2. Traditional Great Groups and Future Model of Great Groups

Distilling Warren Bennis' work, there appear to be some common attributes most great groups share.
The first is that they have a good, strong leader.
These leaders can have widely varying styles, but they share a handful of qualities.
They have a passion for hiring excellent people and a knack for finding the right work for those people.
They provide a good working environment for the group and shield the group from external pressures of bureaucracy.
In addition to such a leader, the Great Group is composed of very talented, highly motivated people.
Often driven by a common enemy or unifying principle, these people are extremely focused on the task at hand.
Personal relationships and social life outside the group are often neglected.
Being able to work on such a team is such a rush that the work is its own reward and lends its participants a sense of being important.

These talented workers need to work in an environment conducive to their success.
This frequently isn't a nice office building with air conditioning.
Many of these groups succeeded in fairly unpleasant places to work.
The critical factor is that they are isolated from the pressures of the larger organization and left with substantial freedom.
They also need the proper equipment and funding to get their work done.
When a Great Group is rolling, the environment is energized, usually fun and often sexually charged.
A final part of this environment is the need for group members to be able to have their work reviewed.
In some instances, like at Disney, the leader would be the most demanding critic and drive the group towards excellence.
Other groups would rely on peer reviews to accomplish the same goal, Xerox PARC for instance.
In this sort of environment, a Great Group can flourish.

Unfortunately, creating this sort of environment is rather difficult.
The greatest challenge is finding and recruiting the top talent that is needed.
This requires hiring many people who are near the top of their respective fields and are able to work well with others.
For an organization intent on creating such a group, they are often burdened with putting this team in an area separated from the rest of the organization.
Often this means getting a separate building.
If this is achieved and the leader is good enough to create the proper environment, there is a chance that the group will click and do extraordinary things.
However, the obstacles are very serious.

One recent Great Group stands in contrast to many of the others.
This is the group that is responsible for Linux.
The Linux operating environment started as a side project for Linus Torvalds ten years ago and has since become a serious competitor for everything that needs an operating system from embedded devices to desktop computers to mainframes.
This was done while deeply entrenched players in the market were considered unbeatably, except perhaps by one another.

This group operates in a different manner than the other great groups that have been discussed.
It does not require its leader to protect it from a greater organization because no such organization exists.
This group is an organization unto itself.
There is a central core of workers who appear somewhat similar to a normal great group.
They are fiendishly passionate about the work and make considerable sacrifices for it.
They are also a collection of highly talented people.

However, unlike other Great Groups, this rich talent pool does not dominate the group.
Instead, widespread use and code contributions from individuals around the world provide the bulk of contributions to Linux.
Some of this comes from interested individuals; other work comes from interested corporations.
The key principal is that the Linux group is able to achieve great things using a relatively small core of people who would otherwise be the right type, but insufficient in number and resources to form a Great Group.
These people, build on the work of many, many others who are not as devoted and would likely not qualify as potential members of a great group.
They are much more normal people.
The recent growth of the internet, logged email threads on the like has provided the environment where this organization model can thrive.

The Linux model is attractive and on the surface may make it seem like Great Group creation is much easier than it once was.
It has several detractors though.
The first is that seed work must be created by a talented group of people.
A cause must be either addressed by this seed or created by it.
This cause has to be powerful enough to attract the large extension of the group that will provide large amounts of effort.
If such a large group can be put together, the core members must be able to manage the information flow to and from the wide flung contributors.
Perhaps the largest disadvantage though, is that the corporate world will struggle to take advantage of creating such groups.
The need for openness of information will destroy efforts to generate the intellectual property Great Groups are so good at creating.

What is needed is a new model of Great Group that can leverage the benefits of the Linux model while functioning well in corporate environment.
This model must support a small core of excellent people with Great Group mentality who are lifted to greatness by the work of normal people.
The core group will function as leaders and thinkers for their project, but will need to use the entire company as an extended pool to receive ideas from.
This model should be able to take advantage of people who are both physically near the core group and those who are distant.
The core group must be set up as special enough that they will devote themselves wholly to the project but their support staff will be more effective if they are also able to feel special because of their relationship to the team.

Another motivational problem in the corporation is finding an enemy or cause for the group.
Frequently, the big enemies of a corporation are irrelevant to some parts of the company and so corporate propaganda must be customized to best motivate each group it wants to be Great.
The final challenge in the corporate world will be to somehow bring the group within the corporation while leaving it in tact.
It will have to be protected from the interference of upper management as well as too much input from other people in the company.
A basic outline of the differences between the models discussed is below.

Figure 1 is hypothesical curve of group's size and performance.

Figure 1: hypothesical curves of groups' size and performance

The dashed curve indicates normal group's performance.
Because there are limited number of great talented individual, a normal group has to add average or less talented individuals into the group.
Scalability issue also prevent them from increasing performance by adding more participants.
In Great Groups indicated by thin curve and thick curve, when their performance reaches a certain threshold, the performance booms off by being a Great Group: high motivation and intensed collaboration.
Their high motivation and intensed collaboration lead positive feedback to the group so that their participants get higher motivation and more intensed collaboration to achieve greater works.

The difference between traditional Great Groups indicated by thin curve and new Great Groups in thick curve is scalability.
In a traditional Great Group, the leader has to find and add good talents to the group and keep the group reasonably compact to prevent individual from bereaucracy and overhead of communication.
Because overhead of communication is proportion to squared number of people[MMM], it virtually defines the maximum size of the group.
On the other hand, new Great Groups, such as Linux Community, take benefits of scale using computational technologies such as the internet.
Even if a new Great Group cannot break the "Great Group" threshold by only great talented individuals, the group incorporate normal people with supressing communication cost by using collaborative technologies, to hit the threshold.
However, after they reach the limit of scalability of collaborative technologies, they will loose their performance or get another technology to adapt to larger scale.
For example, Linux community has gotten too big to capture bugs and todo lists by hands. Their solution is to use computational support tool to keep tracking problems so called bugtrack system.

In future, computational technologies must support (1) taking benefit of scale , diversity and distribution of a group and (2) reducing overhead of the scale. Such computational technologies will be discussed in the next chapter.
The table 1 shown below describes the summary of comparision among traditional Great Groups, Linux Community and future Great Groups.