Beyond Bias: Improving Workplace Diversity in the Age of Algorithms

Sophisticated algorithms can help companies achieve workplace diversity with just a few clicks of a mouse. New research from Summer Jackson, SM ’18, PhD ’21, reveals what happens when hiring managers find these advancements more distasteful than helpful.

From September 2018 to April 2020, Jackson followed a tech startup that was undergoing both exponential growth and calls within the rank and file for greater racial representation. To address both issues, leaders from ShopCo, a pseudonym, needed a new recruiting platform.

Jackson immersed herself in ShopCo’s work culture as decision-makers selected a new vendor for the recruiting platform. The process tested the company’s stakeholders as they dealt with issues of objectification, exploitation, and the specter of affirmative action. Their decisions illuminate how even well-meaning companies can undermine diversity efforts and reinforce inequality within their power structure.

“I had an amazing opportunity to be on the ground with employees and observe in real time their own sensemaking about diversity issues,” says Jackson, who is revising and resubmitting her ShopCo study for publication. During her time at MIT Sloan, she was advised by Kate Kellogg, PhD ’05 (David J. McGrath jr. (1959) Professor of Management and Innovation; Professor, Work and Organization Studies), Ray Reagans (Associate Dean for Diversity, Equity, and Inclusion; Alfred P. Sloan Professor of Management; Professor, Work and Organization Studies), and Ezra W. Zuckerman Sivan (Former Associate Dean for Teaching and Learning; Alvin J. Siteman (1948) Professor of Entrepreneurship and Strategy; Professor, Technological Innovation, Entrepreneurship, and Strategic Management and Work and Organization Studies).

Jackson found that vendors offering recruiting platforms that specialized in the recruitment of underrepresented minority candidates employed either a market-exchange or developmental approach for company hiring managers. The former offered an e-commerce interface that could quickly narrow candidate searches based on several factors, including desired salary and skill set. The market-exchange approach also used an algorithm that predicted people’s race, even if candidates did not self-identify on their applications. ShopCo executives and hiring managers disliked this technique, with some likening the process to “shopping for diversity,” Jackson says.

Vendors that offered the same market-exchange approach, but did not focus specifically on recruitment based on race, elicited no objections from ShopCo leaders. “It was only once race was made salient for them that they started to feel uncomfortable,” notes Jackson.

Vendors offering a developmental approach used a more time-consuming tactic. To create a large multiracial pool of tech workers, these vendors established relationships with affinity groups and historically Black colleges and universities. ShopCo leaders preferred this method and hired a vendor that offered a developmental approach. They would later discover that this approach attracted entry-level candidates, not managerial ones.

The market-exchange model, when applied to diversity recruitment, was morphing into a repugnant market. (Financial exchanges that society finds unpleasant or immoral are known as repugnant markets.) ShopCo leaders were willing to slow their diversity efforts and concentrate their diversity hiring to early-career positions if it meant that they could avoid making market-exchange transactions. Eighteen months after ShopCo hired its new vendor, underrepresented ShopCo tech workers rose from 8 percent to only 10 percent.

Jackson’s ShopCo study continues to inform her research as a faculty member at Harvard Business School. Her unique insights will help business leaders who are ready to take on the hard work of achieving diversity, equity, and inclusion in the workplace.

“When you are faced with a universe of options, paralysis can set in,” Jackson says. “It is easy to write off setbacks as failures and think nothing can be done. I want my research to help decision-makers overcome these challenges.”