For years, managers have been urged to make knowledge sharing between employees easy and prevalent. One popular way is through “social media” – networking, blogs, wikis, and other group tools – which are already prevalent in half of global businesses and expected to reach $4.6 billion in sales by 2013, according to Forrester. In the physical world, budgets were set aside, retreats scheduled and office spaces redesigned to encourage collaboration through informal yet “structured” interaction in employee “communities”. But is knowledge sharing good for the bottom line? Not necessarily, say Sheen S. Levine (MIT & Columbia) and Michael Prietula (Emory), who just completed a broad study on the topic. “It all depends on the specific characteristics of individuals, the organization and its environment,” said Levine, “and we discovered some elements that must be assessed before recommending any of these tools or even collaboration at all.” After documenting hundreds of instances of knowledge sharing in global consulting firm, Levine and Prietula abstracted the underlying principles and created a computer model of corporate knowledge exchange, taking into account factors such as opportunity cost and the motivation to share. “Reliable data are extremely rare and much of the discussion relies on anecdotes,” said Prietula, “creating a computer model based on a real organization allowed us to experiment with many different settings and observe the outcome on performance.” The results, forthcoming in the scholarly journal Organization Science, show that sharing can boost performance, matter little or even harm the company. The authors show, for instance, that investment in individual learning can substitute for collaboration, not complement it; companies with deep organizational memory will benefit less from encouraging collaboration; and businesses operating in a turbulent environment have little to gain from it. “Anyone who spent time on Facebook or twitter knows that too much sharing is bad for you,” concluded Levine “we simply quantified it.”

The full article is freely available for a limited time, until its appearance in print. Read.