learning requires acknowledging failures. That is not an easy thing at the individual level (psychologists refer to the fundamental attribution error as the tendency to attribute unfavorable outcomes to external causes and favorable outcomes to our own actions). It is even more difficult in an organizational context. Feedback tends to be incorporated formally, or informally, to evaluation processes. As a consequence, there is a natural tendency to repress negative outcomes which in turn deprives the organization of the feedback that could enhance future efforts. Learning requires some degree of trust and support among the participants. Fear about one's position, next period's evaluation or compensation will likely repress the possibilities of organizational learning.

1. Learning Organizations Case Study
2. ICMR Case Collection
3. Case Study Volumes