Howard Gerver is a self-proclaimed human capital data geek. His “day job” specializes in finding innovative and practical ways to save money by identifying “golden nuggets” mined from Big HR Data sets, such as claims and human capital data. A lot of this work includes analytics, claim auditing and eligibility auditing. His “nights and weekend” job focuses on helping clients leverage their HR, Benefits, Leave and Time & Attendance data to help improve compliance with the Affordable Care Act (Obamacare). Throughout his career, he has focused on improving the financial performance of the Payroll, Human Resources and Benefits functions of his clients through advanced technology, process improvement and auditing. In his spare time, he researches new and exciting ways to use Big HR Data to address broader business issues vis-à-vis predictive analytics.
MC: How do you define Big HR Data?
HG: In my humble opinion, Big HR Data applies to the leveraging of “wide and “deep” HR data assets in conjunction with non-HR specific enterprise data as well as external, third party data. Examples of non-HR specific data include: sales, production output, production quality, customer satisfaction surveys and financial results. Examples of third party data include: consumer (e.g. household composition, home ownership, shopping, interests and hobbies), census, health rankings and competitors (i.e., the local labor market).
The key idea is to leverage these collective data sets to:
1) Better understand what has happened (analytics), and
2) Identify what is likely to happen (predictive analytics).
From a business perspective, Big HR Data can be applied to virtually any HR functional area such as talent management, employee engagement, and health benefits.
For example, Big HR Data can address important talent management questions, such as which candidate is likely to succeed? Or, which employees pose a “flight risk?” Regarding employee engagement, Big HR Data can be utilized to identify which employees are more likely to have higher levels of productivity and conversely, which employees are likely to have lower levels of productivity. This information can then be applied to creating budget estimates, for example. Lastly, in the context of health benefits, Big HR Data can be leveraged to identify which employees are likely to have ineligible dependents.
MC: Which types of “golden nuggets” might an organization uncover by mining Big HR data sets?
HG: “Golden nuggets?!” Can I get some fries with that? In all seriousness, “golden nuggets” can be found in many places. To get the best yield, forensic, or CSI-type tactics need to be employed. Leveraging all data, including written information stored in filing cabinets needs to be included.
Example 1 – High Turnover
Case-in-point, while performing a turnover analysis for a manufacturing client, we initially zeroed in on locations with the highest turnover. Interestingly, it turns out 24/7 plants had the highest turnover. Upon further review, we discovered new hires working the “graveyard shift” had the highest resignation rates. The average “newly resigned” employee lasted only 8 weeks. Naturally, HQ sensed the likely suspects were either environmental, the “job” or local management. This was not the case, it turned out the root cause was never documented in any system.
So, what was the culprit? Drum roll please…During the exit interviews it was learned that these “newly resigned” employees NEVER worked the “graveyard shift” before; these employees had no idea how different the “graveyard shift was from their own day-to-day routine and the impact it would have on their family and social life. While each of these same people as candidates needed a job, they didn’t think through the lifestyle difference between working a traditional 9 to 5 job and a “graveyard shift” job. To remedy the problem, management improved the selection process which included adding a “do you have “graveyard shift”” experience question, as well as the inclusion of probing related questions during the interview process.
Net, net – management recognized the value and importance of a richer HR dataset. Moreover, the new owner (which was a private equity firm) enjoyed the productivity and financial improvements derived from these improvements.
Example 2 – Lowering Healthcare Costs
Another “golden nugget” example pertains to reducing healthcare costs (yes, that’s not a typo – Big HR Data can be used to save money in a transparent, immediate and recurring manner!). For example, a large employer with several thousand employees decided to confirm the eligibility of the dependents enrolled in the medical plan. In spite of the compelling ROI, management sensed the audit would be disruptive and costly. Rather than require 100% of the employees to submit supporting documentation, management sensed there would be a way to leverage its HR and claims data. Essentially, this data would be used to audit only those that made most sense to audit.
To bring the vision to life, we were hired to calculate a risk score for each employee and to stratify the population. Inputs to the risk score included two major categories 1) Demographic outliers and 2) Dependents whose medical/Rx claim costs were higher than the respective per capita costs for their dependent category (spouse, domestic partner, young adult, child). External, third party consumer data was also integrated. Employees representing all geographies, divisions and departments were included. “True” random employees were also added to balance the model. The results were stunning. Approximately 90% of the savings were realized simply by auditing 25% of the population. The “icing on the cake” was an interesting discovery. It turns out about 30 of the spouses had gastric bypasses. Ironically, gastric bypasses were excluded from the plan design. At $30,000 each, this drove the savings even higher!
Net, net – management became a strong proponent for Big HR Data.
Example 3 – Insider Threats
Cybercrime continues to be a material threat for ALL employers. Basically, no firm is safe – even from its own employees! While employers have increasingly strengthened physical controls, fortified processes, updated data security programs, and provided employees with requisite training as an effort to mitigate enterprise risk, cybercrime continues to be an area where employers simply continue to feel exposed.
To that end, the C-suite and the Board are under continuous pressure to make sure tangible and intangible assets are not compromised. Ironically, the same employees who are touted as the “number one asset” are under scrutiny. Here’s where leveraging human capital data assets in conjunction with enterprise as well as external, third party data comes in real handy!
First, please allow me to illustrate a realistic scenario. John, a loyal 12-year veteran of the company did not get the promotion he was counting on. Needless to say he didn’t get the big bonus either. In the short 12 years he worked there, John always got top reviews and got good bonuses. The word on the street was he was a strong contender. While his historical performance was solid, his recent results were off. John attributed his recent performance to stronger competitors and management’s unwillingness to make deals.
Much to everyone’s surprise, John abruptly resigned. To exacerbate the issue, not only did he take valuable clients with him (and millions of dollars in business), but he also took trade secrets and all the pertinent client data files. The sad thing is, part of the problem could have been avoided. Here’s how Big HR Data could have tipped off management that John was at-risk.
Using external legal data, management would have seen that John filed for bankruptcy earlier in the year. He also had a DUI just a few months earlier. A pattern analysis of his network usage also would have shown that he accessed folders that he never previously accessed. Moreover, his visits to social media sites, such as Linkedin and job sites including Indeed would have tipped management off that John was potentially, looking for a job.
Again, the use of data could have lessened the severity of what became a big issue. Imagine a world where Big HR Data in conjunction with legal data, network usage data and website visit data co-existed! While it would not change John’s promotability, management could have leveraged the data to then take appropriate measures.
MC: What types of systems might an organization need to organize and cross check their data to confirm it is accurate?
HG: Hmmmmm, those are two great questions! Let’s first explore the systems an organization needs. The answer varies based upon 1) the business question you’re trying to answer, and 2) the data that’s available. At a minimum, we only require a minimum amount of indicative data, such as employee name, address, birth date, hire date, department and title. We can then append third party data to get a more comprehensive understanding of each employee’s demographics, interests and even legal history (legal history includes arrests, bankruptcies, liens and judgments). Other HR and non-HR data as listed below can also provide value.
- Applicant (e.g. previous addresses, work history)
- Time & attendance
- Medical Claims (self-insured plans only)
- Pharmacy Claims (self-insured plans only)
- Workers’ Compensation Claims
- Disability Claims
- Retirement Elections
- Stock Purchase Plan Activity
- Voluntary Insurance Elections
- 401(K) Loans
- Production (e.g. sales, units produced, quality metrics)
- Exit Interviews
While this may appear to be a lot of data, that’s the point! Big HR Data is by default, BIG. It is only when disparate data sets are linked that give real gems the opportunity to “pop.” Ultimately, management will learn which data types have positive and negative affinities; this will enable management to only work with data that provides value.
The second question pertains to data quality. As everyone knows, it’s critical the foundation of the house is solid before the first and second floors are built. The same applies here. The first thing that comes to mind is the use of internal controls. Since many different datasets are likely to be involved, management should first take an inventory of each dataset. This includes taking a point-in-time record count by business unit and/or geography. This will help establish data compatibility. In the event there’s a data gap, either a replacement data set should be created or the gap needs to be accounted for in the analysis.
For employers that don’t have the requisite controls in place, “approximate math” could be used. For example, an employer embarks on a workforce planning exercise and the goal of the project is to identify future workforce gaps. A critical input is skills inventory data. A quick computation reveals the average employee has 10 different skills. Management could then determine whether 10 skills “makes sense.” If it does, great! If not, management would need the employees to update their respective skills before the analysis started. Please note, in this scenario it would also be prudent for management to review sample employee skill inventories to make sure they’re current.
MC: How has the use of Big HR Data by organizations changed over the last five years and how do you see it being most useful to an organization going forward?
HG: Big HR Data itself has not really changed at all. What has changed is the mindset of the HR community. Whereas 5 years ago most analyses were limited to data that was sourced from one system due to system constraints as well as limited IT resources, today power HR users can do their own analyses by using intuitive data visualization tools.
Going forward expanded HR data sets will continue to be leveraged by best practice organizations. Given the pervasive use of analytics in every part of the enterprise, it will be “data or die” as the C-suite will no longer accept we don’t have the data or we don’t have the technology to access the data. Net, net – Big HR Data will continue to play a critical role in helping employers maintain a best practice human capital ecosystem.
Editor’s Note: Proper analysis of Big HR Data can assist organizations in achieving cost-savings across a myriad of programs and departments. It can also provide great insights into the composition and actions of the workforce itself. As technology, and its usage, advances, it will be important for employers to monitor and comply with changing laws and regulations as well as ensure that any personally identifiable information is secured and protected. Employers should also take care that they are not violating applicable laws, such as employment-related or privacy laws, when obtaining data and implementing decisions based on data analytics.