People Analytics: Tapping into the Promise, and Warding off the Potential Peril

Data Literacy and the Promise

People analytics—the collection, cleaning, analysis, and utilization of talent-related data to improve business performance through data-driven solution-crafting—has exploded in popularity over the past decade. Be it in healthcare, sports, or even the notoriously slow-moving legal industry, now more than ever, people and organizations are relying upon data to guide workplace decision-making. To an extent, rightfully so, as the increased collection and analysis of organizational performance data can aid in the development of novel, organization-sensitive interventions in hiring, retention, promotion, work assignment, and other workforce processes that serve to boost efficiency and productivity. The promise of people analytics is that, with additional organizational information and a team of sophisticated people analytics practitioners, organizations will be better equipped to develop a business and people strategy that effectively services their employees and clientele, thereby improving the performance of the former and the total population of the latter. For the promise of people analytics to be realized on a consistent basis, data literacy—an understanding of how to interrogate and interpret data—is vital, especially as the people analytics’ popularity continues to soar.

In a recent Forbes interview, Dr. Tyrone Smith, Jr., Udemy’s Director of People Analytics, discussed the rapidly increasing importance of people-oriented statistical insights. He noted that, in the wake of the pandemic and in the midst of increased calls for workforce diversity, analytics can help organizations identify and address the evolving needs of their employees in order to “keep their talent practices and pipeline strong.” Moreover, in creating a “data-informed culture,” organizations stand to optimize their workforce through employee-tailored interventions. However, in extolling the virtues of people-centered leadership and the general promise of people analytics, Dr. Smith, Jr. also cautioned fellow practitioners and organizations that, in order to leverage data effectively, “one must understand the problem [that one seeks to solve] and clearly think through the potential solutions before jumping into the data.” In other words, data can be indeterminate and lend itself to a variety of interpretations, so one of primary responsibilities—if not the primary responsibility—of people and organizations that aim to use people analytics to guide decision-making is to do so judiciously. Data literacy is key to that end.

Data Literacy and the Peril

The peril of people analytics—and data-driven decision-making more broadly—lies in the lack of judicious implementation, and this threat is two-fold.

First, data and algorithmic decision-making, though more malleable than human decision-making, can also be biased, so a lack of judicious implementation can contribute to the covert perpetuation of biases and the inequalities that spring therefrom. For example, a recent Maddyness article advocating for more diversity and inclusivity in workforce composition and practices to combat algorithmic bias featured an anecdote about how Kodak’s use of the Shirley Card—a picture of a White female model used to calibrate skin tones, shadows, and light when developing customers’ photos—resulted in Kodak’s film failing to capture the full range of human skin tones accurately in the 1950s and ensuing decades.

Second, the ability of data-driven insights to be implemented effectively without one necessarily understanding the inputs of the insight-producing algorithm, while convenient, can serve to stymie critical thinking, render algorithms prone to abuse by individuals keenly aware of their inputs, and allow for the proliferation of bias as well. In the parlance of HLS professor Jonathan Zittrain, this latter fold of the peril of people analytics is the accrual of an unsustainable amount of intellectual debt—“[a]nswers without theory”—as data literacy lags behind data reliance.

Conclusion: Data Literacy as a Solution

Data literacy empowers people and organizations to root out algorithmic bias, to routinely evaluate data inputs and outputs critically, and to ensure that the design of the analytical models that one uses is properly designed to achieve one’s objectives. Accordingly, to tap into the promise of people analytics and ward off the peril, it is imperative that, in addition to providing business-relevant insights, people analytics practitioners promote the development of the data literacy of both their organizations and society in general. This course and this assignment in particular are fantastic examples of data literacy promotion, which will serve our class well in a world increasingly reliant on data.