Zelle Hofmann – Kate Bischoff & Judith Langevin
– Big data, data analytics, talent analytics, HR analytics, people analytics. There are many names to describe the use of data science in human resources, and as the list grows, the promise of this technology continues to increase. Data analytics, which quickly gather and analyze vast amounts of historical employee data, can help find the perfect candidate, diagnose performance flaws, identify who might be ready to resign or about to steal your trade secrets, and improve workflow and productivity.
Another list is growing, too: the list of respected institutions and scholars expressing concern about the use of historical employee data in personnel decisions. The New York Times
, The Atlantic
, Harvard Business Review
have all written about the hidden bias contained in historical employee data, and have pointed out that this bias is replicated when the data is used in present day or predictive decision making. These commentators and others argue that the use of historical data may result in discrimination and lack of innovation, precisely what employers want to avoid.
So what should employers interested in this technology do? First, take a deep breath and prepare for some work. Data analytics is very complicated. It is not easy to use well, even though it may eventually make your work easier. HR and IT cannot simply hook up their databases and expect that magical things will happen. Adoption of HR tech requires patience and tenacity from both the employer and the vendor.
What purpose do you want the HR tech to serve? Are you trying to save time? Broaden your ability to search for candidates? Do a better job of providing feedback? Identify performance problems early? There are lots of vendors, lots of products, and lots of HR tasks to accomplish. Be sure that you match up the tasks with the tech.
What does the HR tech require to be effective? Do you have the people and the IT system needed to use the tech well? Is the tech vendor ready to help you figure that out, and ready to help you get the tech up and running?
What promises is the HR tech vendor prepared to make about the effectiveness of the product? What happens if it doesn’t function as promised? Is there a guarantee?
If the HR tech you want to acquire is going to assist in personnel decision making (like hiring, evaluation, promotion or termination) has it been validated? Can you be confident that it uses only unbiased historical data and unbiased criteria? When you use it, will the results withstandEEOC scrutiny
? The Uniform Guidelines on Employee Selection Procedures
describe how to ensure that employment decision tools are not discriminatory. The UGESP helps employers avoid discrimination liability by using only factors or criteria that are “job related” and a “business necessity.” HR tech vendors must
be aware of the need for validation, must know how to validate their product using the UGESP, and must understand that you can’t just validate once. If the HR tech vendor you’re working with hasn’t heard of the UGESP, move on.
Finally, remember that HR can’t do analytics alone. You will need IT and management to be on board with your plans. They need to understand that resources in the form of people, time and money will be required before HR tech will be able to work any magic.
Data analytics is cool. HR tech has the ability to change and improve how HR functions. That said, remember the risks and be prepared to mitigate them.