Insights & Publications

Article|McKinsey Quarterly

How we do it: Three executives reflect on strategic decision making

Sir Martin Sorrell, Randy Komisar, and Anne Mulcahy describe how they balance the importance of timely action with the need for thorough, unbiased decision processes.

March 2010

Over the course of their careers, Sir Martin Sorrell, CEO of WPP; Randy Komisar, a partner at Kleiner Perkins Caufield & Byers; and Anne Mulcahy, Xerox’s chairman and former CEO, have made strategic decisions of all shapes and sizes. Their experiences, which the executives share in these three commentaries, highlight a critical challenge: striking the right balance between thorough, unbiased decision-making processes, on the one hand, and timely action, on the other. While there’s no silver bullet, taking concrete steps to cultivate internal critics, safeguard diversity of thought, clarify assumptions underlying different points of view, and force tough choices between business priorities can help.

Read the commentaries from each executive

Sir Martin Sorrell

Learn from mistakes and listen to feedback
—WPP’s Sir Martin Sorrell

Randy Komisar

Balance out biases
—Kleiner Perkins’ Randy Komisar

Anne Mulcahy

Timeliness trumps perfection
—Xerox’s Anne Mulcahy

WPP’s Sir Martin Sorrell: ‘Learn from mistakes and listen to feedback’

The reality is that leaders must, on the spur of the moment, be able to react rapidly and grasp opportunities. Ultimately, therefore, I think that the best process to reduce the risk of bad decisions—whatever series of tests, hurdles, and measuring sticks one applies—should be quick, flexible, and largely informal. It’s important to experiment, to be open to intuition, and to listen to flashes of inspiration. This is not to say the process shouldn’t be rigorous: run the analyses, suck up all the data, and include some formal processes as well. But don’t ask hundreds of people. Carefully sound out the relevant constituencies—clients, suppliers, competitors—and try to find someone you trust who has no agenda about the issue at hand.

There will be mistakes, of course. The truth is we all make mistakes all the time. For instance, I know it’s true that decision makers risk escalating their commitment to losing endeavors that they have an emotional stake in. I know because I’ve been guilty of that myself. However, the only way to avoid making mistakes is to avoid making decisions (or, at least, very few). But then the company would grind to a halt. Instead, learn from mistakes and listen to feedback.

About the author

Sir Martin Sorrell is chief executive officer of WPP, a leading advertising and marketing-services group. Sir Martin actively supports the advancement of international business schools, advising Harvard, IESE, the London Business School, and Indian School of Business, among others.

Kleiner Perkins’ Randy Komisar: ‘Balance out biases’

Before behavioral economics even had a name, it shook up Randy Komisar’s career. He became aware of the then-nascent field while contemplating a graduate degree in economics, losing confidence in the dismal science as a result. Komisar ultimately shifted gears, becoming a lawyer and later pursuing a career in commerce. He cofounded Claris,1 served as CEO for LucasArts Entertainment and Crystal Dynamics, served as “virtual CEO” for a host of companies such as WebTV 2 and TiVo and, since 2005, has been a partner at Kleiner Perkins Caufield & Byers, the Silicon Valley venture capital fund. Along the way, he has developed a distinct point of view on how to create executive teams and cultural environments that are conducive to good decision making. In a recent interview with McKinsey’s Olivier Sibony and Allen Webb, Komisar provided practical advice for senior executives hoping to make good decisions in a world where bias is inevitable.

Harness bias

Rather than trying to tune out bias, my focus is on recognizing, encouraging, and balancing bias within effective decision making. I came to that conclusion as I was starting my career, when I had a chance to work with Bill Campbell, who is well known, particularly in Silicon Valley, as a leader and coach. Bill was the CEO of Intuit (where he’s now chairman), he’s on the Apple board, and he’s a consigliere to Google.3

What I observed back then was that Bill had this amazing ability to bring together a ragtag team of exceptionally talented people. Some had worked for successful companies, some had not. Some had been senior managers. Some had been individual contributors. Everybody brought to the table biases borne out of their domains and their experiences. Those experience-based biases probably are not that different at the psychological level from the behavioral biases that economists focus on today.

Bill was very capable at balancing out the biases around the table and coming up with really effective decisions and, more important, the groundwork for consensus—not necessarily unanimity, but consensus. I liken it to what I have always understood, true or false, about how President Kennedy ran his cabinet: that he used to assemble the smartest people he could, throw a difficult issue on the table, and watch them debate it. Then at some point he would end the debate, make a decision, and move on. It’s also similar to the judicial process, where advocates come together to present every facet of a case, and a judge makes an informed determination. The advocates’ biases actually work to the benefit of a good decision, rather than being something that needs to be mitigated.

Make a balance sheet

There’s a methodology I’ve used within companies for making big, hard decisions that I introduced into Kleiner Perkins and that we have been using lately to help decide whether or not to invest in new ventures. It starts with assembling a group that is very diverse. If you look at my partners, you’d see an unruly gang of talented people with very different experiences, very different domain skills, and, consequently, very different opinions.

Starting with that, the notion is to put together a simple balance sheet where everybody around the table is asked to list points on both sides: “Tell me what is good about this opportunity; tell me what is bad about it. Do not tell me your judgment yet. I don’t want to know.” They start the process without having to justify and thereby freeze their opinions and instead are allowed to give their best insights and consider the ideas of others. Not surprisingly, smart people will uncover many of the same issues. But they may weigh them differently depending on their biases.

We do not ask for anyone’s bottom line until everybody has spoken and the full balance sheet is revealed. I have noticed my own judgment often changes as I see the balance sheet fill out. I might have said, “We shouldn’t do the deal for the following three reasons.” But after creating a balance sheet, I might well look at it and say, “You know, it’s probably worth doing for these two reasons.”

The balance sheet process mitigates a lot of the friction that typically arises when people marshal the facts that support their case while ignoring those that don’t. It also emphasizes to the group that each participant is smart and knowledgeable, that it was a difficult decision, and that there is ample room for the other judgment. By assembling everyone’s insights rather than their conclusions, the discussion can focus on the biases and assumptions that lead to the opinions. An added bonus is that people start to see their own biases. Somebody will stand up and say, “You’re expecting me to say the following three things, and I will. But I’ve also got to tell you about these other four things, which are probably not what you’d expect from me.” Finally, opinion leaders have less sway because they don’t signal their conclusions too early.

Although this may sound tedious and slow, we’re able to move quickly. One reason is that we never try to achieve perfection—meaning 100 percent certainty—around a decision. We just can’t get there in the timeframe necessary. The corollary is that we assume every decision needs to be tested, measured, and refined. If the test results come back positive, we proceed; if they’re negative, we “course correct” quickly.

Create a culture where ‘failure’ is not a wrong answer

The book John Mullins and I recently wrote, Getting to Plan B, presents a way of building a culture of good decision making.4 The very simple premise is that Plan A most often fails, so we need a process by which to methodically test assumptions to get to a better Plan B.

The process starts with an acknowledgment that Plan A probably is based upon flawed assumptions, and that certain leap-of-faith questions are fundamental to arriving at a better answer. If we disagree on the decision, it’s very likely that we have different assumptions on those critical questions—and we need to decide which assumptions are stronger, yours or mine. You end up teasing apart these assumptions through analogs: someone will say, “Joe did something like this.” And then someone else will say, “Yes, but Joe’s situation was different for the following reasons. Sally did something like this, and it failed.” In that process, you don’t get points for being right about your assumption, and I don’t lose points for being wrong. We both get points for identifying the assumption, working on it, and agreeing that the facts have come in one way or the other.

What makes this culturally difficult in larger companies is that there is often a sense that Plan A is going to succeed. It’s well analyzed. It’s vetted. It’s crisp. It looks great on an Excel spreadsheet. It becomes the plan of record to which everybody executes. And the execution of that plan does not usually contemplate testing assumptions on an ongoing basis to permit a course correction. So if the plan is wrong, which it most often is, then it is a total failure. The work has gone on too long. Too much money has been spent. Too many people have invested their time and attention on it. And careers can be hurt in the process. To create the right culture, you have to make very clear that a wrong answer is not “failure” unless it is ignored or uncorrectable.

Intuit, for instance, has found that many early-stage R&D projects went on too long. As in most companies, there was a belief that “we just need to put a little more time and money into these things.” Within about 90 days after I had explained the Plan B process to Intuit, they had broken a set of projects into smaller hypotheses, put together a dashboard process for testing assumptions, and were starting to make go-no-go decisions at each step along the way. They reported back that teams were killing their own projects early because they now had metrics to guide them. And most important, they were not being blamed. Intuit’s culture allows for rapid testing and “failure,” and those who prove responsible and accountable in course correcting are rewarded with new projects.

Listen to the little voice

I think comfort with uncertainty and ambiguity is an important trait in a leader. That’s not to say that they’re ambiguous or uncertain or unclear, but they’re not hiding behind some notion of black or white. When somebody’s shutting down conversations because he is uncomfortable with the points of view in the room or with where the decision may be going, it usually leads to a culture where the best ideas no longer come to the top.

Now, there are cultures where that does seem to work, but I think those are exceptions. Steve Jobs seems to be able to run Apple exceedingly well in large part because Steve Jobs is an extraordinary person. But he’s not a guy who tolerates a lot of diversity of opinion. Frankly, few leaders I meet, no matter how important they are in the press or how big their paychecks are, are that comfortable with diversity of opinion.

I love a leader who changes his or her opinion based upon the strength of the arguments around the table. It’s great to see a leader concede that the decision’s a hard one and may have to be retested. It’s great to see a leader who will echo the little voice in the back of the room that has a different point of view—and thereby change the complexion of the discussion.

When I went to LucasArts, I can remember sitting down one day with a young woman two levels down in the sales organization. I said, “Do you think we could build our own sales force and distribution here? We’ve been going through distributors for a long time. Our margins are a lot smaller as a result. What do you think?” She shut the door, looked at me, and said, “I know that my boss would disagree with me and I know that my peers in marketing disagree with me, but I think we can do it.” And so we did it, and the company’s gross margin line probably grew fivefold in 12 months—all based upon this one little voice in the back of the room. You’ve got to be able to hear that voice.

About the author

Randy Komisar is a partner with Kleiner Perkins Caufield & Byers. He is the author of The Monk and the Riddle and coauthor, with John Mullins, of Getting to Plan B: Breaking Through to a Better Business Model. Randy has been a consulting professor of entrepreneurship at Stanford, where he still lectures.

Xerox’s Anne Mulcahy: ‘Timeliness trumps perfection’

When Anne Mulcahy became CEO of Xerox in 2001—as the company teetered on the edge of bankruptcy—she dove in with the confidence and decisiveness that had typified her career to date. But as she began to engineer the company’s dramatic turnaround, something unexpected happened: Mulcahy started hearing rumblings that her leadership style was too decisive. As she recounts, “I got feedback that between my directness and my body language, within three nanoseconds people knew where I stood on everything and lined up to follow, and that if I didn’t work on it, it really would be a problem.” So Mulcahy listened. “I stopped getting on my feet,” she explains, “and I worked hard at not jumping in, at making people express a point of view.”

This was the first of many lessons about how to ensure high-quality decision making that Mulcahy would go on to learn during her nine years as CEO. In a recent interview with McKinsey’s Rik Kirkland, she distilled five suggestions for other senior leaders.

Cultivate internal critics

My own management style probably hasn’t changed much in 20 years, but I learned to compensate for this by building a team that could counter some of my own weaknesses. You need internal critics: people who know what impact you’re having and who have the courage to give you that feedback. I learned how to groom those critics early on, and that was really, really useful. This requires a certain comfort with confrontation, though, so it’s a skill that has to be developed.

I started making a point of saying, “All right, John-Noel, what are you thinking? I need to hear.” And this started to demonstrate that even if I did show my colors quickly, they could still take me on and I could still change my mind. The decisions that come out of allowing people to have different views—and treasuring the diversity of those views—are often harder to implement than what comes out of consensus decision making, but they’re also better.

Force tough people choices

If you’re sitting around the table with the wrong group of people, no process is going to drive good decision making. You need to lead with people decisions first. One of the easiest mistakes you can make is to compromise on people. It’s very easy to close your eyes and say warm bodies are better than no bodies. The way to counter this bias is to introduce a “forced choice” process. What I mean by this is, you need a disciplined process for forcing discussion about a set of candidates and a position. At Xerox, we developed an HR process that required three candidates for every job.

We also established a group-assessment process, which helped us avoid what I call lazy people decisions, that is, biases against confrontation that could have marginalized the effectiveness of our team. You need to look for people who can strike the hard balance between courage and learning—people who have audacity in their convictions and know when to be unyielding but who are also good listeners and capable of adapting. That is the single most important leadership trait, outside of pure competence.

However you do it, you need to set a context for choice. Once you’ve done that, you must make sure you understand your own criteria for what first-class talent means, and you need to hold yourself accountable for creating a dialogue about it in a very honest way.

Force tough R&D choices

One of the rules of the road should be never to evaluate R&D programs individually. You should always decide on them within the context of an R&D portfolio. There needs to be an “is this better than that?” conversation—no one should get to personally champion his program in a vacuum. Any single idea can look great in isolation.

The portfolio process, like the “forced choice” process for people decisions, is really important because it gives you choices in context. It also takes some of the difficulty of killing individual projects out of the way. And it helps you hold yourself accountable for the full resourcing of the idea. If you decide to invest in a growth opportunity, it’s because you’ve spent a lot of time making sure that it’s resourced properly, that you’ve got the right skill sets to execute it, and that you’re not just saying, “Sure, go off and do it” before you’ve thought through all those considerations.

This process was particularly important for us at Xerox. We kept an investment going for ten years in a technology called Solid Ink, which just came to market this year. We did this by putting a fence around it and a few other strategic priorities that we knew we wanted to protect. Portfolio decision making helped us drive those priorities forward even though most of the people who made the decisions wouldn’t still be in their jobs to see the returns.

Know when to let go

One of the most important types of decision making is deciding what you are not going to do, what you need to eliminate in order to make room for strategic investments. This could mean shutting down a program. It could mean outsourcing part of the business. These are often the hardest decisions to make, and the ones that don’t get nearly enough focus. Making a decision not to fund a new project is not painful. Making a decision to take out a historical program or investment is. It means taking out people and competencies and expertise. That’s much, much harder.

The most difficult decisions are these legacy ones—the historical investments, the things that are just easier to chip away at rather than make a tough decision. This is where we make the most compromises—at the expense of our focus. A great example from Xerox was that it took too long to move from legacy investments in black-and-white imaging to future strategic investments in color and services.

An approach that can help this process involves establishing a decision framework (one akin to a zero-based budgeting philosophy) that says there’s no preconceived commitment to a legacy business. It will get discussed in the context of opportunities for future investment like all the rest. But to make this decision process work, you need to make sure to create a balance between the people who can champion and advocate the future and those who own—and are very invested in—the past.

Strike the right risk balance

Decisiveness is about timeliness. And timeliness trumps perfection. The most damaging decisions are the missed opportunities, the decisions that didn’t get made in time. If you’re creating a category of bad decisions you’ve made, you need to include with it all the decisions you didn’t get to make because you missed the window of time that existed to take advantage of an opportunity.

These days, everyone is risk averse. Unfortunately, people define risk as something you avoid rather than something you take. But taking risks is critical to your decision-making effectiveness and growth, and most companies have taken a large step backwards because of the current climate. I was CEO of Xerox for five years before we really got back into the acquisition market, even though we knew we needed to acquire some things rather than develop them internally. But we got very conservative, very risk averse, and also too data driven. By the time we would reach a decision that some technology was going to be a home run, it had either already been bought or was so expensive we couldn’t afford it.

Decisions have shelf lives, so you really need to put tight timeframes on your process. I would so much rather live with the outcome of making a few bad decisions than miss a boatload of good ones. Some of it flies in the face of good process and just requires good gut. So when trying to take bias out of decision making, you need to be really cautious not to take instinct, courage, and gut out as well.

About the author

Anne Mulcahy is chairman and former CEO of Xerox. She is a director on the boards of Catalyst, Citigroup, Johnson & Johnson, and the Washington Post Company, as well as chair of the board of trustees for Save the Children.

McKinsey Quarterly

McKinsey Quarterly 50th Anniversary art

The Quarterly has been shaping the senior-management agenda since 1964.more