As organizations continue to adopt data analytics and automation across their enterprise, including in marketing, finance, and operations, the use of people analytics is heating up in the human resources (HR) department. According to Gartner, people analytics is defined as “the collection and application of talent data to improve critical talent and business outcomes. People analytics leaders enable HR leaders to develop data-driven insights to inform talent decisions, improve workforce processes and promote positive employee experiences.”

While the management concept of people analytics can be traced back more than 100 years, only recently has the practice taken hold as a tool for HR management. A recent Harvard Business Review (HBR) article examines some newer implementations of HR analytics and how it’s impacting virtually every aspect of the human resource function, including building an equitable workplace. According to the article, “By automating the collection and analysis of large datasets, AI and other analytics tools offer the promise of improving every phase of the HR pipeline, from recruitment and compensation to promotion, training, and evaluation.”

Overcoming Data Analytics Challenges in HR

However, people analytics has not come without its mistakes and mishaps. For example, in 2018, Amazon implemented data analytics into its resume screening process, only for it to be biased against women. Amazon is not alone, as, despite good intentions, many companies have developed new analytics tools and processes that have created biased outcomes. Algorithm bias in HR is a serious issue that business leaders must address by putting safeguards and proper human interventions in place to prevent it. 

To help you further, below are two best practices for reducing bias and ensuring success in your people analytics efforts:

  1. Pay close attention to data demographics – When using AI applications for HR, it’s important to constantly look at “who” is represented in your data from the start. According to the HBR article, HR professionals are often forced to use the same biased data that they are trying to fix as part of the process of building new models. This requires them to continuously look at training data and ensure that its makeup is not skewing results. Tools that can help HR leaders better monitor this process include bias dashboards, which examine how different analytics tools perform across groups when detecting bias.
  1. Test for bias – Increasingly, HR leaders are exploring new methods to test for bias. The HBR article states that “One way to do this is to exclude a particular demographic variable (e.g., gender) in training the AI-based tool but then explicitly include that variable in a subsequent analysis of outcomes.” If a particular demographic correlates highly with certain outcomes, then there’s a possibility that the algorithms are creating bias, requiring further examination. 

In conclusion, it should be noted that humans are still a very critical part of the decision-making process in the HR department – with data analytics playing an innovative, but supporting role. 

If you’d like to learn more, read the complete HBR article.