A lot of HR decisions are based on lag data and maybe the time has come for us to get into predictive indicators for our decision making
"Big Data’ has emerged as the new buzzword in management and industry who look to tame the volume, velocity and variability of massive data to reveal hitherto insights, which would provide organizations with a competitive edge. After all, analysis and insights from Big Data is what has been the genesis of a large number of service and product organizations. Trend patterns emerging from user actions have resulted in crafting highly personalized user experience and the new wave of advertising. This has been leveraged successfully by the large web-based startups like Amazon, Facebook, Google and Yahoo.
The buzz around Big Data is enough to make sensible people head for the hills! Terms in the Hadoop bestiary like HDFS, Mapreduce, Pig, Zookeeper, Hive, Mahout , Sqoop, Flume are enough to deter most HR professionals from testing these waters. So what competencies are required to become a true Big Data master? Since the differentiation emerges from the speed of the insights provided, agility is a key requirement. Most HR functions are sitting on goldmines of large employee data and each function within HR can independently explore what insights could be gleaned from the volume, velocity and variety of data accessible to them. The variety aspect of the data would refer to the multiple sources of data, which does not necessarily fall into neat relational structures. Moreover, experienced practitioners have an intuitive understanding of which levers in their operations could produce strategic outcomes for the business and the organization. All it requires is a questioning bent of mind, to challenge existing paradigms of efficiency and effectiveness and evaluate the impact of these initiatives through the arsenal of Big Data.
I could share a personal example of how a Talent Acquisition function I was part of, was actually working blindly against non-validated perceptions and biases which were actually counter-productive to the organizations’ interests. The business head had given us a mandate to improve quality of hire, with a clear directive to increase the intake from top tiered organizations. The rationale behind this mandate seemed intuitive, with talent at top tiered organizations having gone through sufficient screening process to ensure high caliber talent. Additionally, having worked at these top tier organizations, the assumption was that these hires would produce non-linear outcomes in terms of quality of performance and productivity. The Talent Acquisition function took affirmative action and congratulations were generously awarded by all, when the intake of top tier organizations was increased from 23 percent of lateral hires to 78 percent of lateral hiring. Naturally, this came at a significant increase in the salary cost. However, reviewing the hiring data of over 5000 hires over a couple of years and correlating with performance and retention indicated that there was no significant correlation of pedigree of company hired from with the performance. What this meant was that with no significant productivity increase, we had, in effect, increased the cost base for the organization by 5 percent and actually got appreciation for it! Naturally, corrective strategies were applied and hiring was then focused on increasing tier hire, but within the same salary grids applicable to the larger organization and no deviations in cost allowed.
The fundamental insight from such an example is the strong feedback loop used to link data streams to build powerful and actionable insights. The purpose of Big Data is not to generate more data but to produce actionable outcomes. The example cited here is not necessarily restricted to the domain of talent acquisition. For instance, HR Managers could be encouraged to look at trends within performance management systems, link with reward and recognition or correlate with training interventions and retention statistics. Depending on the business problem critical to the specific organization, HR managers should start use the “drivetrain approach” in conjunction with a model assembly line and this can be used to link a predictive model to actionable outcomes. A lot of HR decisions are based on lag data and maybe the time has come for us to get into predictive indicators for our decision making. We can all start with asking simple questions around “What objectives is our HR function trying to achieve? What levers do we have at our disposal to achieve this objective?”
All of us would be amazed at the possibilities which emerge from these simple questions.