people build mental models Sinan Aral, Erik Brynjolfsson, and Marshall Van Alstyne, “Information, Technology, and Information Worker Productivity,” Information Systems Research 23, no. 3 (2012): 849–67; Sinan Aral and Marshall Van Alstyne, “The Diversity-Bandwidth Trade-Off,” American Journal of Sociology 117, no. 1 (2011): 90–171; Nathaniel Bulkley and Marshall W. Van Alstyne, “Why Information Should Influence Productivity” (2004); Nathaniel Bulkley and Marshall W. Van Alstyne, “An Empirical Analysis of Strategies and Efficiencies in Social Networks,” Boston U. School of Management research paper no. 2010-29, MIT Sloan research paper no. 4682-08, February 1, 2006, http://ssrn.com/abstract=887406; Neil Gandal, Charles King, and Marshall Van Alstyne, “The Social Network Within a Management Recruiting Firm: Network Structure and Output,” Review of Network Economics 8, no. 4 (2009): 302–24.
leveraged existing skills In response to a fact-checking email, Van Alstyne expanded upon his comments: “One of the original hypotheses attributed the gains of the smaller project load to the efficacy associated with economies of specialization. Doing a singular, focused activity can make you very good at that activity. The idea goes all the way back to Adam Smith and the efficiency associated with focused tasks at a pin factory. Generalization, or pursuing diverse work in our context, meant spreading projects across finance, education, and commercial IT. These are very different industries. Running projects across them requires different knowledge and it also means tapping different social networks. Specialization, in these consulting projects, meant focusing on, say, just the finance projects. Knowledge could be deepened within this focal area and the social network could be adapted to finance contacts alone. At least this is one theory as to why specialization might be better. Obviously, specialization can restrict the number of possible projects—there might not be a new finance project when there does happen to be one, or several, in education or IT. But perhaps if you wait, you’ll get another finance project.”
deemed a success In response to a fact-checking email, Van Alstyne identified other reasons why joining small numbers of projects, and a project at its start, had benefits: “The first is multitasking. Initially, taking on new projects strictly increases output, in this case revenues generated by these consultants. Revenue growth can continue even past the point where the productivity on a given project starts to fall. Consider a project as a collection of tasks (assessing client needs, generating target candidates, selecting candidates, vetting resumes, presenting options to clients, closing the deal…).As a person takes on new work, its tasks displace some tasks of the existing work. So an existing project can take longer when a person takes on a new project, drawing out the period over which he/she gets paid. Total throughput, however, can still rise for awhile as a person takes on new projects. The stream of revenues brought in by a person juggling 6 projects tends to be higher than the stream of revenues brought in by a person juggling 4 even though each of the 6 projects takes longer than it would have taken if it were only in a group of 4. At some point, however, this relationship trends completely downward. New projects take too long and revenues decline. Taking on another project strictly decreases productivity. As one consultant put it, ‘There are too many balls in the air and then too many get dropped.’ It takes too long to complete tasks, some tasks are not completed at all, and the flow of revenues dribbles out over a really long period. So there is an optimal number of projects to take on and this is below 12. The second consideration, as you suggest, is access to rich information. This exhibits a similar invert-U pattern. We were able to judge how much novel information each person received by tracking their actual email communication. We measured this both in a sense of ‘variance,’ i.e., how unusual was a fact relative to other received facts, and also in terms of ‘volume,’ i.e., how many new facts a person received….Initially, greater access to more novel information strictly increased productivity. Superstars did receive about 25% more novel information than their typical peer and this access to novelty helped predict their success. Eventually, however, those outlying people who received the absolute highest novelty—about twice that of the superstars—were less productive than the superstars. Either excess information was too weird, off-topic, and not actionable or excess information was too much to process. A massive volume of novelty introduces the white-collar worker’s equivalent of the ‘Where’s Waldo’ problem: You can’t find the important information in all the noise. Both of these factors were statistically significant predictors of the superstars.”
bright morning sky Richard De Crespigny, QF32 (Sydney: Pan Macmillan Australia, 2012); Aviation Safety Investigation Report 089: In-Flight Uncontained Engine Failure Airbus A380–842, VH-OQA (Canberra: Australian Transport Safety Bureau, Department of Transport and Regional Services, 2013); Jordan Chong, “Repaired Qantas A380 Arrives in Sydney,” The Sydney Morning Herald, April 22, 2012; Tim Robinson, “Qantas QF32 Flight from the Cockpit,” The Royal Aeronautical Society, December 8, 2010; “Qantas Airbus A380 Inflight Engine Failure,” Australian Transport Safety Bureau, December 8, 2010; “Aviation Occurrence Investigation AO-2010–089 Interim-Factual,” Australian Transport Safety Bureau, May 18, 2011; “In-Flight Uncontained Engine Failure—Overhead Batam Island, Indonesia, November 4, 2010, VH-OQA, Airbus A380–842,” Australian Transport Safety Bureau, investigation no. AO-2010–089, Sydney.
de Crespigny later told me I am indebted to Captain de Crespigny for his time as well as his book, QF32. In an interview, de Crespigny emphasized that he is speaking for himself, and not for Qantas, in recalling and describing these events.