This paper is a summary of an Aspen Institute sponsored in-depth roundtable session, written from the perspective of one informed conference observer (Bollier). The participants are leading thinkers in the many complex areas this paper covers (economics, systems theory, human behavior, human futures, information technology evolution, etc) and are listed on page 57. A selection of their key insights shared in the paper are listed below:

A "push" economy is geared towards mass production, anticipating consumer demand, and routing resources to the right place at the right time, to create standardized and mass produced products. By contrast, a "pull" economy is based on open, flexible production platforms that are used to orchestrate a broad range of resources. Instead of producing standardized products, "pull" model companies are demand-driven, and assemble products in customized ways that serve specialized or local needs, usually using "rapid" or "on the fly" processes.

Several global corporations are moving towards "pull" methods, and away from "push" models; ie., Toyota, Dell, Cisco, Li & Fung. These companies employ different variations of Value Network models, that share information about overall network performance and best practices for serving specialized needs, among hundreds or even thousands of partner companies that make up the network. This creates an intra-network knowledge commons. Some companies also work closely with Open Source Software projects, thereby expanding their "pull" network, and expanding their knowledge commons into a broader Open Commons via Open Source Software project contributions. Thus, "pull" business models also tend to be Network Value-Increasing, and Commons-based business models as well.

"Pull" models can also be platforms for creating "increasing returns dynamics." This is due to "pull" models being based around loose and flexible networks that are already configured to scale as growth occurs. So, growth does not incur the huge overhead costs in administration that "push" models must contend with. Pull platform key characteristics include modular and loosely-coupled networks, open channels that better harness the passion and commitment of innovation communities. "Pull" platforms also will tend to influence public policy with regards to education and innovation, as more companies tend to gravitate towards the "pull" models.

The areas where "push" models tend to succeed in business are in areas where people do not know what they want, and prefer to shop from pre-made selections (Ikea, Home Depot). However, there are even "pull" models to found here, in the form of user-driven innovation, such as mountain biking, extreme skiing, hot rodding, etc. In these pro-amateur niches, customers don't necessarily know what they want, but do want to be a participant in the "pull" network that creates the product.

How do you tax a product that is made in 23 different countries? "Pull" models are going to change the way that governments create policy as more companies gravitate toward them. This will influence laws about intellectual property, education, taxation and more.

"Pull" economies are not just centered around finding creative ways to "outsource/offshore jobs" away from one place and to the places where "labor" is "cheaper". Successful "pull" models have encouraged and aided "insourcing", where more jobs are created, for instance in the United States by "foreign sources (a total of 7 million cited by this paper), than are out sourced (a total of 600,000+ cited by this paper). This is because pull models seek out, not just the "cheapest" labor, but the best ways to add value to the production networks. So, they can scale to many participants around the world, regardless of local labor costs, to find the best participants needed for specific specialized productions.

The social dynamics of "pull" models are highly centered around creating relationships of trust, sharing knowledge, and close cooperation among network participants. In "pull" models, non-market value creation (tacit knowledge, intangible value) is generally steered towards a commons-based model. A commons is used as a "collective governance regime for managing shared resources sustainably and equitably." Many of these commons are made possible by networked information technologies (the internet).

Bollier suggests that "if online commons are going to be useful to business, companies will need to do more work to develop protocols for identity and reputation management". This is because the use of the commons is based around trust. It also due to the need for ways to measure qualitative value in intangible assets beyond money, like knowledge, individual performance and value multiplication, and network wide performance/value multiplication.

Roundtable participants also noted that "pull" models will pose challenges to current education regimes that are centered around training people to participate in "push" economies. One of the participants mentions that " Computers, software tools, and Internet resources make possible some radically new styles of learning. By using pull-based systems, students can function much like businesses in the pull environment: They can access resources they don't control and put themselves into flows of activity, rather than just building inventories of static, objectified "knowledge."

Dan Bricklin examines ways to induce a pool of users to contribute to a commons without extra effort, using the architecture of the commons (as in Napster's default to sharing in the way download directories are available) and leveraging user's self-interest. The key to understanding the success of Napster and other file-sharing technologies resides not in their 'peer-to-peer' nature but in the fact that they provide users with access to a database of desirable things and enable people to create a public good in the process of seeking their own interests.

Bricklin identifies three ways to fill a database: organized manual, organized mechanical, and volunteer manual.

CDDB succeeded at motivating volunteer manual data entry because it leveraged the desire for users to have their data in the database so that CDDB-aware programs could access it, for example when a user would insert a CD into their computer.

Bricklin calls this "harnessing the power of individual selfishness."

Napster cleverly avoided manual data entry by automatically indexing anything in the user's 'Shared Music' directory. Thus "storing the copy in the shared music directory [was] a natural by-product of the user's work with the songs."

Sharing is the default. This results in users "adding to the value of the database without doing any extra work."

Six Degrees begins in the beginning. Stanley Milgram's initial small world studies are analyzed. His findings in seeing if a group of people in Nebraska can get a letter to someone in Massachusetts are scrutinized. Milgram left a puzzle. Mathematically, six degrees of separation can be shown and intuitively it is appealing. But do social networks actually work that way?

Initially, Watts steps into the world of pure mathematic theory. Graph theory and random graphs are employed to build potential worlds in which connections can be made. These tools are detailed and their histories explained.

Watts and his colleagues then take the science to new levels, by introducing sociology, epidemiology, economics, and business models into this new multi-disciplinary science. Immediately, each new field of study brings with it new insights into network dynamics.

This convergence of disciplines reveals the social, transportation and technological networks that make up our world. These networks are, ultimately, made up of individuals. Individuals in turn relate back to the networks and define how they operate.

Socially, people relate to their network by clustering. Clusters are logical organizations of network elements. In a social context, we might cluster in terms of a religion, a favorite author, a school we are attending or an affinity for a type of food. Some of these have very close physical distance, while others have a social distance with members spread out over a large area.

Networks of this type are, to various extents, “scale-free” networks. If graphed these networks roughly follow a classic power law trend where the level of connectivity between two nodes in a network increases dramatically as more nodes are connected. Real-world scale-free networks tend to have highly connected hubs which rapidly, purposely, and efficiently transmit pertinent or pervasive content from one location to another. In social circles, these are networkers. In the airline network these are hub airports. In traffic they would be freeway interchanges.

Due to this architecture, the Internet and modern air transport have combined to greatly decrease the role of proximity in our social networks. This has had great impacts on commerce, tourism, cultural sensitivity and other social factors. However, it has also led to great risks in the transmission of diseases, sensitivity to distant economic fluctuations, and rapid spread of misinformation.

These dynamics create a type of network that Duncan calls simultaneously robust and vulnerable. Their strength and weakness is that, with rapid transmission from cluster to cluster, anything can move quickly from one location or group to another. He uses the example of Toyota, whose network of suppliers was organized in such a way as to quickly compensate for and recover from a potential economic catastrophe.

Stable scale-free networks do not rely on a rigid hierarchy to provide direction in times of crisis. Rather, the structure of the network itself can rapidly respond to an unforeseen situation.

Their network was arranged in such a way as to foster and reward communication. This communication helped cope with ambiguous or unplanned situations. Rather than paralyzing Toyota while people waited for a decision from a rigid hierarchy, the contractors in the network were able to analyze the calamity and provide a rapid response to it.

As mentioned above, this robustness also rapidly transmits malicious content as well. The Melissa Virus, SARS and Ebola are analyzed to show why the network did or did not transmit them and, when it did, how they eventually died out.

Watts ends this book by summarizing that the multidimensional nature of social distance is sometimes counterintuitive and subjective. People can feel close in a network sense to people they are physically distant from and, conversely, socially distant from people physically nearby.

He continues by warning that social and physical distances have shrunk. People can quickly travel from place to place and economies are highly interdependent. The sheer number of dependencies in the modern world may yield surprising results from seemingly insignificant actions.

He finishes by showing the stability of our networks with the example of how New York adapted to the 9-11 attacks. The City bounced back to semi-normal operations within a week. During the disaster, the best laid plans of emergency operations staff were scuttled by the utter unavailability of facilities and services designed to copy with disasters. The network will provide.

Network forms of organization, with reciprocal patterns of communication and exchange, are alternatives to hierarchically or market based governance structures; they are more suited to describing companies involved in an intricate latticework of collaborative ventures with other firms over extended periods of time.

Hierarchies are suited to transactions that involve uncertainty, recur frequently, and require substantial “transaction-specific investments”. Markets are suited to exchanges that are straightforward, non-repetitive, and require no transaction specific investments.

These “alliances” aim at creating indebtedness and reliance over the long haul: your current collaborator will be your competitor in other domains (or in the same domain) over time. In markets, the strategy is to drive the hardest possible bargain in the immediate exchange. Commitment is low.

Network organizations are more social than markets and hierarchies, they are dependent on relationships, mutual interests, and reputation. They are less guided by a formal structure of authority. Successful networks involve complementarity and accommodation. Reputation, friendship, interdependence, and altruism are integral. The most useful information comes from people you have dealt with in the past rather than from the formal chain of command.

Conflicts are resolved by haggling in markets; administrative fiats in hierarchies; norms of reciprocity and reputational concerns in networks.

Markets offer choice, flexibility, and opportunity. Prices determine production and exchange. Hierarchies are well-suited for mass production and distribution. Networks are more flexible than hierarchies. Transactions occur through networks of individuals engages in reciprocal, preferential, mutually supportive actions.

Reduction of uncertainty, fast access to information, reliability, and responsiveness are paramount concerns that motivate participants in network organizations.

Know-how, the demand for speed, and trust are critical components of successful network organizations.

Examples of network forms:

• Craft industries (construction, publishing, film and recording industries)—facilitated by loyalty to the profession and to project teams. Informal trading of proprietary expertise is common among members of a profession in different organizations.
• Regional Economies and Industrial Districts (German textiles, Silicon Valley)—Rich array of support services. Pooling resources on basic research through consortia and trade associations. Encouragement by local government. Proximity to centers of higher education.
• Strategic alliances and Partnerships (Oil and gas, chemical and pharmaceuticals, commercial aircraft)—share risk and expense. Cooperative relationships with suppliers. Gain fast access to new technologies or markets. Needs to deal with anti-trust concerns.
• Vertical disaggregation. (Downsizing, outsourcing)—Increased flexibility in responding to technological change and commodifcation of products.

Know-how, the demand for speed, and trust are critical components of successful network organizations.

• Know-how and detailed knowledge of the abilities of others who possess similar or complementary skills thrive in networks. Exchange of competencies is more likely to occur in networks; exchange of tangible resources is more likely to occur in market transactions or among units in a hierarchy.
• Demand for speed is facilitated by a network’s strengths in offering fast access to information, flexibility, and responsiveness to changing tastes. Information flow through a network is freer and richer than in more tightly controlled markets or hierarchical organizations.
• Trust develops when there is a high probability of future association. There is a higher probability of cooperation and also a willingness to punish those who do not cooperate.