Predictive Analytics in HR: Part One – Beware the Risks!

Predictive AnalyticsPredictive Analytics. It’s the Holy Grail for the future success of HR. Oh, to be able to successfully, consistently and efficiently hire the best candidates for the right jobs! To retain top talent! To be able to identify who the best managers are and what makes them the best, and conversely identify those who are hampering the business’ progress. To identify unintended business practices such as gender pay gap concerns, unconscious biases, and productivity drains.

Believe it or not, data scientists have the “sexiest job of the 21st century” according to a Harvard Business Review article by Thomas Davenport and D.J. Patil. Predictive analytics and big data is hot. Rife with opportunities and excitement for actionable business insights. Yes, there are tremendous opportunities for process improvements, efficiencies, and improved results using predictive analytics.

Yet, caution is needed as predictive analytics comes with risks that can undermine the upside benefits if not addressed and managed.

The Predictive Analytics World Workforce conference held in San Francisco in May 2017 was run on the premise that “every single product built, service provided, goal achieved, sale made, or error made, every single thing that happens at your company happens because of an employee.” While early predictive analytics approaches were used to drive enterprise performance by predicting customer, voter, debtor and other human behaviors, now predictive analytics can be similarly applied to the workplace to drive performance and lifetime value of the workforce.

This first of two blogs addresses the key risks of predictive analytics in the workforce. Part two will address the opportunities.

Legal Pitfalls

When the likes of top employment law firms Littler and Jackson Lewis roll out major business practices in Big Data (Littler has its Big Data Initiative and Jackson Lewis recently launched its Data Analytics Practice), it’s clear there are significant employment law risks to how big data is used and applied in the workplace.

Eric Felsberg, National Director of the Jackson Lewis Data Analytics Group, spoke to a packed crowd at the Predictive Analytics World  conference on “Legal and Ethical Issues when using Predictive Analytics for Talent Acquisition.”

“The word is out guys,” he proclaimed. “The plaintiff-side is aware of predictive analytic algorithms and lawsuits coming!”

To avoid such lawsuits and risky use of big data, Felsberg outlined key legal issues that must be considered in any predictive big data initiative:

  1. Disparate treatment. An employer can’t take into consideration an individual’s protected characteristics when making a selection decision. Big data can provide details such as race, gender, age and other protected characteristics that can result in decisions causing disparate treatment.
  2. Disparate impact. An apparently neutral practice that results in disproportionate impact on a protected group, such as some preemployment tests and data analysis, can result in disparate impact on certain groups.
  3. Privacy concerns. Big data initiatives may include sensitive or protected data, impacted by such laws as GINA, HIPAA and the ADA.  Employers must be vigilant to protect employee privacy and comply with the myriad of international, federal, state and even local laws in this area.

There are additional risks and concerns that data initiatives must factor. HR and counsel should ensure that the data scientists and other team members do not put the business at risk by a lack of understanding of the legal frameworks relating to EEO, privacy and other areas.

XpertHR’s How to Use Analytics in Recruiting and Talent Management and How to Reduce Unconscious Bias in Recruiting and Hiring provide step-by-step guidance on effectively using analytics in your talent acquisition efforts and reducing unconscious bias.

Getting “Fooled by Randomness”

With all the data, it’s easy to misinterpret, draw (and act on) faulty conclusions.

Eric Siegel, arguably the godfather of predictive analytics, author of Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, and founder of the Predictive Analytics World conference, blogged about Avoiding the most pernicious prediction pitfall and warns “Before you dive in, be warned: This spree of data exploration must be tamed with strict quality control. It’s easy to get it wrong, crash and burn – or at least end up with egg on your face.”

Nassim Nicholas Taleb, author of Fooled by Randomness: The Hidden Role of Chance in Life and in the Markets highlights the unscientific tendency people have to believe in unfounded explanations for their own successes and failures, rather than recognizing the significant role of randomness.  People have a tendency to assume trends, causality, and actions where randomness is actually the primary factor.

With the plethora of data comes responsibility to understand the underpinnings of the data, ensure algorithms are tested and validated, and sanity checking is conducted on resulting decisions.  The clearer and more focused the initial objective, the more likely the results can be targeted and objectives achieved.

Be Prepared to Act on the Results

There is the “creepy side” of predictive workforce analytics – the sense of “big brother” in the kinds of insights the data can provide.

Results and insights may well indicate that certain managers are more effective than others, belying the results of their performance reviews.  Talent acquisition may point to a different profile of person who can be most effective in a role, contradicting the insights and beliefs of the hiring manager. Unexpected benefits, perks and work activities might be a stronger indicator of talent retention than previously understood.

Significant changes in processes, policies, benefits, training and more may result from these insights. Strongly held beliefs, workplace culture and more might be challenged by the insights and need to be adjusted. These might not be simple or easy. Don’t underestimate the follow up actions and commitment to acting on the insights, or expect them to remain static.

XpertHR’s best practice guide on HR Transformation can help plan for and implement transformational activities relating to big data.

Coming Soon –  Predictive Analytics in HR: Part Two – Reap the Rewards!

You may also be interested in our related blogs:
Josh Bersin on HR Analytics 2016: Moneyball Meets the Workplace
Analyze This: Unconscious Bias in Recruiting and What to Do About It
The ROI of Talent Management: Tying the Intangible to the Tangible

, , ,

Leave a comment: