United Airlines discovers the complexity of talent management

On Sunday, April 9, 2017, passenger David Dao was forcibly removed from United Airlines flight 3411 after being randomly selected because the flight was overbooked. Videos of the incident went viral, causing immediate backlash. The public response was swift and voluminous: according to Google Trends, the number of Google searches for the term “United Airlines” spiked to a ten-fold level relative to normal.

In addition to the potential brand impact, this incident may have a substantial economic impact on United itself and on the entire airline industry: just one day after settling with the passenger for an undisclosed amount, on April 28 United CEO Oscar Munoz issued a letter to customers apologizing for the incident and indicating that United will increase incentives up to $10,000 for voluntary rebooking. Given that more than half a million passengers were bumped from U.S. airline flights in 2015, costs could skyrocket if the industry follows United’s example.

But my interest in the incident is less about the brand and financial impact, and more about the operational implications, especially in the context of talent management. How can it be that an organization whose customer commitment statement opens with “We are committed to providing a level of service to our customers that makes us a leader in the airline industry” allow such an incident to take place?

To explain this, I want to introduce two concepts that are at the root of a new approach to talent management: complexity and emergent behaviors. We define a system as being complex if it consists of many interacting parts, and the behavior of the system as a whole emerges from the interactions in ways that are not easy to predict. Examples of complex systems are all around us: from natural systems such as social insects and flocks of birds, to human systems like stock markets and sports teams.

Emergent behaviors are those behaviors observable at the system level, which result in sometimes unpredictable ways from the behaviors of individual elements. For example, geese flying in a V formation are not following the orders of the goose captain telling each goose what to do. And to use a more familiar example of emergent phenomena, traffic jams emerge even when everyone on the highway is trying to get to work as soon as possible.

The notions of complexity and emergence are very useful in analyzing the behavior of organizations, and in understanding why there can be such a dramatic dissociation between an organization’s stated mission, and the behavior of its employees. Consider a baseball team: There are only nine players on the field, each with a clearly defined role. There are terabytes of data collected over decades that can track the outcome of a relatively small set of possible actions (throw a ball, swing a bat, catch a ball), and there is a well-established field of analytics – as eloquently described in Michael Lewis’ best-selling book, Moneyball – that provides a quantitative link between the characteristics of each player and their contribution to the overall success of each team. In spite of this, predicting the outcome of any single match, let alone an entire season, is virtually impossible.

Now take this example and expand it to a company like United, with some 86,000 employees holding hundreds of different roles to transport more than 100 million passengers per year to dozens of countries around the world. Clearly, the overall behavior of the company is impossible to predict from the bottom up.

The real problem comes when we try to go in the other direction, from the top down. Expecting a company’s mission statement to influence the behavior of individual employees is like expecting the manager of a baseball team to influence the outcome of a single play by telling the media that his team is playing to win the World Series. The best that a manager can do is to instate procedures and policies that encourage the “right behavior” in individuals. But as we already mentioned, the link between individual behaviors and the overall behavior of an organization is extremely difficult to predict, especially in today’s business environment, which is subject to rapid change and even more rapid and uncontrolled diffusion of information.

Most tools used to manage corporate talent are based on analytical methodologies that cannot deal with the increasing complexities of today’s business environment. For the past two decades I have had the fortune of researching, teaching and applying the concepts of complexity science, which relies on an inside-out approach to analyzing organizations, and uses computer simulations to capture the behavior of an organization based on the behaviors and interactions of its employees.

Although these methodologies cannot predict a specific event, such as a passenger being forcibly removed from an airplane, thinking about talent management from the inside out can help organizations understand the impact that its policies and processes have on the behavior of individual employees, and how this in turn can impact the overall behavior of the organization, as well as its culture and reputation.