In early 2007 Ursula Burns and I had dinner at a French restaurant on the Upper East Side of Manhattan. It was quiet and expensive—which was uncharacteristic for us. But we needed a place where we could talk. At the time, Ursula was a senior vice president at Xerox and the leading candidate to succeed me as CEO. We were about to announce that she was being promoted to president and would join the board of directors—a move intended to signal that the CEO job was hers to win. It seemed like a smooth transition on the surface. But it’s fair to say that we were arguing behind the scenes, though I’d prefer to call it debating.
We were having trouble drawing a line between our roles for the period until she became CEO. We were playing the org-chart game, debating who would report to whom, and who’d have responsibility for what. Ursula rightly wanted as much responsibility as possible as soon as possible so that she’d be well prepared to lead Xerox. As president, she wanted most of the company reporting to her, and that just didn’t work for me. We never let our team see that there was conflict between us, but there was—so we called a time-out. We asked our CFO to come to dinner and act as a facilitator. We talked honestly about how to provide the best transition experience for Ursula while being sensitive to the fact that it was hard for me to step back and give up power. And we talked about how, throughout the process, we would put the company first.
That dinner was just one moment in a long process that ended in July 2009, when Ursula succeeded me as chief executive. Ours was the first woman-to-woman CEO succession in the Fortune 500—and Ursula became the first African-American woman to lead a large U.S. company. I was determined to have a smooth succession, because I’d come into the job the wrong way: I’d taken over for an outsider who resigned after only 13 months as CEO. So I wanted Ursula to be as well prepared as she could. I wanted to move little by little, to make it almost a no-brainer when the time came. And that’s how it felt.
That doesn’t mean it was easy. Succession is very difficult, which is why companies so often fail at it. It takes a tremendous amount of discipline, focus, and support to do well. People have no road map for the process—they don’t know what to expect or how to prepare for it. They also don’t realize how hard succession is on the incumbent CEO. It’s designed to make you able to go away without causing a big impact, and that doesn’t come naturally. Giving up power and accountability is a challenge, but it’s essential to preparing the future CEO for the role.
“Who Is This Person?”
Ursula’s story is extraordinary, considering her background. She grew up in the projects, raised by a single mother who ironed and did chores to support her kids. Her mother used to tell her, “This is where you’re going to grow up, but this is not what defines you.” Ursula eventually received a master’s degree from Columbia.
I first encountered her around 1991. She was serving as executive assistant to a senior Xerox executive, a developmental role for high-potential employees. I was running HR. Ursula stood out. In many meetings she was the most junior person present, and people in that role are expected to listen and be invisible. Not Ursula. She offered opinions. She challenged points of view. Among the senior team there was a sense of “Who is this person?” She was just so vocal. But I liked her authenticity and directness, even if she was a little rough around the edges.
During the mid-1990s I was chief of staff, one level above Ursula. She ran a product-development team. During those years I never looked at her and thought, She could be CEO someday. That’s partly because she was young—still in her thirties. But I certainly knew she was smart and courageous, and that she’d do very well at Xerox if she stayed.
As Ursula rose in the company, she got on recruiters’ radar screens. I don’t know how many outside job offers she received, but I do know we came close to losing her around 2000. She’d received a great offer from a competitor, and she was a bit disillusioned with our company. An accounting scandal was just starting to unfold, and business results were tanking. She also felt that Xerox’s value system and culture were changing. At the time, we had a relatively new CEO, Rick Thoman, who’d been brought in from IBM. Paul Allaire, Thoman’s predecessor as CEO and one of Ursula’s mentors, talked with her about whether she should leave. He told her a management change was likely (Rick resigned in May 2000) and asked if that would give her the confidence to stay at Xerox. She said it would. I became Xerox’s president, and a year later I became CEO. I expanded Ursula’s responsibilities to include manufacturing and the supply chain. I wanted her to be a significant member of my team, and I also wanted to give her an early signal that I believed in her and trusted her with a big part of the operation.
From the moment I stepped into the job, the board began discussing who might succeed me. At that point, in 2001, we focused on four candidates and were considering two scenarios. The first was: What if something unexpected happens tomorrow, and we need someone really experienced who can step in immediately? The second was: What if we had the luxury of developing someone naturally over a long period of time? Ursula fit the second scenario—she needed more time to develop.
Over the next few years she totally reengineered our manufacturing and supply chain. By the time she finished, our operating costs were probably a third of what they’d been five years before. She also overhauled our product-development process. My predecessor used to say that during the 1990s Xerox didn’t launch products, it allowed them to escape. It was true: We brought out only a handful of products in that decade. But Ursula made that her mission, and by 2005 we were introducing 30 or 40 new products a year.
She also built a terrific team of people. Xerox had lost a lot of research and engineering talent during the 1990s, and Ursula groomed a new generation of technology leaders. She racked up an impressive list of accomplishments.
We had regular conversations about her progress. We’re always very direct with each other; there’s little subtlety. If I wasn’t giving Ursula the assurance she needed, she’d be frank about wanting to be told “You’re on track,” which meant “You’re a really good candidate for CEO.” There were never any guarantees, but by mid-decade she was running half the company.
In the conversations I had with the other candidates during this time, I tried to be realistic about their status. I don’t believe in having people face off against each other for the CEO job in a classic horse race. I know that’s the General Electric model, but I didn’t want to lose the three players who weren’t going to get the job. I sent them honest signals about whether they were lead candidates or not, and what the possibilities would be if they didn’t get the job of chief executive, so that they could feel good about their opportunities and the contribution they would make. I worked hard to avoid a contest that would be dysfunctional for the company.
Beware the Tapping Pen
Ursula had appeared before the board of directors in the past, but I made sure she had a lot of visibility in the boardroom over the next few years. She was on her feet a lot—that’s the formal part of it. But board events also include dinners and lunches, and I would seat her next to the members she needed to get to know. We made sure she had a chance to engage with them at other social events, too, and she did very well.
I also tried to work with her on weaknesses. Earlier in her career she didn’t have a good poker face—all her emotions were visible. That’s a big thing for a CEO, because everybody is looking at you. You can destroy someone by showing your emotions, particularly negative ones. It just shuts people down. If you come into the office looking like you’re having a very bad day, everyone reacts to your mood. As chief executive, you have to consciously set the right tone, and Ursula worked to develop that. I’m not perfect at it myself. My team will tell you that when my pen starts tapping in a meeting, it’s time to run for the hills. But CEOs have to manage those unintended displays, because of how much impact they have on other people.
Something else Ursula needed to develop was an ability to engage her team. When you’re really smart and tough and have a lot of responsibility, it can be hard to sit through a long process where everyone voices his or her views—especially if you’ve already reached a decision. A person like Ursula, who really hates to waste time, can be tempted to cut through the crap and reach a conclusion that seems obvious to her. But it’s important to realize that if, as CEO, you make your call too early, your team won’t feel that it owns the outcome. The process can seem like a waste of time, but it’s important. Ursula is very good at it now, but it didn’t come naturally to her early on.
Over the years we’ve worked together, some people have gotten the idea that Ursula and I are close personal friends. We’re not. We don’t socialize away from work. Our families don’t get together. Misconceptions like this seem to be tied to women in leadership positions; people generally don’t assume that male executives are best friends. People tend to look for characteristics in women leaders that would make them seem like really nice people and good friends so that those leaders won’t fit the stereotype of ambitious women.
But we are good business colleagues, and we’ve been able to talk honestly and work through difficult issues. At dinner that night in 2007, we debated various scenarios for splitting responsibility. One involved flipping duties: I’d take over the parts of the company she knew well, and she’d take over the parts I was running. It sounded like a good developmental experience, but it was actually a pretty dumb idea. Putting your two top people on a learning curve isn’t smart, and the CEO isn’t expected to have intimate experience with every piece of the company.
Instead, we decided that Ursula would keep her current responsibilities and would take on all the corporate staff except the CFO. I kept customer operations and our geographic regions, with the understanding that she’d spend more time with the leaders of those units. It was a way of not moving too quickly but giving her access to everything. In some ways we were saying we’d run the company together—make decisions together with no definitive line of separation. We arrived at a really good place.
Smoothing the Path
Under the plan we came up with in 2007, Ursula would have been CEO by early 2009—possibly as early as mid-2008. In fact it took longer. Xerox had been facing securities litigation since 2000, and I wanted to settle that case before Ursula took over so that she could start with a clean slate. We wanted to get our services business up and running, and to have a road map for the acquisition of Affiliated Computer Services. A couple of our senior people were retiring, and we wanted to give their replacements time to get their feet wet so that Ursula wouldn’t have a bunch of newbies on her team. I think she understood that there is no perfect time to become a CEO or to leave the position, but that it would be good to have some of those milestones behind her.
By mid-2008 the board had agreed that Ursula would succeed me. This was largely a formality; the process had been iterative. I’d been discussing our progress with directors over breakfast before every single board meeting, and Ursula was present for many of these discussions. The directors weren’t still asking, “Is she ready?” or “Is this the right decision?” Their confidence in the answer had been established long ago. At that point they were concerned with timelines, communications, and how long I would stay on as chairman.
Looking back on the process, I think the succession conversation between chief executives and their boards needs to start a lot earlier than might feel comfortable. The board must make sure the discussion gets going—it’s unnatural for a CEO to want to initiate it early in his or her tenure. And I think succession should play out over three to five years at a minimum, so the “early and often” nature of the conversation is crucial. Your actions have to support your plans. You have to give the candidates developmental responsibilities and visibility with the board. You also need to establish very clear guidelines with regard to timelines and expectations. I’ve seen lots of companies allow ambiguity on those points, and it’s a problem. And although there may be exceptions, generally you shouldn’t be CEO for more than a decade. You do a disservice to the company if you don’t rotate the leadership.
There may be times when a company finds itself with two equally strong candidates who are in a horse race for the job. I’m glad that wasn’t the case at Xerox. Having winners and losers isn’t good for the company: You wind up losing top talent. (Of the three other candidates for CEO back in 2001, one has retired but two are still at Xerox.) If you can develop one very strong player and some peripheral candidates who could be groomed if necessary, and if you can put your energy into making sure that one person is absolutely the best person for the job, you’ll be in a good place.
By the time Ursula took over, I knew we’d succeeded, because I was needed less and less. That came home to me most vividly at a meeting right after she was named CEO. Everyone was looking at her rather than at me. The whole team’s attention had just shifted, without a lot of drama. That’s the way it should be.