New Rules For Team Building-The times are changing–and so are the guidelines for maximizing team performance.

Whenever I board a commercial airline flight, I have the impulse to stick my head in the cockpit and ask, "First trip
flying together?" I would never actually do such a thing, of course. But it comes to mind because of some analysis
by the National Transportation Safety Board that has profound implications for airlines, their passengers, and
anyone who creates or leads organizational teams.

by J. Richard Hackman Optimize Magazine

In a 1994 study, the NTSB staff combed its database to identify the circumstances under which cockpit crews are
most likely to get into trouble. Surprisingly, they found that 73% of all incidents occurred on a crew’s first day of
flying together, and almost half (44%) took place on a crew’s very first flight. The lesson for all businesses?
Teams are most vulnerable to mishaps when they’re starting out, before they’ve had the chance to learn through
experience how to best work together.

The NTSB analysis reinforced earlier findings by NASA’s Ames Research Center. Attempting to assess the effects
of fatigue on pilot behavior, the researchers unexpectedly wound up with evidence suggesting that there are real
benefits to keeping teams together. They recruited a number of crews as they returned from several days on the
job, and compared their performance on a moderately demanding scenario in an aircraft simulator to that of
crews whose members had been off duty for several days.

As one would expect, the pilots returning from multiple-day trips were tired–and individual members of those
crews made more mistakes than rested pilots did. But the surprising finding was that fatigued crews made
significantly fewer errors than did crews composed of rested pilots who had not yet flown together. The
experience they gained working together as a team more than overcame the debilitating effects of individual
fatigue.

There are many reasons why reasonably stable teams perform better. Members develop familiarity with one
another and with their collective task, so they can quickly settle in and get to work rather than waste time and
energy getting oriented. They build a shared pool of knowledge that’s more comprehensive than that of any
individual team member. They learn who is especially knowledgeable or skilled in different aspects of the work,
and how to deal with members who are less skilled without excessively disrupting the team’s progress. And
gradually, they grow a shared commitment to the team and a measure of caring for one another.

The NASA findings, along with those of many other studies, confirm the advantages of stable, well-bounded
teams. Yet most airlines, like many companies, still train team members as individuals and change crew
composition constantly because of long-standing practices enforced by labor contracts. Why have airline
managements, pilot unions, and federal regulators–despite their commitment to flight safety–not jumped to
implement policies based on the new research findings? For one thing, they don’t want to believe them.
Conventional wisdom says teams learn to work together early on, but that the learning eventually plateaus, and
finally yields to overfamiliarity. "Everyone knows" that if a team stays together too long, members become too
comfortable, too lax in enforcing standard procedures, and too ready to forgive teammates’ mistakes. Only a
constant flow of new members keeps teams on their toes.

Everyone knows these things–but they simply aren’t true. Members of competently designed teams do learn fairly
rapidly how to work together. But there’s little evidence that the learning stops at some point. Like the Guarneri
String Quartet, which has continuously improved its music-making over more than three decades of playing
together, the best teams get better and better indefinitely.

Unclear team boundaries and membership instability are pervasive and pernicious problems not just for airlines,
but for many types of work teams, including project task forces, senior-management teams, sales teams, and
increasingly, virtual teams whose membership may be as fluid and uncertain as it is dispersed.

Because the work of managers and professionals invariably involves extensive engagement with a wide variety of
individuals and groups, teams made up of these creative personalities are at special risk for what organizational
psychologists call "underboundedness," or a lack of clear definition. Case in point: I was invited to meet with the
top-management team of a large insurance company, about two dozen senior executives. The CEO was
increasingly frustrated as he tried to get the team to take on collective responsibility for the well-being of the
company as a whole, and the team members were equally unhappy. Everyone wanted to identify just what was
keeping these talented and committed individuals from pulling together.

I would argue that these senior managers weren’t a team at all, but merely co-actors. Each member’s main
responsibility and accountability was for the performance of his or her own unit. No one was clear on why they
were meeting or what they were supposed to do as a team. Nor was it clear exactly who was on the team; there
were so many members that it was hard to keep track, especially since substitutes often attended in place of
team members. Only one thing was certain: At heart, the CEO believed that running the organization was his job.
He really didn’t want the team to make important decisions at all.

Setting boundaries
Teams of many different sizes and durations dot the landscape of every company, and indeed, managerial and
professional work does not lend itself to working for a single team for extended periods. Instead, one person is
likely to serve simultaneously on a variety of teams that form, reform, and disappear like sand dunes on a windy
beach. In one financial institution, almost all professionals serve on multiple teams, some with life spans that
extend indefinitely, some created to accomplish a particular task, and some created on the fly to solve an
unanticipated problem. But "sand-dune" teams need clear boundaries as much as, or even more than,
permanent teams.

Managers and professionals often feel they don’t need the clear team boundaries upon which they insist for the
rank-and-file teams that report to them. "We’re experienced at this kind of thing. It’s what we do every day," one
senior manager told me. "We don’t need all the i’s dotted and t’s crossed." It may be true–but I doubt it–that
managers and professionals are skilled at handling ambiguity and uncertainty. But the underbounded character
of many such groups makes it hard for them to demonstrate how good they really are at teamwork.

Before a group of co-workers can develop into a superb team, it first must actually be a team–a reasonably stable
unit with shared responsibility for a defined piece of work. Beyond that prerequisite, my research collaborators
and I have identified three things leaders can put in place to increase the chances of team effectiveness. These
are: a compelling direction; an enabling structure and context; and available, expert coaching.

A compelling direction. It’s up to the leader to establish a team’s direction by specifying and communicating its
overall purpose. Direction is critical to energize the team, get it properly oriented toward its objectives, and engage
the full range of members’ talents. Everything else depends upon this first step–the team’s design, the kinds of
organizational supports provided, and the character of leaders’ hands-on coaching.

Good direction is clear, so people can orient their work properly; it’s consequential for the organization and its
customers; and it’s appropriately challenging, neither impossibly demanding nor so easy that it’s uninteresting.

Managers sometimes use rhetorical devices to try to make a team’s direction seem more compelling than it really
is. But this is akin to trying to convince brick carriers that they are building a cathedral. It may work for a while, but
eventually it becomes clear that they really are simply carrying bricks. When the work is essentially trivial, it’s
impossible to fully engage members’ talents.

In setting direction for a team, managers can focus on the ends to be achieved or the means for pursuing them.
Good leaders exercise their authority to specify end states, but not dictate the details. Specifying both ends and
means mitigates the challenge to the team and underuses its resources; specifying neither invites anarchy rather
than purposive work; and specifying means but not ends is the worst kind of micromanagement.

An enabling structure and supportive context. Some teams have difficulty because they’re not set up
correctly, or their structures and systems undermine members’ efforts. Overly elaborate structures just get
in the way, while the belief that self-managing teams can work out everything on their own may lead to
insufficient structure and support.

Probably the single biggest mistake team leaders make is believing that "the bigger, the better" is the right
approach. That’s what IBM did in the 1960s when it was creating OS/360, then the largest systems-programming
effort ever undertaken. As generally happens with large-scale projects, it was running behind schedule. The
temptation in such cases is always to compute how far behind you are and then add staff to make up time. If the
project is a dozen person-months behind, perhaps assigning a dozen extra people to it for one month will get it
back on track. In retrospect, Frederick Brooks, IBM’s manager of systems programming at the time, acknowledged
that it’s simply not possible to produce a baby quickly by asking nine women to carry it for a month each. Brooks’
Law tells us that adding people to a late software project just makes it later.

The best group size depends on the task, of course, but our data points to a simple rule of thumb: No team should
have more than six members. I believe in this rule because the number of group-process problems a team
experiences tracks not the number of members, but the number of links among members–that is, the total
number of paired relations, which compounds at an accelerating rate as size increases. A six-person team has 15
pairs among its members, but a seven-person team has 21–and the difference is noticeable. Twelve people have
66 links. And imagine the process losses in a 24-person board of directors (552 links).

The politics of size
So why do we see so many large teams in the workplace? It likely has more to do with emotional issues than
team performance. A large team spreads accountability and is politically expedient, giving all relevant
stakeholders a voice so they accept its final product. But its very size hampers its ability to create something
usable.

My university has taken a two-pronged approach to creating a functional management team. Harvard has a
30-person Board of Overseers on whom it depends for ideas, perspectives, and, of course, contributions. But the
overseers don’t govern the university. The decision-making group is the Harvard Corporation, which consists of
five outside members plus the university president and treasurer–just about the right size to make the
consequential decisions.

Another example: A startup organization once grew to the point at which its dozen founding officers could no
longer make decisions in the informal manner that had worked well in the early days. The CEO considered
devising a senior-management structure himself. Then, in light of the company’s democratic spirit, he considered
asking the dozen officers to come up with a structure they all could accept. But he wisely realized that with such
enormous personal stakes, the team was unlikely to be able to put the company’s needs first.

Finally, the CEO formed a reorganization task force of four highly respected officers to develop a proposal for the
new structure. Each member also was responsible for staying closely in touch with three other officers who were
not on the task force. The first agenda item at every meeting was a report on the views of the nonmembers, and
the last item was an explicit review of what should be communicated to them. Although there were many rough
spots along the way, the team eventually came up with a reorganization plan that was accepted by the CEO and
the other officers.

Expert coaching. A leader can promote team effectiveness by helping team members learn to work
interdependently and manage themselves. Like teaching, coaching is done best when leaders exploit their own
personalities and style to get the lessons across, rather than try to follow some preset formula.

Still, there are particular times when team members are likely to be especially open to coaching. Although I’m
uneasy about applying examples of sports teams to the corporate world, as their contexts are so different, the
behavior of good athletic coaches illustrates the coaching functions that can be performed at different times in a
team’s life. In the locker room before the game, coaches focus on motivational issues, establishing that the
contest ahead is challenging but that the team has a real chance to win if members play their best. In the locker
room at halftime, coaches move to a consultative role, helping to revise strategy based on how things are going.
The next day, as the team reviews the game, coaches emphasize educational interventions, helping the team
learn from experience and build proficiency for its next contest.

Even a first-rate coach can make little constructive difference in a team that’s fundamentally flawed. Indeed, a
coach may do more harm than good in such a case, by distracting members’ attention from the fundamental
aspects of structure or context that they ought to be addressing.

Two questions come to the fore when a team’s direction or design is fundamentally flawed. First, can the purpose
and format of the team be restructured to give members a reasonable chance to perform well? And, if not, is it
imperative that the team exist at all? There are some kinds of work for which teams are a wholly inappropriate
design choice (see chart, p. 56). And there are some settings in which teams can never succeed.

Creative writing, for example, often is assigned to a team, but should not be. Expressing and organizing ideas in
writing are tasks inherently better suited to individual than collective performance. Committee reports invariably
turn out better when written by one talented individual on behalf of the group rather than by the group as a whole.
The same is true for many aspects of executive leadership. The most powerful statements of corporate vision
invariably are the product of a single intelligence willing to establish a collective purpose that goes beyond what
others believe to be the limits of the organization’s capability.

Design strengths
A team’s basic design also strongly conditions the impact of leaders’ coaching, as was shown in a 2001 study of
self-managing field-service teams at Xerox by organizational psychologist Ruth Wageman. For each team
studied, Wageman obtained assessments of the team’s design, its level of self-management, the coaching
behaviors of its leader, and its measured performance. She predicted that a team’s design features would affect
its level of self-management and performance more than the leader’s coaching behaviors would–and she was
right. Design was four times as powerful as coaching in affecting a team’s level of self-management, and almost
40 times as powerful in affecting team performance.

Perhaps the most fascinating finding of the Wageman study turned up when she compared the effects of "good"
coaching–such as helping a team develop a task-appropriate performance strategy–with those of "bad"
coaching–such as identifying a team’s problems and telling members exactly how to fix them. Good coaching
significantly helped well designed teams, but made almost no difference for poorly designed teams. Bad
coaching, on the other hand, significantly compromised the ability of poorly designed teams to manage
themselves, worsening an already difficult situation, but did not much affect the self-management of well
designed teams.

A good team design yields a double benefit: Teams are likely to have less need for coaching interventions
because they encounter fewer problems that lie beyond their own capabilities, and the coaching they do receive is
likely to be more helpful because the team is not preoccupied with more basic, structurally rooted difficulties. Over
time, such teams may become skilled at coaching themselves and perhaps even enter into a self-fueling spiral of
ever-increasing team capability and performance effectiveness.

The challenge of designing the optimal team structure for every purpose will always be with us. That’s as it should
be, since a team’s basic design is what provides the platform on which members do their work. Devoting time,
thought, and energy to making that platform as high and sturdy as it can be is always a good investment, whether
people will be working together around the same table or dispersed around the globe, communicating and
coordinating their activities electronically.

J. Richard Hackman is professor of social and organizational psychology at Harvard University and a consultant on teams and
leadership. His latest book is Leading Teams: Setting the Stage for Great Performances (Harvard Business School
Press, July 2002).

http://optimizemag.com/issue/009/culture.htm