"The Peter Principal Revisited"The Peter Principle, whose fortieth anniversary arrived this year,1 has diminished relevance in our era of flattened organizations.2
The Peter Principle, devised by Laurence Peter, an education professor at the University of Southern California, is geared to hierarchical organizations. It states that employees in such organizations are promoted until they reach a level at which they are unable to perform well. In flattened organizations, opportunities for upward career moves are, by definition, limited, so the scope for misquided promotions is also limited.
All the same, it would be interesting to know how to counteract the Peter Principle in an organization which still has numerous levels to which people can be promoted. Alessandro Pluchino (a professor in the physics and astronomy department of the University of Catania in Italy), Andrea Rapisarda (ditto), and Cesare Garofalo (a member of the Chaos And Complexity Theoretical University Study Group at Catania) have recently published a paper (pdf) that addresses this question. Its counterintuitive results have attracted considerable attention.3
The model the researchers use is illustrated in the graphic below.
Pluchino, Rapisarda, and Garofalo's hierarchical organization model.
There are 160 positions, 81 at Level 6, 41 at Level 5, 21 at Level 4, 11 at Level 3, 5 at Level 2, and 1 at Level 1.Pluchino, Rapisarda, and Garofalo (PRG) use an agent-based simulation to show that, if an organization promotes people to positions whose skill requirements are independent of the skill requirements in their previous positions (the "Peter hypothesis"), the organization should make a point of choosing the most incompetent person at a given level for promotion to the next level up.
The numbers in the lefthand column are responsibility factors for the respective levels. The responsibility factors increase as one moves up the pyramid, meaning that an individual's impact positive or negative on the organization's efficiency increases as the individual rises through the ranks.
The Competence distribution at the upper right indicates that competence is distributed normally. Rising competence is represented by deepening color of the employee figures in the pyramid from pink to dark red. Emply positions are yellow.
Although PRG do not state this explicitly, the idea behind this "Worst" promotion strategy is that the organization moves someone out of a position where he/she is creating minimal value (what I call "value" PRG call "efficiency"), while retaining all those at the same level who are doing a better job. Over time, this strategy is optimal (again, assuming skill requirements of jobs at the various level are independent of each other), as shown in the graphic below.
Organizational efficiency in the PRG simulation model.
Efficiency over time is plotted for each of six cases, three based on the "Common Sense hypothesis," in which a promoted employee is assumed to be roughly as competent in his/her new position as in his/her previous position; and three based on the "Peter hypothesis," in which a promoted employee's competence in the new position is independent of his/her competence in the previous position.Note that if skills in the new position do reasonably closely match those in the previous position the Common Sense hypothesis the best promotion strategy, as you'd expect, is to pick the strongest performer, i.e., to adopt the "Best" promotion strategy.
The three variants for each hypothesis are (1) the "Best" strategy for promotion, in which the most competent person is the one promoted to the next level; (2) the "Worst" strategy, in which the least competent person is the one promoted; and (3) the "Random" strategy, in which the choice of whom to promote is made randomly.
The greatest positive impact on efficiency occurs if the Peter hypothesis is correct and the Worst promotion strategy is adopted. Next, in descending order of impact on efficiency are: the combination of Common Sense hypothesis and Best strategy, the combination of Common Sense hypothesis and Random strategy; the combination of Peter hypothesis and Random strategy; the combination of Common Sense hypothesis and Worst strategy (negative impact); and the combination of Peter hypothesis and Best strategy (negative impact).
I'd further note that if an organization follows the rational approach of promoting people whose skills match the requirements of the new position (with a plan to fill manageable gaps through appropriate training), they should realize their intended value creation more often than not.
And, as pointed out at the beginning of this post, in a flattened organization promotion is less of a focus than assembling teams with complementary skills. Effective organizations will plan individual employees' development so that people are assisted in filling new roles and responsibilties competently.
To follow PRG's suggestion of promoting randomly, or else alternating between promoting the best performers and the worst performers, would not only be unnecessarily defeatist, but would also generate disgruntlement that would have its own serious negative effects on value creation.
1 Laurence J. Peter and Raymond Hull, The Peter Principle: Why Things Always Go Wrong (William Morrow, 1969).
2 For an opposing view, see Robert Sutton, "A New Look at The Peter Principle," Business Week, March 31, 2009.
3 Alessandro Pluchino, Andrea Rapisarda, and Cesare Garofalo, "The Peter Principle Revisited: A Computational Study," Physica A: Statistical Mechanics and its Applications, Vol. 389, No. 3 (Feb. 2010), pp. 467-472.
If you want run the PRG simulation yourself, you can do so here.
I was made aware of this article by an item in the "Ninth Annual Year in Ideas" feature published in the New York Times Magazine on December 13, 2009: "Random Promotions," by Clive Thompson. There is also a short write-up at Technology Review.