What Makes Leaders Successful: Survey of 1000 CEOs

What Makes Leaders Successful: Revealed by a Survey of How 1,000 CEOs Spend Their Day Reveals

What makes a CEO effective? The question has been studied extensively. Yet we still know fairly little about how CEOs behave day-to-day and how their behavior relates to the success or failure of the companies they run. Previous studies have typically had limitations. Some have been of small samples, or relied heavily on the researchers’ interpretation to classify different “types” of executive.

In new research, we use survey data from over 1,000 CEOs across six countries and the financial performance of their companies to explore these questions. And our evidence suggests that hands-on managerial CEOs are, on average, less effective than leaders who stay more high-level.

Our data set includes every activity a CEO undertakes in a week, as well as whether it was planned ahead of time and who else was involved. We used machine learning to determine which differences in CEO behavior are most important. In effect, we asked the algorithm: If you had to explain CEO behavior by dividing them into two types, how would you do it?

Although the algorithm is completely agnostic, the classification it generates closely resembles John Kotter’s distinction between “managers” and “leaders.” The first type of behavior — managers — includes relatively more plant visits, interactions with employees in supply chain management, and meetings with clients and suppliers. The other type — leaders — includes relatively more interactions with C-suite executives, personal and virtual communications and planning, and meetings with a wide variety of internal functions and external stakeholders. Our data doesn’t insist on classifying CEOs strictly as one type. Instead, we use an index that classifies each CEO as a mix of the two types.

What Do CEOs Do All Day?

On average, about one-quarter of CEOs’ days are spent alone, including sending emails. Another 10% is spent on personal matters, and 8% is spent traveling. The remainder (56%) is spent with at least one other person, which mostly involves meetings, most of which are planned ahead of time. About one-third of the time CEOs spend with others is one-on-one; two-thirds is with more than one other person. (This data includes a CEO’s entire workday, not just time in the office.)

The most common departments for CEOs to meet with are production (35% of time spent with others), marketing (22%), and finance (17%). The most common meetings with outside functions are clients (10%) and suppliers (7%).

But CEOs vary considerably on each of these, and our model divides CEO behavior into the two groups mentioned above — leaders and managers — and then scores each CEO as being degrees of each.

Which CEO Type Is Better for Companies?

When we analyzed CEO type and companies’ financial performance, accounting for other variables including industry, country, and firm size, we found that CEOs who tilt more toward “leader” than “manager” run more-productive and more-profitable companies. And, to our surprise, these previously ignored behavioral differences across CEOs have quite a large association with firm productivity, about one-fifth as big as the impact of a firm’s capital inputs (machinery, equipment, buildings, and so on). Do leader CEOs just happen to work at better companies? We looked at before and after data for firms where a new CEO was appointed, and we found that the appointment of a leader CEO was followed by higher productivity. The effect showed up three years later, which suggests that leaders are doing the hard work of changing companies.

Is One CEO Type Always Better?

So far, you might conclude that the best CEOs don’t get too bogged down in the details of day-to-day management, and instead focus on higher-level leadership tasks, such as convening the heads of the different functions and communicating strategy and vision.

But the picture painted by the data is actually different from this one-size-fits-all approach. Leaders tend to be more prevalent in larger firms and in industries that are, on average, more skill-intensive and complex, while managers tend to run smaller and, to some extent, simpler organizations (i.e., industries characterized by a greater intensity of routine tasks). And plenty of manager CEOs in our data set do run successful firms.

These observations led us to hypothesize that the performance differentials we captured in the data might instead be due to imperfections in the CEO-firm fit. Some companies need great in-the-weeds managers as CEOs, and others need high-level, vision-setting communicators. But, because the market for CEOs is far from perfect, sometimes managers — who are more abundant in our sample than leaders — end up in a leader role, and thus negatively affect the performance of the firm they run.

In support of this hypothesis, we saw that places with less-effective labor markets for CEOs were typically associated with a greater disparity in the performance of firms run by managers, relative to firms run by leaders. Although we can’t say exactly what might drive these allocation frictions, empirically they are important and suggest that the fit between company and CEOs’ behavioral traits really matters for firm performance.

Leaders who set the vision, convene key functions, and communicate effectively can, overall, have a meaningful impact on firm performance, when the setting requires these skills. But just as important is understanding and finding the right fit between the CEO’s leadership style and what the company actually needs.


This blog is authored by Oriana Bandiera, Stephen Hansen, Andrea Prat & Raffaella Sadun and published on HBR. Oriana Bandiera is the Sir Anthony Atkinson Professor of Economics and the Director of the Suntory and Toyota Centre for Economics and Related Disciplines (STICERD) at the London School of Economics. Stephen Hansen is Associate Professor in the Economics Department at the University of Oxford (Oxford, UK). Andrea Prat is the Richard Paul Richman Professor of Business and Professor of Economics at Columbia University. Raffaella Sadun is the Thomas S. Murphy Associate Professor of Business Administration at Harvard Business School, where she studies the economics of productivity, organization, management practices, and information technology.