Published on: March 30, 2026 at 12:18 pm
When dealing with people analytics, relying on the notion that numbers don’t lie can be misleading. The truth is that data and algorithms are not always neutral and unbiased.
Academy of Management Scholar Sekou Bermiss of the University of North Carolina’s Kenan- Flagler Business School said that the biggest hurdle that organizations face is “recognizing that you might have bias in your data.”
“This is a big leap, because if you admit to that, then you are admitting to decisions that you’ve made in the past based on this data could have been faulty,” Bermiss said. “To improve, organizations need to recognize that the data that we had available to us about the market was not nearly as good and maybe was biased, and now it’s gotten better.”
Bermiss cited a study examining the promotion rates of men versus women as an example of unintentional biases in an organization’s data.
“The research found that men and women network differently by nature,” Bermiss said. “Those differences can negatively impact some people during promotion time.”
While useful, people analytics can often reflect the biases inherent in the system from which they originate. To confront biases in data, Bermiss suggested “questioning the virtues of your data.”
According to Bermiss, confronting data biases is the biggest hurdle that organizations face, and it can feel daunting.
“If most of your data is probably pretty good, but there are some parts of the data that are probably prone to some sorts of biases, just focus on that,” Bermiss said. “We don’t throw the baby out with the bath water.”
