Human Resources professionals often have very good intuition about what strategy will work—and what won’t—in implementing sound initiatives related to improving the productivity, recruitment, retention, leadership development, and the management of benefit costs related to the workforce. Their professional experience, instincts, and daily interactions with employees in identifying and resolving their concerns often provide good insight to inform their decision-making. The workforce of today is so diverse that HR leaders need better information to support their decision-making by recognizing the value of differences, combating discrimination, and promoting inclusiveness in order to avoid major pitfalls. In an era where the underlying data across all of the HR sectors, both internal and external to the organization, is increasing in amount, variety, and type (also known as “big data”), few HR organizations are poised to harness these valuable assets.
At Rensselaer Polytechnic Institute, I, as the vice president for human resources, have harnessed the power behind big data through a collaborative project with Professor Peter Fox, Professor and Tetherless World Research Constellation Chair, who led a team of graduate students who analyzed data on HR sector trends and practices at Rensselaer over a 10-year period (2004-2014). The effort focused on three major areas: compensation and benefits, training and job performance, and paid time-off and insurance claims. The process involved:
- data identification, collection, assembly, and cleaning
- pattern recognition and relationship analysis of the multidimensional data
- analytic model development and statement of findings
- validation of models and practices in the data collection and assembly, and
- suggested actions and process improvements going forward
Even today, HR professionals are typically oriented toward spreadsheets and bar graphs. The Division of Human Resource at Rensselaer went far beyond the typical bar graphs and charts and sought to incorporate as much multidimensional data as possible into an HR process that was predictive and decision-making oriented. Using big data, my staff and I are now able to construct models of what-if scenarios more accurately to evaluate and then identify candidate actions based on the models’ predictions. They have significant, well-structured, and cleaned internal data to supplement their intuition. External data (e.g., from New York’s Capital Region) are yet to be explored.
The collection and analysis of the data resulted in more than 125 discoveries, nearly 50 of which were deemed to be significantly impactful and a few that were considered startling. The purpose of the project was to use big data from a 10-year period to determine myriad factors that could lead to both validation of current HR programs, initiatives, identification of new programs, revisions of HR policies, and precision in identification of key factors that support the recruitment and retention of top talent and leadership at the Institute.
Curtis Powell, SPHR
Vice President for Human Resources