Human Resources

Big Data: HR Must Ask the Right Questions

In yesterday’s Advisor, guest columnist Kate McGovern Tornone discussed how improperly trained HR professionals could be buying a lawsuit with big data. Today, Tornone goes over the key risks surrounding the issue and questions HR needs to ask.

Key Risks

During a presentation at the Society for Human Resource Management’s (SHRM) employment law and legislative conference last month, Marko J. Mrkonich, a shareholder at Littler Mendelson, and two other Littler attorneys discussed the risks employers face when using analytics, especially the disparate impact its use could have on protected classes.

To prevent that and other problems, HR departments must be both responsible consumers and responsible users. Employers have access to high-quality information and great tools to analyze it, but they’ll always need a human decision maker involved, Corinn Jackson said. These tools are being developed by data scientists who may not even know about the EEOC, Jackson explained. Zev J. Eigen agreed: “Data scientists don’t know HR.”

When adopting analytics, employers should avoid copying competitors. Just because another company is using a certain product or vendor doesn’t mean it’s safe, the Littler attorneys said. And when you’re considering vendors, ask the right questions (more on this below). Finally, employers also need to watch for future developments, the speakers said. The EEOC is very likely to get involved, they warned, and Congress may as well.

Questions to Ask

While an employer may already use big data in other departments like finance and marketing, its use must be adopted differently in HR, the experts from SHRM and the EEOC meeting said. Employers must remember that this information comes from real people who may be protected by federal nondiscrimination laws.

This means that HR must be included early in the process and must ask the right questions. As Commissioner Charlotte A. Burrows said during the agency’s meeting, you need to know: “Am I getting a good product, or am I buying a lawsuit?”

Questions for Decision Makers

  • What are the questions we’re trying to answer with data?
  • Are managers properly trained on using data?
  • Is our legal counsel trained on its use?
  • Will we inform candidates of its use?

Questions for Prospective Vendors

  • What is being measured?
  • How do we know that we’re increasing diversity?
  • What safeguards can we put in place to prevent discrimination?
  • What evidence has been collected to establish the job-relatedness of the algorithm? Can we see the study?
  • Does the product or service include an audit?
  • What kind of ongoing monitoring do you provide?
  • Will you help us defend discrimination claims or indemnify us against claims?

Eigen noted that you may not get the answers you want to every question—particularly the indemnification request—but you must at least ask the questions to fully understand your risk. Selecting a vendor requires a lot of diligence, he said.

“Don’t let people hand-wave because you don’t fully understand it,” Eigen cautioned. “Ultimately, the responsibility is on you.”