Employee surveys, focus groups, benchmarking … employers have used numerous helpful methods to gather input from and about their employee population for decades. Over the past few years, many employers have begun using new data sources along with traditional ones to understand their employees just that much more, arming C-Suites with better information to drive decision making on employee programs.
This will be the first in a series of posts on using new approaches to uncover actionable insights into the workforce. We’ll begin by discussing what those new data sources are and how employers are taking action based on what they learn.
I don’t think we’re in comp & benefits land anymore
Compensation and benefits data is a foundational launching point for designing rewards programs and identifying levers available to address needs. An example of one such lever is salary-based premium tiers for health plan elections, now used by 18% of all US organizations with 500 or more employees, according to Mercer’s National Survey of Employer-Sponsored Health Plans. But moving beyond comp and benefits with additional data points can offer new insights into segments of your workforce. Here are a couple of examples:
- Zip code – Layering in zip code information often reveals unique benefit enrollment and utilization patterns between communities of different income levels or access options. For example, a tendency toward self-care in affluent suburban areas – going to the gym, visiting salons and spas, etc. – is typically linked to higher preventative care visits and screening compliance in health plans. In this example, a planned communication campaign to increase preventive care visits and screening may become an initiative focusing on access points for these services.
- Gender, race & ethnicity – Many employers are actively working toward more inclusive rewards programs in support of the broader goal of creating an inclusive culture and working environment. With detailed demographic data, we can learn whether efforts to provide equitable access have resulted in equitable outcomes – and using these new insights to refine initiatives and move closer to the goal. We will dig deeper into this in a future post.
The scientific method in employee data
To have a seat in just about any executive boardroom, you need to bring data-driven decision making to the table. Be prepared to answer questions such as “Are we interpreting the data correctly?” and perhaps: “Are we exploring and analyzing the right data to get to a concrete answer?” Calling in a data scientist specializing in workforce data can bring the scientific method to human resources – form the hypothesis, make a prediction, test that prediction and iterate. Ensure the actions you take solve for a problem you have.
Consider this case (study) of a data interpretation mistake. One employer reviewed their 401(k) plan data and observed females were contributing less and had lower accumulated balances. Their hypothesis was that females in the workforce could benefit from additional education and engagement in planning for their future, so they devised a campaign directed at females to increase savings rates. Simple enough – except that the hypothesis was based on a misinterpretation of the data. When our data scientist created groups of similarly situated employees and then looked at gender differences, we found that the savings rates and account balances of female employees were about the same as those of males.
So what happened? The workforce overall had more females among their younger and lower-paid segments. In a simple comparison of all females vs all males, the lower savings rates and balances among females were a result of lower tenure and pay rather than gender.
A gender-based communication campaign would have been a waste of time and money. Instead, the employer adjusted their engagement strategy to make the correct data-driven decision for their population.
Developing employee personas
Having access to the same benefits does not provide the same outcomes for employees. Take a closer look at your employee population to understand what is driving different outcomes. While it’s common to work with reporting by department or line of business, or by a single demographic dimension such as age or gender, you will get richer context from your data by grouping people who are similar in multiple aspects.
At Mercer, we use machine learning to create distinct “personas” across the workforce. Those personas may cross job profiles, or there may be multiple personas within a job category. Taking that deeper dive into your population is a key first step. Then assess what you can impact as an employer based on what the data is telling you … after you’ve tested your hypothesis of course!
Stay tuned as we will dive deeper into these topics in future posts.