Embracing Technology: The Application of Complexity Theory to Business

by Norman N. Axelrod

Reprinted from

Strategy & Leadership
October/November/December 1999


The metaphor of complexity has been applied productively to many different planning processes, to the balance between control and freedom, and to corporate organizations and values that are structured to deal with rapid change. It serves as a means to generate quantitative models and simulations that have been used in service and manufacturing industries and in reviewing social problems. This article is the author�s report on discussions at "Embracing Complexity: the Adaptive Enterprise," the 4th Annual Colloquium on the Application of Complex Adaptive Systems to Business which was held in Cambridge, Massachusetts, in July 1999 and was co-sponsored by Ernst & Young and Complexity Magazine.

Complexity theory, Adaptive systems, Simulation, Creativity, Innovation

Norman Axelrod is president of New York-based Axelrod Associates, which consults on the effective use of technology to enhance growth and productivity. He has worked with over 100 of the Fortune 500 companies as well as with emerging companies. Dr. Axelrod reports here on "Embracing Complexity: The Adaptive Enterprise," the 4th Annual Colloquium on the Application of Complex Adaptive Systems to Business. The meeting was held in Cambridge, Massachusetts, USA, in July 1999 and was co-sponsored by Ernst & Young and Complexity Magazine.

Complexity theory deals with complex adaptive systems. These systems are made up of independent "agents that interact with each other according to a set of rules. Iterative learning by populations of these agents can modify the rules as the environment changes. People operating as independent buyers and sellers are technically agents who have self-organized an economic system. Vast numbers of organisms in our biosphere are constantly adapting one another as a way of evolving for survival. Now we have software programs that can simulate the action of populations of agents according to many different sets of roles. These programs are being used to explain, and sometimes predict, agent behavior.

The primary characteristics of complex adaptive systems are learning and adaptability, spontaneous self-organization, and phenomena that emerge from interactions among the agents but are not features of the independent agents working alone. Together these characteristics allow the system to evolve and improve its performance in changing environments. These features are particularly attractive for large or complex organizations that need to readily adapt to changing situations. Understanding the mechanism and structures that drive the dynamic improvements we observe in physical systems can often be useful in helping corporations, industries, and societies to improve their performance. Ernst & Young's colloquium provided an interdisciplinary approach to the discussion of this topic.


In a dialogue about creativity and innovation, Antonio Damasio, a neurologist at the University of Iowa and the Salk Institute, pointed out the apparent contradiction embedded in the fact that freedom of humans to create is dependent on a rigid, tightly controlled physiology that maintains homeostatis within very narrow bounds. Within the constraints of this contradiction, cognitive creative organisms have produced many physical, social, and cultural devices that helped them survive throughout millions of years of evolution. Curiosity and reasoning, both of which provide tools for dealing with change, are survival features in themselves and undoubtably helped early peoples to deal with changes in grazing conditions, predator strategies, markets, and so forth.

The concept of a balance or tension between a controlled framework and local freedom was discussed from different standpoints during the conference. Mihaly Csikszentmihalyi, a psychologist at Lake Forest College and the University of Chicago, reported on his personality studies, which revealed that conflict also exists within the personal characteristics of creative people. These people can exhibit divergent and convergent thinking, energy and productivityand apparent laziness, perserverance and playfulness, rebelliousness and traditionalism, risk-taking and caution, passion and detachment, extroversion and introversion. The emotional condition of flow - or being "in the zone" - experienced by many people when they are innovating, tends to emerge when their high levels of skill are placed under tension as they face large challenges.

Danny Hillis, a computer pioneer, researcher in emergent properties and co-evolutionary phenomena, and vice president of the Walt Disney Company, considered the drivers of flow and creativity. He sees creativity as a cognitive process, while flow is an emotion experienced during that process. Flow is a necessary but not sufficient condition for real productivity. Creativity occurs in people, and the organizations they work within can either encourage or surpress it. the organization imposes a set of templates or guidelines for acceptable behavior that create an internal tension as the creative person asks questions that can initiate behaviors that fall outside the templates.

Thomas Petzinger, Washington economics editor for the Wall Street Journal, discussed the "paradox of freedom and control." He told the story of a large chemical company in which leadership articulated a short, simple set of rules based on corporate values. Members of the corporation then had the freedom to do what was required as long as they stayed within "the bowl" or the parameter of the rules. This method preserves trial and error and provides freedom within a circumference of control. Petzinger described the sweet spot between freedom and control as "the edge of chaos." He observed that complexity often appears rigorous in the laboratory, but is more dubious in the workplace. E-coli bacteria are not equivalent to IBM. Individuals and organizations both are indeed complex adaptive systems, but each also contributes unknown factors such as emotional response and specific conditions as part of the equation. Thus, it is risky to use a one-to-one correlation with metaphors from lower species to predict higher level systems.


The successful adaptation of complex systems would seem to require some level of cooperation among the agents within the system. Francis Fukuyama, a political scientist at Johns Hopkins University, asked the question, "How can leaders nurture the emergence of social capital, i.e., the norms and values that communities share for cooperative behavior? He commented that strong leaders can create general rules that push authority down to lower levels of the organization, which allows them to adapt and take advantage of local opportunities. The radius of trust - the scope of the group willing to work together - strongly influences the type of structure and results that can be expected.

Fukuyama noted that traditional hierarchies are based on the assumption that information will flow automatically, but this is not necessarily true. Hierarchies produce information delays and distortions that severely limit the ability of managers and employees to deal effectively with rapidly changing environments. He offered as examples three large, complex hierarchies - the USSR, AT&T, and IBM - that had failed. In particular, the USSR had been predicted to fail based on the inherent impossibility of up/down communication as the system became more complex.

The limits of spontaneity within an organization are determined by size, boundries, learning from repeated contacts, existence of shared norms, and the persistence of bad choices. Hierarchical control is often necessary because creating social capital is costly. Networks don't work if one node can "bet the company" as happened in the recent failure of a major UK investment bank when its Singapore branch hemorrhaged.

Stephen Haeckel, director of strategic studies at IBM, explained his sense-and-respond model that enables large companies to systematically create and execute customer-centric business designs that adapt rapidly in the face of discontinuous change. the system provides signals or tools that enable a leader to make sense out of apparent noise and to extract meaning quickly. Adaptability is important when the inputs are not predictable, and knowing early is important for adaptive response. He noted that scenario planning and role-specific technology can be useful in allowing a faster response, and enterprise resource planning can provide forecasts, but the predictability of input and previous outcomes are also important.

George Overholser, vice president of new business development at Capital One Services, described how his corporation had used a complexity model to structure the organization. This was done in the rapidly changing credit-card industry in which massive tests are run daily to obtain data, and market requirements can change dramatically within months. They had several comparative advantages in their domestic operations and then extended them globally. He observed that evolved effectiveness replaced and outperformed engineered effectiveness; connectivity enabled unprecedented forms of self-organizing enterprise; and competency traps or local learning minimums can be escaped by maintaining separate values (e.g. in the USA and UK) to accomodate differing cultures while permitting cross-pollination.


The application sessions provided a sampling of how complexity theory can be applied to different levels of accomodation or deviation in order to deal with practical problems. Swiss Re and the Santa Fe Institute have been exploring models to reduce risk and to increase productivity. They used complexity theory to create a computer simulation to provide insight into the interaction between catastrophic events - earthquakes, cyclones, floods - and the balance between investing income and paying the insured. By assuming different conditions, observers can learn from the various results obtained.

The US Marines use complexity theory to compute mathematical models as a way of finding new, more adaptive ways of organizing troops. Increasing amounts of information have become available on the battlefield, and network forms of organization are being considered to uncover forms of organization that will best support communication and teamwork. The models are used to explore different scenarios and their consequences.

British Telecommunications a computerized allocation procedure to increase the effciency of its service calls. It integrated constraint programming and simulated annealing algorithms in a hybrid approach. the system was put into operation in 1997, reaching 20,000 engineers in 1998 and saving BT $150 million per year in operational costs.

In a second example of scheduling, simulations on cargo handling were created for Southwest Airlines by the Santa Fe Institute. They used computational agents to represent decision-makers. It was reported that the recommended solutions seemed remarkedly intuitive to those who have a big-picture view of the system.

In summary, the metaphor of complexity has been applied productively to many different planning processes, to the balance between control and freedom, and to corporate organizations and values that are structured to deal with rapid change. It serves as a means to generate quantitative methods and simulations that have been used in service and manufacturing industries and in reviewing social problems. It enables meaningful communication between people in very different enterprises. Complexity theory provides a flexible paradigm with much potential.


Return Home

Axelrod Associates

445 East 86th Street
New York, NY 10028