talent acquisition

People Analytics: Overcoming HR Data Challenges for Program Success

People Analytics: Overcoming HR Data Challenges for Program Success

Recently, I completed the Harvard Business Analytics Program, a combination of online and in-person curriculum exploring current issues and topics in business analytics. A cross-disciplinary program, we studied how analytics are being used to solve challenges and realize efficiencies in areas from supply chain and logistics to marketing and human resources. While each area is defined by unique needs and considerations, they all share a common denominator: the quality and quantity of available data will determine the success of your analytics program.

Yet the need for quality data—and lots of it—is what often sets HR apart from other applications of data analytics. To begin with, data about an organization’s people has rarely been centralized. Instead, it resides across a myriad of departments, such as recruiting, training and development, diversity and inclusion, performance management, workforce planning, benefits administration, labor relations, and organizational development. As such, data collection and storage practices are typically inconsistent, which means that individuals charged with managing a people analytics program will need to spend considerable time just gathering available data and assessing quality prior to creating a central database from which models and analyses can be run.

Clearly Defined Outcomes Must Lead the Way

However, while data availability is paramount to the success of a people analytics program, there’s one other essential step that HR teams need to take before moving forward with data collection and analysis, and that step is clearly defining the outcomes they want to achieve.

Take sourcing, for example, which represents one of the more common applications of data-driven AI in use today, in large part, because the extremely time-consuming task of scanning thousands of resumes makes it ideal to hand over to machines. On the surface, AI appears well-suited to match the aspects of a candidate’s resume not open to interpretation, such as specific skills and certifications, against criteria detailed in a job requisition.

But what if talent acquisition’s desired outcome isn’t simply hiring people with the right skills and certifications but identifying people who also possess the attributes that make it more likely that they’ll stay at the organization over the long term? In this case, what the machines need to learn is now far more subjective because you’re looking to correlate characteristics, such as culture fit, motivation, perseverance and even emotional IQ, with retention.

Matching Outcomes to Available Data

Even if the desired outcome adds a level of complexity, having clearly defined it puts you in a much better position to determine if and where data exists that will support your efforts. In terms of sourcing for skills and retention, an obvious place to go for data might be your organization’s performance assessment records.

At the same time, it’s important to prepare yourself for the reality that the existing data sets may pose several challenges that need to be addressed. First, the performance assessments may not be structured to collect data regarding aspects such as culture fit or emotional IQ. In addition, even the most standardized review methods, including 360-degree feedback reviews, are subject to the biases of reviewers. As a result, few employees believe in either the fairness or accuracy of the review process, which has negative implications for using performance assessment data in its existing form as a source for people analytics.

Yet challenges such as these can and should be overcome. One option to explore would be to overhaul your organization’s performance management process to both reduce bias and to ensure the collection of information that will serve as the data points needed to source for retention. While a process overhaul will delay the launch of your people analytics effort, waiting for the accumulation of good data is always preferred over the alternative because bad data will only yield bad results.

Another option is to consider a different method of data analysis. Instead of looking to performance assessments to uncover the characteristics and attributes of hires who are more likely to stay, you may find that organizational network analysis can achieve this by analyzing employee emails, collaboration tools and calendars to uncover the “influencers” among your teams. Because influencers are often highly engaged—and therefore more likely to stay—this might produce a less biased set of data that can be used to drive retention through sourcing.

Prepare Today for a Faster Tomorrow

The main takeaway for talent acquisition is that, in addition to ensuring that a people analytics program is framed by clearly defined outcomes, TA teams will need to closely examine the availability of quality data in support of the desired outcomes. Arriving at good data may even involve revamping current processes and operations, as well as the consideration of alternative approaches to data collection.

Investing the time now will also help ensure that HR has sound data collection strategies in place to take advantage of the rapid advances in computing power that will lead to the ability for machines to process huge amounts of data and learn far more quickly than what we’re used to today.

Get more GR8 insights regarding effective people analytics practices. Read our Q&A with Randstad’s Graham Trevor and download our E-Book, How People Analytics Is Transforming the Practice—and Influence—of HR.


Filed under