A Data-Driven Approach to Identifying Future Leaders

Rather than just relying on the subjective opinions of executives, some companies are using assessment tools to identify high-potential talent.

Reading Time: 9 min 



An MIT SMR initiative exploring how technology is reshaping the practice of management.
More in this series
Permissions and PDF

Having a strong pool of talent is critical for success in today’s business environment. After all, weak leadership can be one of the biggest impediments companies face. Diversity in leadership is a potential key to unlocking high performance within organizations. However, despite significant attention and investment, the top management of the largest U.S. corporations remains not very diverse.

Our own study of 245 organizations with operations in North America found that while 71% of organizations aspire toward having a diverse culture, only 11% report having one. Moreover, the higher up you look in most organizations, the less diversity you’re likely to see. According to the Center for American Progress, women represent 59% of the college-educated workforce at the entry level, but only 14.6% of executive officer positions are held by women. What’s more, women of color hold only 11.9% of managerial and professional positions.

Clearly, today’s organizations are missing out. Companies that are serious about making tangible and measurable progress toward attracting and retaining the best talent need to make a fundamental shift in how they address diversity. The shift will require organizations to go beyond providing training that educates employees on their unconscious biases; employers will also need to establish data-oriented processes for identifying and promoting leaders.

Many executives believe they are good at identifying leadership talent. However, when asked how they make their decisions, they often cite intuition or “gut” instincts. Social science research, on the other hand, suggests that it is difficult for an individual to judge who will become an effective leader. When hiring, managers tend to select and seek out people who are likable, attractive, nearby, and often similar to themselves. Moreover, they tend to assume that if a person is highly skilled at one thing, he or she is likely to be good at other (even unrelated) tasks. This bias (commonly known as the “halo effect”) explains why, for example, extroverts capable of giving compelling presentations are likely to be viewed as leaders.

Although biases such as these typically operate at the individual level, collectively they form organizational archetypes of what successful leadership looks like. The archetypes may emphasize identity (such as race, gender, and ethnicity) or functional roles within the organization (such as sales or marketing over operations). These archetypes, which are often prone to biases, form the narrative around successful leadership in organizations.

Identifying Leadership Potential

Despite the influence of biases, there are methods organizations are using today to improve decision-making, make promotions fairer, and increase diversity. One tool HR professionals use is called the “nine-box grid.” It helps managers rank and plot an employee’s performance (on the X-axis) against his or her potential for future leadership (on the Y-axis). However, there’s an inherent problem: While current performance can be measured with some objectivity, potential is highly subjective. As such, it is often influenced by the same biases noted above.

We wanted to find a more objective and measurable method of assessing individual leadership potential. Our research teams have conducted more than 100,000 hours of interviews over two decades with senior leaders in 40 countries and performed more than 22,000 leadership psychometric assessments. In addition to our proprietary interview exercises, we have relied on personality questionnaires that measure individual differences, the Hogan Development Survey (a personality questionnaire measuring 11 performance “derailing” factors), and general mental ability tests.

To test the differences between leaders with varying degrees of experience and responsibility, we compared two subgroups from our database in 2016: those who were “managers of managers” (in other words, senior managers) and those who were “managers/supervisors” (in other words, midlevel managers). The individuals came from a wide range of industry sectors, including utilities, telecommunications, financial services, engineering, manufacturing, and professional services, and worked for international companies in a variety of functions (finance, operations, sales/business development, engineering/technical, and general management). Our goal was to identify significant differences between the two groups to help us understand why some individuals rose to leadership positions faster than others.

We found that “managers of managers” differed from “managers/supervisors” on a number of personality, reasoning, and business and leadership dimensions. Specifically, the senior managers were, on average, more driven toward pursuing goals and more capable of managing stress than middle managers. In addition, the senior management group was better at both analytical thinking and creative thinking.

The results highlighted four specific types of potential that predicted how quickly one could lead and take on increasing levels of responsibility. Leaders who scored high in intellectual ability, motivation, social and emotional competence, and an ability to navigate change were predisposed to becoming senior leaders faster than those who scored lower on these factors. These results cut across gender, region, and industry. (See “Four Types of Leadership Potential.”)

Leaders with high intellectual potential, for example, were able to make good decisions by quickly assessing and balancing complex and ambiguous information. Those with high motivation potential showed resilience and confidence in their capacity to lead; those who scored lower on this dimension were less likely to persevere when faced with new and unknown situations. Those who possessed strong people potential were empathetic and more adept at building relationships than their less people-savvy peers. And leaders with high change potential were able to move out of their comfort zones to experiment and take necessary risks; those who were more averse to change had more difficulty going against the status quo.

Using Data-Driven Assessments

Do the findings mean that managers who score lower on the potential metrics can’t lead? No. That said, the research suggests that it may take lower-scoring individuals more time to develop strong leadership skills. By the same token, we saw that simply scoring high on leadership potential isn’t always a guarantee of success. To be successful over the long run, leaders need to have sound business judgment, an ability to set direction, and the capacity to inspire and influence others. Leaders who score high on potential while also exhibiting the right behaviors tend to have the most success.

Making workplaces more diverse and inclusive will require more than simply understanding biases. You also need to correct the biases. The best way to do this is to apply consistent definitions of potential and have clear data-driven assessments for identifying future leaders. Just as companies use data and controls for financial decisions, they should use data to select leaders.

There are many ways in which data-based methods can be used for assessing potential. For example, it’s common to assess people’s motivation by observable behaviors, such as how early they come into work or how late they stay. However, this method has been shown not to be a particularly reliable or valid predictor of motivation. Behavioral interviews or results from psychometric surveys that assess one’s leadership efficacy often provide more-accurate information.

Finding Hidden Gems

How can organizations scale data-based assessments in their leadership selection process? In selecting employees who can be “fast tracked” into leadership positions, we recommend employing screening tools that can reach a large audience. For example, online assessments, which measure employees’ intelligence, motivation, emotional capacity, and ability to operate under high levels of uncertainty, can be used at scale and at relatively low cost.

One organization we worked with wanted to improve the return on what it spent on leadership development. Previously, it had relied on the nine-box grid. But given the grid’s reputation for inconsistency and inaccuracy, it was seen as inadequate. Another organization we worked with wanted to bring greater rigor to how it went about identifying its next generation of leaders. However, since most of its employees worked remotely, it was difficult for management to observe leadership candidates firsthand. The two organizations faced different challenges, but the results were the same: Their talent pools lacked diversity, and leaders had little confidence in their leadership pipelines.

To move forward, we recommended that they use an assessment tool to screen for leadership potential. Currently, the managers use an online assessment to evaluate their team members on the four dimensions we identified in our research: intelligence, motivation, emotional capacity, and ability to navigate change. The assessment produces a score that is used in succession-planning conversations in conjunction with more typical performance review data.

The assessments have a second benefit. They provide organizations with a consistent definition of what leadership potential looks like, regardless of race, gender, or ethnicity. With the help of the tools, management has found that it is able to find “hidden gems” — employees who have low visibility or who previously were not seen to have leadership potential. Remote employees in low-profile departments have also gained visibility. Perhaps most important, the tools have helped educate managers about the difference between an employee’s experience and his or her aptitude to take on new and different responsibilities. Having shared definitions has led to an increased sense of fairness.

Adopting Data-Driven Methods

How can organizations get started in adopting data-driven methods in the leadership selection process? We offer two recommendations.

First, define leadership potential consistently and carefully across the organization. Too many organizations eliminate talented leaders from consideration because the criteria used to determine potential are subjective and inconsistent. If created carefully, a clear, consistent definition of leadership potential can reduce the potential for bias, increase diversity, and save money by ensuring that the organization invests in high-potential employees early in their careers. Confusing current performance with future leadership potential is a common mistake.

Second, use data from multiple points in time. As long as people are involved, leadership selection will be subjective. However, once you have good data, the focus can change. Rather than evaluating leadership potential with past performance ratings, we recommend using data-based assessment tools that measure talent using the same criteria at multiple points in time. Using data analytics derived from these tools helps reduce bias and changes the conversation. Indeed, having results from multiple assessments has the power to reshape talent planning and alter how talent-development decisions are made.



An MIT SMR initiative exploring how technology is reshaping the practice of management.
More in this series

Reprint #:


More Like This

Add a comment

You must to post a comment.

First time here? Sign up for a free account: Comment on articles and get access to many more articles.

Comments (3)
Gerard Francois
In large corporations, whatever tools you put in place to better assess potential leaders, the ultimate decision will always be political and based on networks. That said, it still can be useful at least to eliminate people who do not go through the tools.
Ernesto Torres
Greatly appreciate information on recommended data based assessment tools. Also would like to contact the authors for a one on one conversation on the overall process for implementing  the data driven approach.

Ernesto Torres
Jonathan Obise
Succinct article!
Indeed, a data-driven approach to selecting leaders in organizations would eliminate the bias that so often cloud human judgement.