Another SMD Prediction Validated: Bad Analytics Leads to Bad Business

At the end of last year, we developed a “Crystal Ball” prediction series about predictive analytics in HR, and how this approach will impact you, your organization and the HR field in the future. Our first prediction came true in February (see here). We’re happy to report yet another has been validated; the full prediction follows.

An article was written by Lisa Milam-Perez, J.D., in SHRM titled “The Promise and Peril of Big Data,” and it confirms our fear that bad analytics can lead to bad business.

“If HR professionals and hiring managers were to ignore these possibilities and take the data at face value, they would risk making unwise hiring decisions based on erroneous—and biased—assumptions,” states Milam-Perez. See our “Manager Mayhem” section below.

“Among the most pressing concerns inherent in relying on big data is that improperly used HR analytics can result in employment discrimination,” said Milam-Perez. See our “Affirmative Action Wins” paragraph below.

And lastly, Milam-Perez comments, “Because a poorly conceived algorithm can produce discriminatory outcomes, it’s important to make sure you validate all algorithms before acting on them. Consider whether data inputs fairly correspond to desired traits or whether the use of certain data sets skews the analysis.” See our “Buyer Beware” thoughts below. Keep in mind, we do encourage leveraging analytics in HR; you can add significant value to your organization when applied correctly. The key: you must be thoughtful in how you approach it.

Stay up-to-date on all our insights by receiving our blog series straight to your inbox. Sign up here.

Bad Analytics Leads to Bad Business

There are lots of vendors jumping on the analytics bandwagon. Some are just re-branding things they were already doing and calling it analytics. Most of these tools/approaches use very basic methods and don’t connect results directly to business results. Others are pushing new approaches based on machine learning and algorithms. These may be truly innovative, but be very careful in this area. Remember that one of HR’s primary roles is risk mitigation and some of these “new” approaches introduce very real risks. Let’s explore some of the potential risks:

  • Manager Mayhem: HR will go overboard with the data they have and lose great employees by giving managers individual turnover risk metrics. We are already seeing this happen in the market. So, what do you think the likely outcome will be if you give managers an individual level turnover risk metric? A manager might fire or not promote someone because they are a “turnover risk.” Can you say lawsuit?
  • Affirmative Action Wins: Affirmative action lawsuits will be won in court based on a “predictive metric” with adverse impact or discrimination. This is where the machine learning component becomes risky. In this approach, correlations are sought for any and all data points. So, if a font style is a predictor of success in the resume screening process, is it possible the algorithms might have adverse impact? Sure it is. We can’t say for sure when this will happen, but it definitely will.
  • Buyer Beware: Many “predictive” analytic technology companies will fail and go out of business by rushing to get unproven things to market that aren’t actually predictive tools. These companies will fail because (1) clients will discover that bad analytics don’t actually drive success and/or (2) lawsuits. So, don’t just assume that something new and innovative will work. Be an educated buyer for your organization. Don’t make the mistake of buying something that will likely fail or result in significant risk for your organization.