The Network Secrets of Great Change Agents

by Julie Battilana and Tiziana Casciaro

CHANGE IS HARD, especially in a large organization. Numerous studies have shown that employees tend instinctively to oppose change initiatives because they disrupt established power structures and ways of getting things done. However, some leaders do succeed—often spectacularly—at transforming their workplaces. What makes them able to exert this sort of influence when the vast majority can’t? So many organizations are contemplating turnarounds, restructurings, and strategic shifts these days that it’s essential to understand what successful change agents do differently. We set out to gain that insight by focusing on organizations in which size, complexity, and tradition make it exceptionally difficult to achieve reform.

There is perhaps no better example than the UK’s National Health Service. Established in 1946, the NHS is an enormous, government-run institution that employs more than a million people in hundreds of units and divisions with deeply rooted, bureaucratic, hierarchical systems. Yet, like other organizations, the NHS has many times attempted to improve the quality, reliability, effectiveness, and value of its services. A recent effort spawned hundreds of initiatives. For each one, a clinical manager—that is, a manager with a background in health care, such as a doctor or a nurse—was responsible for implementation in his or her workplace.

In tracking 68 of these initiatives for one year after their inception, we discovered some striking predictors of change agents’ success. The short story is that their personal networks—their relationships with colleagues—were critical. More specifically, we found that:

1. Change agents who were central in the organization’s informal network had a clear advantage, regardless of their position in the formal hierarchy.

2. People who bridged disconnected groups and individuals were more effective at implementing dramatic reforms, while those with cohesive networks were better at instituting minor changes.

3. Being close to “fence-sitters,” who were ambivalent about a change, was always beneficial. But close relationships with resisters were a double-edged sword: Such ties helped change agents push through minor initiatives but hindered major change attempts.

We’ve seen evidence of these phenomena at work in a variety of organizations and industries, from law firms and consultancies to manufacturers and software companies. These three network “secrets” can be useful for any manager, in any position, trying to effect change in his or her organization.

You Can’t Do It Without the Network

Formal authority is, of course, an important source of influence. Previous research has shown how difficult it is for people at the bottom of a typical organization chart—complete with multiple functional groups, hierarchical levels, and prescribed reporting lines—to drive change. But most scholars and practitioners now also recognize the importance of the informal influence that can come from organizational networks. The exhibit below shows both types of relationships among the employees in a unit of a large company. In any group, formal structure and informal networks coexist, each influencing how people get their jobs done. But when it comes to change agents, our study shows that network centrality is critical to success, whether you’re a middle manager or a high-ranking boss.

Consider John, one of the NHS change agents we studied. He wanted to set up a nurse-led preoperative assessment service that would free up time for the doctors who previously led the assessments, reduce cancelled operations (and costs), and improve patient care. Although John was a senior doctor, near the top of the hospital’s formal hierarchy, he had joined the organization less than a year earlier and was not yet well connected internally. As he started talking to other doctors and to nurses about the change, he encountered a lot of resistance. He was about to give up when Carol, a well-respected nurse, offered to help. She had much less seniority than John, but many colleagues relied on her advice about navigating hospital politics. She knew many of the people whose support John needed, and she eventually converted them to the change.

Another example comes from Gustaf, an equity partner at a U.S. law firm, and Penny, his associate. Gustaf was trying to create a client-file transfer system to ensure continuity in client service during lawyers’ absences. But his seniority was no help in getting other lawyers to support the initiative; they balked at the added coordination the system required. That all changed when Penny took on the project. Because colleagues frequently sought her out for advice and respected her judgment, making her central to the company’s informal network, she quickly succeeded in persuading people to adopt the new system. She reached out to stakeholders individually, with both substantive and personal arguments. Because they liked her and saw her as knowledgeable and authentic, they listened to her.

It’s no shock that centrally positioned people like Carol and Penny make successful change agents; we know that informal connections give people access to information, knowledge, opportunities, and personal support, and thus the ability to mobilize others. But we were surprised in our research by how little formal authority mattered relative to network centrality; among the middle and senior managers we studied, high rank did not improve the odds that their changes would be adopted. That’s not to say hierarchy isn’t important—in most organizations it is. But our findings indicate that people at any level who wish to exert influence as change agents should be central to the organization’s informal network.

The Shape of Your Network Matters

Network position matters. But so does network type. In a cohesive network, the people you are connected to are connected to one another. This can be advantageous because social cohesion leads to high levels of trust and support. Information and ideas are corroborated through multiple channels, maximizing understanding, so it’s easier to coordinate the group. And people are more likely to be consistent in their words and deeds since they know that discrepancies will be spotted. In a bridging network, by contrast, you are connected to people who aren’t connected to one another. There are benefits to that, too, because you get access to novel information and knowledge instead of hearing the same things over and over again. You control when and how you pass information along. And you can adapt your message for different people in the network because they’re unlikely to talk to one another.

Which type of network is better for implementing change? The answer is an academic’s favorite: It depends. It depends on how much the change causes the organization to diverge from its institutional norms or traditional ways of getting work done, and how much resistance it generates as a result.

Consider, for instance, an NHS attempt to transfer some responsibility for patient discharge from doctors to nurses. This is a divergent change: It violates the deeply entrenched role division that gives doctors full authority over such decisions. In the legal profession, a divergent change might be to use a measure other than billable hours to determine compensation. In academia, it might involve the elimination of tenure. Such changes require dramatic shifts in values and practices that have been taken for granted. A nondivergent change builds on rather than disrupts existing norms and practices. Many of the NHS initiatives we studied were nondivergent in that they aimed to give even more power to doctors—for example, by putting them in charge of new quality-control systems.

A cohesive network works well when the change is not particularly divergent. Most people in the change agent’s network will trust his or her intentions. Those who are harder to convince will be pressured by others in the network to cooperate and will probably give in because the change is not too disruptive. But for more-dramatic transformations, a bridging network works better—first, because unconnected resisters are less likely to form a coalition; and second, because the change agent can vary the timing and framing of messages for different contacts, highlighting issues that speak to individuals’ needs and goals.

Consider, for instance, an NHS nurse who implemented the change in discharge decision authority, described above, in her hospital. She explained how her connections to managers, other nurses, and doctors helped her tailor and time her appeals for each constituency:

I first met with the management of the hospital to secure their support. I insisted that nurse-led discharge would help us reduce waiting times for patients, which was one of the key targets that the government had set. I then focused on nurses. I wanted them to understand how important it was to increase their voice in the hospital and to demonstrate how they could contribute to the organizational agenda. Once I had their full support, I turned to doctors. I expected that they would stamp their feet and dig their heels in. To overcome their resistance, I insisted that the new discharge process would reduce their workload, thereby enabling them to focus on complex cases and ensure quicker patient turnover.

By contrast, another nurse, who led the same initiative at her hospital, admitted that she was handicapped by her cohesive network: Instead of supporting her, the key stakeholders she knew quickly joined forces against the effort. She never overcame their resistance.

The cases of two NHS managers, both of whom had to convince colleagues of the merits of a new computerized booking system (a nondivergent change), are also telling. Martin, who had a cohesive network, succeeded in just a few months because his contacts trusted him and one another, even if they were initially reluctant to make the switch. But Robert, whose bridging network meant that his key contacts weren’t connected to one another, struggled for more than six months to build support.

We’ve observed these patterns in other organizations and industries. Sanjay, the CTO of a software company, wanted his R&D department to embrace open innovation and collaborate with outside groups rather than work strictly in-house, as it had always done. Since joining the company four years earlier, Sanjay had developed relationships with people in various siloed departments. His bridging network allowed him to tailor his proposal to each audience. For the CFO, he emphasized lower product development costs; for the VP of sales, the ability to reduce development time and adapt more quickly to client needs; for the marketing director, the resources that could flow into his department; for his own team, a chance to outsource some R&D and focus only on the most enriching projects.

Change agents must be sure that the shape of their networks suits the type of change they want to pursue. If there’s a mismatch, they can enlist people with not just the right skills and competencies but also the right kind of network to act on their behalf. We have seen executives use this approach very successfully by appointing a change initiative “cochair” whose relationships offer a better fit.

Keep Fence-Sitters Close and Beware of Resisters

We know from past research that identifying influential people who can convert others is crucial for successful change. Organizations generally include three types of people who can enable or block an initiative: endorsers, who are positive about the change; resisters, who take a purely negative view; and fence-sitters, who see both potential benefits and potential drawbacks.

Which of these people should change agents be close to—that is, share a personal relationship built on mutual trust, liking, and a sense of social obligation? Should they follow the old adage “Keep your friends close and your enemies closer”? Or focus, as politicians often do, on the swing voters, assuming that the resisters are a lost cause? These questions are important; change initiatives deplete both energy and time, so you have to choose your battles.

Again, our research indicates that the answers often depend on the type of change. We found that being close to endorsers has no impact on the success of either divergent or nondivergent change. Of course, identifying champions and enlisting their help is absolutely crucial to your success. But deepening your relationships with them will not make them more engaged and effective. If people like a new idea, they will help enable it whether they are close to you or not. Several NHS change agents we interviewed were surprised to see doctors and nurses they hardly knew become advocates purely because they believed in the initiative.

With fence-sitters, the opposite is true. Being personally close to them can tip their influence in your favor no matter the type of change—they see not only drawbacks but also benefits, and they will be reluctant to disappoint a friend.

As for resisters, there is no universal rule; again, it depends on how divergent the change is and the intensity of the opposition to it. Because resistance is not always overt or even conscious, change agents must watch closely and infer people’s attitudes. For nondivergent initiatives, close relationships with resisters present an opportunity—their sense of social obligation may cause them to rethink the issue. But in the case of divergent change, resisters typically perceive a significant threat and are much less susceptible to social pressure. It’s also important to note that the relationship works both ways: Change agents might be reluctant to pursue an initiative that’s opposed by people they trust. They might decide that the emotional cost is too high.

An NHS clinical manager who failed in her effort to transfer responsibility for a rehabilitation unit from a physician to a physiotherapist—a divergent change—described her feelings this way: “Some of my colleagues with whom I had worked for a long time continued to oppose the project. Mary, whom I’ve known forever, thought that it was not a good idea. It was a bit hard on me.”

By contrast, a doctor who launched the same initiative in her organization did not try to convert resisters but instead focused on fence-sitters. This strategy was effective. As one of her initially ambivalent colleagues explained, “She came to me early on and asked me to support her. I know her well, and I like her. I could not be one of the people who would prevent her from succeeding.”

Similarly, John, a member of the operating committee of a boutique investment bank, initiated a rebalancing of traditional end-of-year compensation with a deferred component that linked pay to longer-term performance—a particularly divergent change in small banks that rely on annual bonus schemes to attract talent. His close relationships with several fence-sitters enabled him to turn them into proponents. He also heard out the resisters in his network. But having concluded that the change was needed, he maintained his focus by keeping them at a distance until the new system had the green light.

The important point is to be mindful of your relationships with influencers. Being close to endorsers certainly won’t hurt, but it won’t make them more engaged, either. Fence-sitters can always help, so make time to take them out to lunch, express an authentic interest in their opinions, and find similarities with them in order to build goodwill and common purpose. Handle resisters with care: If you’re pursuing a disruptive initiative, you probably won’t change their mind—but they might change yours. By all means, hear them out in order to understand their opposition; the change you’re pursuing may in fact be wrongheaded. But if you’re still convinced of its importance, keep resisters at arm’s length.

All three of our findings underscore the importance of networks in influencing change. First, formal authority may give you the illusion of power, but informal networks always matter, whether you are the boss or a middle manager. Second, think about what kind of network you have—or your appointed change agent has—and make sure it matches the type of change you’re after. A bridging network helps drive divergent change; a cohesive network is preferable for nondivergent change. Third, always identify and cultivate fence-sitters, but handle resisters on a case-by-case basis. We saw clear evidence that these three network factors dramatically improved NHS managers’ odds of successfully implementing all kinds of reforms. We believe they can do the same for change agents in a wide variety of organizations.

Originally published in July–August 2013. Reprint R1307D