Is there any value to subjectivity in the hiring process?
There is a significant gap in the implementation of data analytics within HR, and in understanding the very real advantages that predictive analysis and big data can have in the HR field. This gap is a by-product of many misconceptions.
One such misconception is that big data analysis will take the “human” out of human resources. Naysayers fear the controversy in trusting a machine-learning algorithm to make the best hiring decision.
Then, there are those who feel that HR departments simply don’t have enough big data to make informed decisions about employees; they feel they cannot confidently infer about the population at large.
These factors are just some of the many that skeptics have about the role of big data within HR.
As a data analytics student at Ryerson University who has worked within an HR setting, I want to pursue how data analytics helps organizations meet benchmarks for success. My research brought up several “success stories” as they relate to data analytics and predictive analysis within HR.
As an example, a retail chain in the US analyzed employee surveys to find out the reason behind high turnover rates. Performing cluster analysis on high-risk factors helped them identify reasons for high turnover rates, and it also helped identify stores/areas at high risk. This in turn helped the retail chain reduce its turnover rate by thirty percent, based on the factors analyzed.
Project Oxygen
The big data analytics trend in HR is being adopted by large organizations. Google is one such multinational company that has been at the forefront of implementing big data analytics to bring improvement to its top-level management. Here is a breakdown of how Google did this, based on my research.
Google data-mined performance appraisals, employee surveys and nominations for top manager awards, resulting in 10,000 “manager behavior” observations.¹ This quantitative data was used by the research team in conjunction with qualitative information from interviews.
The research was a two-step data analysis and is popularly referred to as Project Oxygen, as it brought new life to the management style of the organization.
Analysis 1
HR question to solve: Do managers matter?
Data analyzed:
- Performance reviews (top-down review of managers)
- Staff surveys (bottom-up review of managers)
Data results: Regression analysis performed on data from staff reviews revealed that the best managers were also likely to have higher retention rates among their team members. However, it didn’t answer the fundamental question that Google had sought to answer in the first place: What makes a manager good at Google? In order to answer this, the HR team at Google performed a second set of analyses.
Analysis 2
HR question to solve: What makes a manager good at Google?
Data analyzed:
- Implementation of a “Great Manager” award: Employees had to choose managers that they felt were good and provide reasons for why they felt that way. This data was then analyzed.
- Interviews of managers: The HR team conducted interviews of managers at bottom and top quartiles to understand how they were running their teams. (Obviously, the managers were not told which quartiles they were in.) This data was then analyzed along with the the data from the “Great Manager” awards.
Conclusion
The second analysis was used by the Google HR management team to identify the top eight qualities of good managers, as part of Project Oxygen.
This analysis helped the HR team to identify management qualities that make managers successful at Google, and thus retain staff. This example clearly shows that asking the right questions and using the correct data can help make informed HR decisions. It also shows that data analytics can play an important role in making hiring decisions!
datacritics 
0 comments on “Analytics in HR: Google’s Project Oxygen”