5. Connect the Dots
“Creativity is the power to connect the seemingly unconnected.”
—William Plomer, South African novelist
In the wake of the tragic events of September 11, 2001, various inquiries examined whether the U.S. intelligence community had failed to detect signs that a massive terrorist attack would take place. These inquiries determined that assorted bits of information had emerged in the months prior to 9/11, pointing to the possibility of a terrorist attack on American soil. Disparate pieces of information surfaced in parts of the intelligence community, including the Central Intelligence Agency (CIA), Federal Bureau of Investigation (FBI), and National Security Agency (NSA). However, people did not share key information swiftly and efficiently both within and across these organizations. No single entity or individual ever had access to all the data that suggested a possible attack. In a U.S. Congress report on intelligence failures prior to 9/11, Senator Richard Shelby concluded:
“Our Joint Inquiry has highlighted fundamental problems with information-sharing within the Intelligence Community, depriving analysts of the information access they need in order to draw inferences and develop the conclusions necessary to inform decision-making. The Intelligence Community’s abject failure to ‘connect the dots’ before September 11, 2001 illustrates the need to wholly re-think the Community’s approach to these issues.”1
In this chapter, we will examine why sufficient information sharing does not take place in many groups and organizations. Moreover, we will describe how leaders can facilitate more effective information sharing so that people can “connect the dots” among bits of data that, when synthesized and integrated, signal a potentially significant problem for the firm. Let’s begin by taking a closer look at the intelligence community’s actions prior to the 9/11 terrorist attacks. What signals existed, and why did the intelligence community fail to “connect the dots”?
In the months leading up to the 9/11 terrorist attacks, a steady drumbeat of “frequent but fragmentary” reports began to emerge, pointing to the possibility of a terrorist attack.2 The information came in bits and pieces, and it often proved rather ambiguous. Intelligence officials noted the increased level of al Qaeda propaganda, recruiting, and “chatter” during this time period. Various agencies received tips about possible attacks against American interests, both domestic and abroad, but the information often tended to be quite vague. The CIA did not have specific, credible information about the nature and location of possible attacks. Still, CIA Director George Tenet noted that “the system was blinking red” during the summer months.3 As it turns out, much more specific information about the 9/11 attacks did exist that no one managed to synthesize and integrate in the summer of 2001. In particular, three separate investigations produced information that was not shared broadly, thereby making it impossible for any senior intelligence official to “connect the dots” prior to 9/11. The information stemmed from a CIA investigation of a terrorist meeting in Kuala Lumpur, as well as from investigations taking place in the Phoenix and Minneapolis field offices of the FBI in the summer of 2001. We cannot say for certain that the attacks would have been thwarted if all this information had been shared, but the probability of disrupting the plot certainly would have risen with more effective intelligence sharing and synthesis.
In the late 1990s, a man named Khalid Sheikh Mohammed persuaded al Qaeda leader Usama Bin Ladin to go ahead with a plan to crash airplanes into U.S. buildings. Soon thereafter, Bin Ladin recruited two Saudis—Khalid al Mihdhar and Nawaf al Hazmi—to work on this plan. The two men attended al Qaeda training at Afghan camps and then traveled to Kuala Lumpur for a key meeting about the “planes operation.” The National Security Agency (NSA) learned about the terrorist meeting set to take place, and CIA agents tracked Mihdhar to Kuala Lumpur. The agents observed the meeting, after which the terrorists fled to Thailand. At this time, the CIA knew that Mihdhar possessed a multiple-entry visa allowing him to come to the United States. The CIA did not add their names to the TIPOFF terrorist watch list maintained by the State Department, and it did not tell the FBI that these two individuals had visas permitting entry into the United States. The CIA learned that Hazmi traveled from Thailand to the United States in January 2000. The agency did not inform other units of the federal government about his arrival.
No federal agency tracked Mihdhar and Hazmi when they arrived in California in January 2000. The two men used their real names to open bank accounts, obtain driver’s licenses, and attend flight schools. Mihdhar actually left the country in June 2000. While he was overseas, the CIA linked him to the attack on the USS Cole. However, the agency did not inform the State Department, which issued Mihdhar a new visa permitting him to return to the United States in July 2001.
During the summer of 2001, the CIA’s Bin Ladin unit dug deeper into the connection of these two men to the USS Cole bombing. On August 21, 2001, the CIA contacted the Immigration and Naturalization Service (the INS was then located within the State Department) about Mihdhar and Hazmi. The INS reported that the two men had already arrived in the United States.
In late August, an FBI analyst assigned to work in the CIA’s Bin Ladin unit sent a memo about the two men to the FBI’s New York Field Office. The memo requested that agents begin to search for Mihdhar and Hazmi. No one communicated this information to FBI headquarters. Confusion regarding FBI rules about the handling of intelligence investigations hampered the search. The New York agent assigned to the matter began trying to find the two men, but he did not make substantial progress in the few remaining days prior to September 11.
On July 10, 2001, an FBI agent in Phoenix, Arizona sent a memo to FBI headquarters, as well as to the New York Field Office. His memo stated:
“The purpose of this communication is to advise the Bureau and New York of the possibility of a coordinated effort by Usama Bin Ladin to send students to the United States to attend civil aviation universities and colleges. Phoenix has observed an inordinate number of individuals of investigative interest who are attending or who have attended civil aviation universities and colleges in the State of Arizona... These individuals will be in a position in the future to conduct terror activity against civil aviation targets.”4
One year earlier, Phoenix agent Kenneth Williams had learned that Zackaria Soubra had enrolled in aviation courses at a local university. Soubra had organized a number of anti-U.S. and anti-Israeli rallies in which he called for jihad. An organization to which he belonged had issued several statements suggesting airports as potential targets. Williams’ investigation further revealed that Soubra had a connection to a man who tried to gain access to the cockpit of a commercial airliner in 1999. The FBI actually had interrogated Soubra’s associate, but the Bureau released him when the man stated that he believed the door led to the bathroom, not the cockpit.
Williams interviewed Soubra several times in the spring of 2000, and Williams became concerned when he saw a poster of Bin Ladin in Soubra’s home. Soubra admitted to Williams that he believed United States government entities were “legitimate military targets of Islam.”5 Soubra also informed the agent that he endorsed prior al Qaeda attacks against overseas interests of the United States.
Williams later discovered that other Sunni Muslims with connections to Soubra, and holding similar radical views, had enrolled at a flight school in Arizona. His investigation identified six individuals obtaining flight training and one studying aviation security. With this information, Williams crafted his July 2001 memo to FBI headquarters. The agent made four recommendations. He wanted to compile a list of civil aviation schools in the United States, make contact with those schools, obtain visa information for students in those institutions, and discuss his investigation more broadly with other officials in the intelligence community. The agents in the Bin Ladin unit examined the memo, but they did not share it with their bosses, other units within the Bureau, or senior executives. They did not act on his recommendations. The headquarters agents also did not share the memo with other government agencies, such as the CIA.6
In August 2001, FBI agents in the Minneapolis Field Office learned that a man named Zacarias Moussaoui had enrolled at Pan Am International Flight Academy in Eagan, Minnesota. He stood out there for several reasons. First, he wanted the instructors to teach him how to “take off and land” a Boeing 747 aircraft, even though he did not work for a commercial airline, nor did he intend to seek employment at one of those firms.7 He did not even have a pilot’s license of any kind. Moussaoui also paid cash for the training—the rather hefty sum of $6,800.
FBI agents discovered that Moussaoui believed in the notion of waging jihad against the United States. When questioned about his religious beliefs, Moussaoui tended to become perturbed. Based on his past travels, agents believed that Moussaoui may have traveled to al Qaeda training camps in Afghanistan. When they questioned him about the camps, he became agitated. Minneapolis agents began to suspect that Moussaoui “was an Islamic extremist preparing for some future act in furtherance of radical fundamentalist goals.”8 They worried that he wanted to learn to fly Boeing 747 airplanes because he was planning a major hijacking.
The INS detained Moussaoui on August 17 because he had overstayed his visa. The Minneapolis agents wanted a special warrant to search Moussaoui’s computer and other belongings. A conflict arose between the field office and headquarters regarding whether sufficient evidence existed for a special Foreign Intelligence and Surveillance Act (FISA) warrant. A contentious phone call between headquarters and the field office took place in the final days of August. A headquarters official remarked that the Minneapolis agents were overreacting to the Moussaoui situation. Remarkably, a field office supervisor replied that he was simply trying to ensure that Moussaoui “did not take control of a plane and fly it into the World Trade Center.”9 The headquarters agent replied, “That’s not going to happen. We don’t know he’s a terrorist. You don’t have enough to show he is a terrorist. You have a guy interested in this type of aircraft—that is it.”10 No one briefed senior FBI officials about the Minneapolis investigation prior to the 9/11 attacks.
On September 11, 2001, Mihdhar and Hazmi served as “muscle hijackers” on American Airlines Flight 77, which crashed into the Pentagon. They had come into and out of the United States with relative ease over the past two years, despite CIA knowledge of their terrorist activities and concerns about their ties to the USS Cole bombing. The State Department and FBI did not track these individuals due to the lack of information sharing by the CIA. The individuals identified in the Phoenix memo did not take part in the hijackings. However, at least one student investigated by Kenneth Williams had ties to a 9/11 hijacker. Moussaoui remained in custody on September 11 due to his visa violation, while the FBI had not pursued its investigation due to the lack of a FISA warrant. As it turns out, the FBI later learned that Moussaoui had ties to the terrorists onboard the planes that were hijacked. He eventually pled guilty to conspiring with al Qaeda to attack the United States.
Williams had no knowledge that Minneapolis agents had been conducting an investigation into flight training performed by a radical fundamentalist. Likewise, the Minneapolis field office did not know about the Phoenix investigation. The headquarters agents who denied the Minneapolis field office’s warrant request had no knowledge of the Phoenix memo prior to the attacks. The CIA, which had learned of the Moussaoui investigation prior to 9/11, also had no knowledge of the investigation in Arizona. Williams noted, “It’s been my past experience that the smallest bit of information that comes in could later turn out to be the most important piece of the investigation.”11 Coleen Rowley, Chief Division Counsel of the Minneapolis Field Office, testified to Congress, “The need for people at FBI Headquarters who can connect the dots is painfully obvious.”12 The Joint Inquiry by the U.S. Senate and House Intelligence Committees faulted all the agencies of the intelligence community for a lack of adequate information sharing. It concluded that better dissemination of intelligence might have enabled someone to connect these three investigative threads to one another, as well as to the broader increase in threat reporting that took place in the summer of 2001.
In the case of the U.S. intelligence community, many reasons exist for the inadequate information sharing prior to the 9/11 attacks. Some of those reasons prove to be quite specific to intelligence issues. For instance, officials often guard intelligence data closely because they do not want to compromise its security. They worry about leaks, as well as accidental disclosure of the information. Moreover, officials worry that their sources might be compromised. The rules for securing special warrants on intelligence investigations also hindered information sharing prior to 9/11. While these issues certainly are peculiar to the intelligence community, the 9/11 case also highlights many broader obstacles that impede information sharing in a range of organizations, in both the private and public sectors.
All complex organizations must balance the need for differentiation with the need for integration, as scholars Paul Lawrence and Jay Lorsch have explained.13 Differentiation refers to how firms get work done by creating specialized units that become experts at focusing on particular dimensions of the organization’s mission. In a business, corporations might create specialized units that focus on particular products, customers, and/or geographies. Lawrence and Lorsch argue that the most successful firms balance that specialization with integration—alignment and coordination among the firm’s differentiated units.
Striking the right balance can be quite difficult, though, particularly as organizations grow in size and complexity. Specialized units of an organization often develop their own identity, but people sometimes develop a closer affiliation to their unit than to the organization as a whole. They might go so far as to disparage the culture and identity of other units. Moreover, specialized departments hold divergent interests at times. Conflicting goals and objectives among differentiated units inevitably lead to “thick walls” among the “silos” inside some organizations. No organization with sufficient levels of differentiation can ever create subunits whose goals are completely aligned.
Many organizations benefit from healthy competition among differentiated units, and they purposefully induce such rivalry. Witness the many businesses that rank divisions against one another, handsomely rewarding those units that outperform their peers. At times, however, such competition can become destructive, with one unit benefiting at another’s expense, while collaborative opportunities fall by the wayside.
For the intelligence community prior to 9/11, federal officials tended to place a premium on the need for differentiation, with less focus on integration (as my colleague Jan Rivkin pointed out when we conducted this research together).14 They did so because they wanted multiple sources and perspectives on intelligence. Officials believed that some level of competition among disparate parts of the intelligence community would protect against errors arising from “groupthink” behavior within any particular agency. Moreover, each agency had its own goals and objectives, beyond protecting against terrorist threats. The FBI, for instance, maintained a primary mission of fighting crime; intelligence collection proved secondary prior to 9/11. The Defense Department focused on military preparedness, the CIA on intelligence regarding foreign nation-states, and so on.
Many organizations face a similar dilemma, where the need for a high level of differentiation leads to substantial barriers that prevent adequate integration. For many firms, the lack of integration does not prove costly until a substantial shift occurs in the external environment. That turbulence, and the new threats it creates, often require much more information sharing and coordination among differentiated units. A similar problem occurred with regard to the intelligence community and the 9/11 attacks. The high level of differentiation served the country well, for the most part, during the Cold War. However, the new threat posed by the amorphous al Qaeda organization created a need for much more integration.
Concerns about power also impede information sharing within organizations. Everyone has heard the famous phrase “Knowledge is power”—a statement often attributed to Sir Francis Bacon. Organizational subunits do hoard information at times because they recognize that access to critical data brings with it a certain amount of leverage and perhaps glory too. In the case of the intelligence community, Senator Shelby notes that “knowledge literally is power...the bureaucratic importance of an agency depends upon supposedly ‘unique’ contributions to national security it can make by monopolizing control of ‘its’ data stream.”15 The same power dynamics come into play in business enterprises. For instance, managers may not want to provide others with access to a key customer because they want the credit for the revenues generated by that client, and because they want to maximize their personal importance in the eyes of senior executives.
To this point, we have focused on the information-sharing problems that make it difficult for leaders to “connect the dots” in large, complex organizations. However, information-sharing troubles are not the exclusive provenance of big bureaucracies. Research shows that even small teams face substantial hurdles when it comes to the dissemination and synthesis of information possessed by various members.
In 1985, Garold Stasser and William Titus conducted a study that “challenged the idea that group decisions are more informed than individual decisions.”16 They created four-person groups and asked them to make a decision. Stasser and Titus compared the team decision quality with the choices made by individuals given the same information. When all team members possessed the same information, the group decision exceeded the quality of individual choices. However, the scholars then created a scenario in which each member possessed unique information that his or her colleagues did not have. To arrive at the optimal decision, individuals needed to share their privately held information. The results showed that “groups were more likely to endorse an inferior option after discussion than were their individual members before discussion.”17 The scholars surmised that the groups must have had difficulties surfacing all privately held information.18
A subsequent study by Stasser and colleagues further explored the information-sharing difficulties encountered by small groups. In this research project, the scholars recorded the team discussions. Each individual had information common to all of his or her colleagues, as well as pieces of privately held information not provided to other team members. The scholars discovered the following:
“Effective information pooling is not simply a matter of promoting the discussion of unique information. It also requires that decision-making groups use unique information to inform their decisions... About a third of the common items were repeated at least once after they were first discussed but only a fourth of unique items were repeated. The message was clear. Not only were unique items less likely to be mentioned during discussion, but they were also more likely to be ignored once they were mentioned.”19
Other scholars have confirmed Stasser’s findings.20 In groups where information is distributed unequally among the members, people tend to surface and discuss commonly held information much more so than privately held information. The failure to adequately share, discuss, and analyze uniquely held information inhibits the effectiveness of group problem-solving. One should remember that this information-sharing problem exists in small groups, even in these experimental settings where the team members’ interests are completely aligned, and where power concerns do not seem to be substantial. Amy Edmondson, Michael Watkins, and I have argued that these breakdowns in team decision-making only get worse—perhaps far worse—when group members maintain partially divergent goals and objectives, as is often the case with management teams in organizations.21 In sum, the intelligence community’s failure to “connect the dots” should not simply be attributed to the size and complexity of the bureaucracy. The research shows that many teams, as small as four members, have trouble sharing and integrating information.
Why do members of small groups fail to share information, even when their interests are aligned? Psychologists do not know for certain, but Stasser has argued that perhaps “the bearer of unique information, like the bearer of bad news, incurs some social costs... these social costs may include the necessity of establishing the credibility and relevance of the unique information.”22 Stasser also points out that status dynamics may exacerbate the problem. In a study by James Larson and colleagues, medical residents, interns, and third-year medical school students examined information relating to a patient’s symptoms. They found that residents were substantially more likely to repeat unique information than interns or students. They also asked more questions about unique information that came to light. This finding suggests that lower-status members of a group may feel a particularly heavy burden associated with the social costs of surfacing private information.23 In sum, “connecting the dots” should not be viewed as a problem solely of the large, complex organization; it plagues small groups as well due to fundamental psychological processes and interpersonal dynamics common to many workplace teams.
How can leaders overcome information-sharing barriers so that they can “connect the dots” more effectively? How can they surface and synthesize disparate bits of data so as to identify significant problems and threats in their organizations? Let’s begin by focusing on how leaders can “connect the dots” more effectively within the small teams they manage. Then we will turn to the challenge of information sharing and integration across units of a large organization.
One might think that a consensus-oriented, participative leadership style will encourage more communication and information sharing within your team, while a directive approach hinders information flow. However, the research evidence provides an interesting twist on this conventional wisdom. James Larson and his colleagues contrasted participative and directive leaders. They defined the former as leaders who shared power with their subordinates and who withheld their views until others had been given an opportunity to voice their opinions. These scholars defined directive leaders as those who took charge and stated their views at the outset, while often playing devil’s advocate as other opinions surfaced. Larson and his colleagues discovered that participative leaders tend to surface more privately held information. However, more directive leaders tended to repeat unshared information more often, even when that data did not support their viewpoint. Therefore, a directive leadership approach may encourage a group to analyze unshared data more closely and actually incorporate it into the team’s decision-making process.24
Directive leaders have to be careful, though, because a forceful statement of their views at the outset might quash dissenting views and hinder candid dialogue. Deference to the boss can become quite problematic if leaders become overly directive. Amy Edmondson, Michael Watkins, and I have argued that leaders can avoid that problem by adopting a directive approach to facilitating the group process, while not trying to dictate the content of the decision. Rather than stating their views at the outset, leaders might consider focusing on their facilitation skills. They might intervene actively to surface privately held information, drawing people into a lively and vigorous dialogue. Moreover, they can highlight key pieces of private information and ensure that the team pays adequate attention to this data. These process interventions can help a team “connect the dots” while not suppressing the level of constructive debate.25
What types of behaviors can leaders engage in to encourage more effective handling of information? First, they can “manage airtime,” ensuring that a few people do not dominate the discussion. Leaders should actively seek out the quiet members of the group and encourage them to participate. Second, leaders ought to reiterate and/or paraphrase ideas and statements that emerged quickly but perhaps did not receive adequate attention from most members of the team. By playing back what they have heard, leaders can test their interpretation of the new data, and also ensure that others catch ideas they may have missed at first. Third, leaders should ask clarifying questions as people bring forward new information. They also can question other team members to ensure that they have understood the new points that have been raised. Such inquiries not only clarify understanding, but also flush out additional data that may not have yet surfaced. Finally, leaders should encourage people to express alternative viewpoints and, at times, induce debates so as to surface more data and assumptions.
Perhaps most importantly, leaders need to take time near the end of a decision-making process to highlight the areas of remaining uncertainty that would ideally be resolved before making a decision. In so doing, leaders can ask the following types of questions:
• What else would we like to know in order to make a good decision?
• Have we made some assumptions that could be validated through additional information gathering?
• Would additional data resolve the differences of opinion within the group?
• Where could we find that information?
• Does anyone have access to that type of data?
By asking these kinds of questions, leaders can stress the gaps in the team’s knowledge base and stimulate a search to close those gaps. In so doing, leaders can draw out information that still has not surfaced. Throughout this lively dialogue, leaders must explicitly ask the team to consider the links among various bits of information. They should probe for connections as well as for inconsistencies. They must encourage a process of synthesis and integration, not simply a contentious give-and-take among alternative viewpoints.26
In large, complex organizations, many leaders may adopt structural solutions to information-sharing problems. Often, the new structures focus on the vertical flow of information, and they involve increased centralization and an additional level to the organizational hierarchy. Consider the structural reforms made in the intelligence community after the 9/11 attacks. Government officials decided to create two new entities to facilitate the sharing and integration of intelligence: the Department of Homeland Security (DHS) and the Office of the Director of National Intelligence (DNI).
Both new structures involved a substantial increase in centralization, as well as an additional level in the organizational hierarchy. Clearly, some benefits may emerge. The DNI, for instance, may be able to examine and synthesize various bits of intelligence collected and analyzed by the CIA, FBI, and others. Perhaps they can “connect the dots” effectively among disparate pieces of information. However, serious risks emerge with this type of structure. The additional hierarchy and structural complexity can make it even more difficult for “bad news” or dissenting views to rise to the top of the government bureaucracy. Moreover, the DNI may limit the range of perspectives presented to top officials. In synthesizing what they have heard from various intelligence agencies, the DNI may be tempted to speak with “one voice” to the president and other senior officials. It may be more effective, however, if top decision-makers hear competing and divergent stories from various intelligence agencies. In sum, no structural change provides a perfect solution to the information-sharing problem. All structural reforms have costs and benefits.
Leaders do have an alternative to simple and sometimes problematic structural solutions that focus primarily on the vertical flow of information. They can and should foster the expansion and utilization of informal social networks. As Tiziano Casciaro and Miguel Sousa Lobo wrote in a recent Harvard Business Review article:
“How do you ensure that relevant information gets transferred between two parts of an organization that have different cultures... The answers to such questions lie not in an examination of organization charts but largely in an understanding of informal social networks and how they emerge. Certainly, organizations are designed to ensure that people interact in ways necessary to get their jobs done. But all kinds of work-related encounters and relationships exist that only partly reflect these purposefully designed structures. Even in the context of formal structures like cross-functional teams, informal relationships play a major role.”27
Leaders can foster the formation and development of social networks through activities such as job rotation programs, the creation of informal gathering places, off-site retreats, and leadership development programs. For instance, job rotations expose people to other units in the organization. Those rotations help foster better understanding and appreciation of the work being done in other units. They also help individuals build their social networks. Then, people know precisely who they should ask for critical information in the future.
Sociologist Ronald Burt has argued that we should focus our attention on the precise structure of social networks if we hope to foster more effective information sharing. In most organizations, clusters of people exist within which individuals tend to speak with one another and exchange information quite frequently. However, a cluster may not have strong ties to other groups within the organization. Burt described the paucity of connections between any two groups as “structural holes” in the social network of the organization. He argued that a few individuals may span the holes though. In short, several people within a particular group may possess strong ties to other individuals in an otherwise disconnected cluster. Therefore, they have the ability to serve as “brokers” who facilitate the flow of information across the structural holes. To help connect the dots more effectively, we need to identify and leverage these critical brokers within the social networks of our organizations.28
Beyond social networks, leaders also can use technology and mass-collaboration techniques to marshal the collective knowledge and intellect of many people throughout an organization. In so doing, they can take advantage of the “wisdom of crowds.” In his best-selling book by that name, James Surowiecki argues that we can get effective solutions to many problems by pooling the conclusions and insights of many people without even asking them to communicate and interact with one another. He gives the simple example of the game show Who Wants to Be a Millionaire?. As you may know, contestants have several techniques they can employ to help them answer a difficult question. For instance, they can “ask the audience,” in which case the audience votes on which of the four possible answers is correct. Surowiecki points out that the random group of people in the audience selects the correct answer 91% of the time. He provides many examples such as this one, and he argues that the aggregation of information in this manner provides better-quality decisions than most individual experts can make.29
Many companies have taken advantage of mass collaboration and the wisdom of crowds. For instance, in their book titled Wikinomics: How Mass Collaboration Changes Everything, Dan Tapscott and Anthony Williams describe how a Toronto mining company created a contest. People around the world could examine the company’s geological data and offer a recommendation as to where to search for gold on its properties. The contest yielded solutions that had eluded the company’s in-house experts.30
The intelligence community has adopted “wiki” technology and mass-collaboration techniques as well. Wall Street Journal columnist Gordon Crovitz explains:
“The federal government has launched several wikis, which permit staffers to post information and expand on it until a consensus has been reached. Intellipedia lets 37,000 officials at the CIA, FBI, NSA, and other U.S. intelligence agencies share information and even rate one another for accuracy in password-protected wikis, some ‘top secret.’ Users are told, ‘We want your knowledge, not your agency seal’; indeed, the wiki format may be the best last hope for connecting the dots of intelligence across sixteen different agencies.”31
After 9/11, FBI Director Robert Mueller and his top team had to reshape the Bureau.32 Above all else, they had to change the mindset and the culture within the organization. Throughout its history, the Bureau focused primarily on investigating crimes after they occurred. Now, it had to focus on gathering intelligence proactively so as to prevent crimes before they happened. Detecting potential problems, with little public attention or praise, had to become as highly valued and rewarded as heroically solving crimes and arresting perpetrators. Senior counterterrorism official Arthur Cummings explained the shift toward a “prevent first” mindset that would have to occur:
“The agent’s job now isn’t just to arrest bad guys. It is to understand everything in the terrorist’s head, everything around him, so that we can understand his world and the world of those around him... After 9/11, we have to subordinate the desire to arrest and prosecute quickly because often there is more value to waiting, watching, and collecting more intelligence. If the threat is imminent, we might have to make an arrest so as to disrupt the terrorist plot. But, in many cases, it is more productive to wait. Before 9/11, we were a law enforcement agency with the power to gather domestic intelligence. Now, we have to become a domestic intelligence service with law enforcement powers.”33
Many observers expressed skepticism as to whether the Bureau could accomplish this shift in mindset. As one long-time investigator for a Senate committee responsible for FBI oversight put it, “Mueller is essentially waging two wars at the same time: one against terrorism and one against his own bureaucracy. They are not geared up for prevention of anything. They are geared up to arrest someone after a crime has been committed.”34 In fact, the traditional metrics for measuring agent performance focused on the number of arrests and prosecutions. Agents achieved promotions by solving crimes and “putting bad guys behind bars.” Waiting and watching for months, without making arrests, simply did not fit the “can-do” attitude of most agents. The cultural shift proved to be quite substantial, and it has taken years for the Bureau to shift toward a problem-finding mindset, where people feel equally satisfied and rewarded for detecting potential threats as compared to solving crimes that have already occurred. Mueller has made progress, but much work remains to be done.
All leaders who want to become better problem-finders have to accomplish this shift in mindset. They have to make detecting problems a priority, rather than simply making heroes of those who put out the fires in the organization. They must recognize that those who discover and prevent problems often do not get rewarded, nor are they heralded by their peers. The problem-solvers tend to get the most attention. People often do not even know about the problems that have been prevented.
To become more effective at “connecting the dots,” leaders need to embrace one other mindset change. It involves how we think through problems and situations. According to Roger Martin, Dean of the Rotman School of Management, the most effective leaders nurture and develop their integrative thinking skills.35 By that, he means that they hone their ability to synthesize opposing ideas and discordant information. Martin identifies four ways in which the best integrative thinkers distinguish themselves from conventional ones. First, integrative thinkers search extensively and proactively for “less obvious but potentially relevant factors” in a situation.36 They “don’t mind a messy problem. In fact, they welcome complexity, because that’s where the best answers come from.”37 Second, they do not assume simple linear cause-and-effect relationships. They recognize that most outcomes have multiple causes. Third, integrative thinkers “see problems as a whole.”38 They examine situations and organizations from a systemic perspective rather than looking separately at each part. They understand that parts of a system interact in unpredictable ways, and that one element in a system can magnify the effect of a change in other areas. Finally, integrative thinkers do not simply make either-or choices. They try to generate innovative ideas by combining and synthesizing many pieces of information as well as different and perhaps conflicting ideas.
Martin believes that we can nurture our integrative thinking skills; they are not simply a product of our genes. He’s right. To become more effective at “connecting the dots,” leaders must attract, retain, and develop the integrative thinkers in their organizations, while attending to their own integrative skills. We must learn to see problems as a whole, recognizing that most complex failures (or successes) do not have a single cause. We have to recognize that leaders need to do more than surface crucial information; synthesis indeed represents one of the most important responsibilities of a leader.
1 Shelby, R. (2002). “September 11 and the imperative reform in the U.S. intelligence community: Additional views of Senator Richard Shelby, Vice chairman, Senate Select Committee on Intelligence.” pp. 4–5. Malcolm Gladwell has written an interesting essay in response to Shelby’s conclusions. See Gladwell, M. “Connecting the dots: The paradoxes of intelligence reform.” The New Yorker. March 10, 2003.
2 This section draws heavily on the case study, cited in Chapter 1, about the 9/11 terrorist attacks; see Rivkin, J., M. Roberto, and E. Ferlins. (2006). “Managing National Intelligence (A): Before 9/11.” Harvard Business School Case Study No. 9-706-463. That case study, in turn, draws heavily upon the 9/11 Commission’s work: The 9/11 Commission Report: Final Report of the National Commission on Terrorist Attacks Upon the United States. (2004). New York: W.W. Norton & Company.
3 The 9/11 Commission Report: Final Report of the National Commission on Terrorist Attacks Upon the United States. (2004). New York: W.W. Norton & Company. p. 254.
4 http://www.thememoryhole.org/911/phoenix-memo/.
5 Ibid.
6 The 9/11 Commission Report: Final Report of the National Commission on Terrorist Attacks Upon the United States. (2004). New York: W.W. Norton & Company, p. 272.
7 Ibid, p. 273.
8 Ibid.
9 Ibid, p. 275.
10 Ibid.
11 Poniewozik, J. “The man behind the memo.” Time. November 26, 2001.
12 Colleen Rowley, Testimony to the United States Senate Committee on the Judiciary, Oversight Hearing on Counterterrorism, June 6, 2002.
13 Lawrence, P. and J. Lorsch. (1967). Organization and Environment: Managing Differentiation and Integration. Boston: Harvard Business School Division of Research.
14 Jan Rivkin drew these connections as we discussed how to teach our case study on the 9/11 attacks. We both owe a debt of gratitude to Jay Lorsch, who we have had the privilege of learning from over the years. Lawrence and Lorsch’s contingency theory continues to stand out as one of the seminal events in the history of the evolution of organizational theory. In my view, every doctoral student in management ought to read their 1967 book during their first semester in graduate school. In my case, I had the privilege of taking a course with Jay Lorsch in graduate school, in which he took us through many of the classic readings in organizational theory, stretching all the way back to the works of Frederick Taylor, Mary Parker Follett, Chester Barnard, and others of their time.
15 Shelby, R. (2002). p. 22.
16 Stasser, G. and W. Titus. (2003). “Hidden profiles: A brief history.” Psychological Inquiry. 14(3/4): 304–313. The quotation is from page 304 of that article. The original research was published in 1985. Stasser, G. and W. Titus. (1985). “Pooling of unshared information in group decision making: Biased information sampling during discussion.” Journal of Personality and Social Psychology. (48): 1467–1478.
17 Stasser, G. and W. Titus. (2003). p. 305.
18 Many people hail the promise of teamwork and collaboration in organizations. When individuals pool their knowledge and expertise, presumably a synergy may occur, in which the whole proves greater than the sum of the parts. Unfortunately, research shows that many teams do not achieve this expected synergy. Team researchers use the term “process losses” to describe how a group’s actual performance may fall short of its potential performance. See Steiner, I. D. (1972). Group Process and Productivity. New York: Academic Press. The inability to share and integrate privately held information constitutes one important type of group process loss.
19 Stasser, G. and W. Titus. (2003). p. 308. The study in which the conversations were recorded was published in 1989. Stasser, G., L. Taylor, and C. Hanna. (1989). “Information sampling in structured and unstructured discussions of three and six-person groups.” Journal of Personality and Social Psychology. 57: 67–78.
20 For example, see Larson, J., C. Christensen, A. Abbott, and T. Franz. (1996). “Diagnosing groups: Charting the flow of information in medical decision making teams.” Journal of Personality and Social Psychology. (71): 315–330.
21 Edmondson, A., M. Roberto, and M. Watkins. (2003). “A Dynamic Model of Top Management Team Effectiveness: Managing Unstructured Task Streams.” Leadership Quarterly. 14(3): 297–325.
22 Stasser, G. and W. Titus. (2003). p. 308.
23 Larson, J., C. Christensen, A. Abbott, and T. Franz. (1996).
24 Larson, J., P. Foster-Fishman, and T. Franz. (1998). “Leadership style and the discussion of shared and unshared information in decision-making groups.” Personality and Social Psychology Bulletin. 24(5): 482–495.
25 Edmondson, A., M. Roberto, and M. Watkins. (2003). See also Nadler, D. A. (1996). “Managing the team at the top.” Strategy and Business. 2: 42–51.
26 For those who would like to teach these ideas around how to facilitate the sharing and discussion of shared and unshared information, Amy Edmondson and I have created a web-based leadership and team dynamics simulation. Roberto, M. and A. Edmondson. (2008). Everest Leadership and Team Simulation. Boston: Harvard Business School Publishing. We have published a comprehensive teaching note to help instructors run and then debrief this simulation. During the simulation, teams of students set out to “climb Mount Everest” and must solve a series of problems along the path to the summit. To solve the problems correctly, the team members must share private information with one another and integrate that information effectively.
27 Casciaro, T. and M. Lobo. (2005). “Competent jerks, lovable fools and the formation of social networks.” Harvard Business Review. June: 92–99.
28 Burt, R. (1992). Structural Holes: The Social Structure of Competition. Cambridge, MA: Harvard University Press.
29 Surowiecki, J. (2004). The Wisdom of Crowds: Why the Many Are Smarter Than the Few and How Collective Wisdom Shapes Business, Economies, Societies and Nations. New York: Anchor Books.
30 Tapscott, D. and A. Williams. (2006). Wikinomics: How Mass Collaboration Changes Everything. New York: Penguin.
31 Crovitz, G. “From Wikinomics to Government 2.0.” Wall Street Journal. May 12, 2008. p. A13.
32 This section draws on research that Jan Rivkin and I conducted at the Federal Bureau of Investigation in 2006–2007. We conducted interviews with many officials at the Bureau, stretching from FBI Director Robert Mueller to special agents working in field offices. Based on this research, we published two case studies about the FBI. The first takes the Bureau up through November 2001, with the case ending as Mueller ponders how to restructure the organization after the 9/11 attacks. The second case takes the student through the end of 2006; it charts many of the changes that Mueller enacted in the five years after the 9/11 attacks. The cases provide an unvarnished examination of the successes and failures that the FBI has experienced. We are very grateful to Director Mueller and his colleagues, particularly senior FBI national security branch executive Phil Mudd, for their willingness to be so candid with us about the challenges that the FBI faces in this transformation process. Rivkin, J. and M. Roberto. (2007). “Federal Bureau of Investigation (A).” Harvard Business School Case Study No. 9-707-500; Rivkin, J. and M. Roberto. (2007). “Federal Bureau of Investigation (B).” Harvard Business School Case Study No. 9-707-553.
33 Rivkin, J. and M. Roberto. (2007). “Federal Bureau of Investigation (B).” Harvard Business School Case Study No. 9-707-553. p. 2.
34 Eggen, D. and J. McGee. “FBI rushes to remake its mission.” Washington Post. November 12, 2001. p. A1.
35 Martin, R. (2007). “How successful leaders think.” Harvard Business Review. June: 60–69.
36 Ibid.
37 Ibid.
38 Ibid.