HBS Working Knowledge has a new working paper by Deishin Lee and Eric Van den Steen, Managing Know-How. They look at companies that keep best practices and model employee use of the best practices and company decisions about recording new practices with an economic model. The results are interesting, but the extreme model the authors use gives me some trouble.
For many firms, the ability to create, organize, and disseminate know-how is a key factor in their ability to succeed. But should all companies engage in formal knowledge management? If not, which companies derive most value from a formal knowledge system? Conditional on implementing such a system, should the company focus more on learning from successes or learning from failures? Should such knowledge systems simply capture all experience, or should they be more selective? This paper develops and applies an economic framework to examine these questions. Key concepts include:
- Supporting firms' focus on best practice, information about successes is typically more useful than information about failures. Past successes can guide future successes, while past failures only point out certain pitfalls.
- Recording mediocre know-how can be counter-productive by inefficiently reducing employees' incentive to experiment.
- Larger firms with high turnover potentially gain the most from knowledge systems, but should also be the most selective when encoding information.
[via Allan Engelhardt at Cybaea.net]
The first result strikes me as odd on first reading. I've always been under the impression that failures provide more guidance / information than do successes. But in this context, the authors are looking at the value of what is recorded in a best practice / knowledge system. In this light, failures aren't terribly helpful when someone is looking for an answer of "what should I do?" On the other hand, people developing the next-best-thing need to know more about what works and what doesn't.
The second result is another odd one, but makes much more sense in light of their model. Essentially, if the firm is looking to make big leaps in performance of the operations in question, then preserving small improvements for future use is going could be recording improvements that are within the normal variation of the system. (This is my take on the discussion, not the authors'.) The authors have some interesting comments about the "competency trap" that causes people to repeat the moderately good actions and not seek even better actions that will benefit the firm even more. This suggestion in this result - that firms should consider recording and promoting only "successes" that are demonstrably better than the status quo - should go some way to placating the common criticism of best practices.
The third result appear to be a no-brainer. The larger the firm (with common problems across the firm), the more potential benefit of disseminating this kind of best practice. At the same time, it becomes even more important to ensure that the new practices actually make sense and apply across the firm rather than being locally optimal practices.
The critique I have of the central model (a mathematical representation of the best practice choices employees and the firm makes). The model makes assumptions about best practices that make many people criticize the whole idea of managing and recording best practices - that best practices are clear representations of what people should do in a given situation. It also has a very idealistic view of how people use best practices and how best practices are typically recorded in corporate history. The model simplifies a best practice to a strict action-payoff pair without adding all the nuance that is at the core of the value behind best practice and lessons learned literature. The authors acknowledge some of these and discuss how the model might be modified and where "failure" information could be important.