People analytics—a data-driven approach to improving people-related decisions in organizations—is an increasingly popular concept. But what does it really entail in practice? And what do managers need to know about employing people analytics in their work?
In the course of organizing and moderating a recent webinar on people analytics at the MIT Sloan School of Management with other MIT Sloan MBA students, I had the opportunity to listen to and learn from three experts on the subject: Jennifer Kurkoski, Director, People Analytics at Google; Stephanie Lampkin, founder and CEO of Blendoor; and Emilio Castilla, the NTU Professor of Management and a Professor of Work and Organization Studies at the MIT Sloan School of Management.
Here are five of lessons for managers I took away from the insights the expert panelists shared. The webinar, “Managing With Fairness: The Role of People Analytics,” was cosponsored by the Good Companies, Good Jobs Initiative at MIT Sloan, the MIT Sloan People and Organizations Club, and the MIT Sloan Office of Student Life. It was the third installment of the groups’ “Redefining Management: Leadership for Social Progress in Troubled Times” speaker series.
Be skeptical. Continuously question the data. As Jennifer Kurkoski shared, the kinds of people she looks to hire are those who “love data, but don't trust data.” Asking questions allows analysts and decision makers to understand where analyses of data can be useful and where they are likely to be flawed. Active skepticism of data illuminates potential biases and blind spots, and encourages individuals to think through how the data can be misconstrued in decision making. Those looking to leverage people analytics should ask questions about how the data are generated, what might be missing, and why data are incomplete. Proactively altering one’s frame of reference when considering data by asking “what do I not know” or “why might this not make sense” is crucial in order to fully understand both the strengths and shortcomings of data analyses.
Numbers are not everything. Decision makers and managers often want “the number” that will allow for an easy yes/no decision. However, such an approach often leads people to see the “the number” as the truth, when in reality any numerical outcome depends on the underlying process of collecting data and making assumptions. As a result, there can be no definitive number and all outcomes should be evaluated in the context of how they were created. In addition, there is rarely one “right” answer when it comes to making decisions about people. Professor Emilio Castilla emphasized the benefit of qualitative data to complement any quantitative analysis. He noted that people analytics should be as much a qualitative endeavor as it is quantitative and encouraged managers to remember to talk to a wide variety of people when considering people-related decisions.
Consider the larger context. When working with people analytics, it is necessary to have what Stephanie Lampkin, CEO and Founder of Blendoor, called “sociological awareness.” Understanding the factors that have created the world we live in is necessary when evaluating employee outcomes and trying to address problems such as pay equity. Understanding the role race, gender, and the ZIP code people live in play in determining the opportunities afforded them is necessary when evaluating people-related data. Additionally, work in people analytics, and data science in general, should be grounded in a strong ethical foundation. As Lampkin noted, if ethics and sociological contexts are not considered when developing and employing algorithms, we risk proliferating inequity across the globe.
Encourage transparency. People analytics has the potential to make organizations fairer by for example, making disparities in the rates at which people from different demographic groups are hired, compensated, and promoted by an organization visible to stakeholders. Lampkin also observed that having transparency in terms of employee career outcomes and openly defining, and quantifying, what “success” looks like at an organization can aid the creation of true meritocracy over time. The panelists warned, however, that outcomes and data presented are susceptible to being distorted or presented in a way that supports a particular narrative the organization wants to promote. Therefore, there must also be transparency with regards to the process by which the benchmarks are created. Castilla added that it is not only about being transparent with employees but also being held accountable to ensure that the benchmarks, processes, and ultimate outcomes within an organization are actually operating in a fair and unbiased manner.
Focus on effective communication. Ultimately, any people analytics team needs to be able to communicate not only the numbers, but also the process that created those numbers, to decision makers. As Kurkoski pointed out, the mechanics of dealing with data are only half the job. Being able to effectively explain analyses in a way that makes others better informed is paramount to effectively using people analytics. Managers leveraging people analytics should work on developing data communication skills to ensure that nothing gets lost in translation when moving from analysis to action.
The realm of people analytics is dynamic and evolving at a fast pace. However, there was a consensus among panelists that innovation in this space must be matched with a high standard of intellectual rigor. Professionals looking to use people analytics to aid decision making must bring a strong sense of skepticism, a solid ethical foundation, and an understanding of the broader context both within their organizations and in society. It is necessary that those using these tools are able to effectively communicate and be held accountable to ensure that the benchmarks, processes, and outcomes used to drive decision making are fair and unbiased.