Employers are getting serious about HR Analytics (aka People Analytics). At the same time, many of our wellness industry colleagues demonize data, often cloaking their anxieties behind advocacy of humanization

We’ll hear wellness leaders denigrate data because, for example, “it reduces people to numbers” (which could be the slogan for the International Society of Dataphobes).

But if we let our fears, insecurities, or aversions get the better of us, resisting data as a primary language of business, we’ll get left behind in a world where employers, even their HR departments, increasingly see the promise of analytics.

We don’t need to be data scientists. But we can adopt data mindsets, develop our analytical skills, and actively obtain support from data experts and resources. Though I’m by no means a data expert I’ve adopted skills that already deliver value to my clients and their employees. As I mention in a recent Redesigning Wellness podcast:

“I endeavor to be a skills leader, not a thought leader.”

Here are practical examples (ranging from the simple survey data we all collect to integrated big data) of how a data mindset helps enhance wellbeing strategies, without fears of dehumanization:

Small Data, Big Insights

Wellness coordinators voice exasperation with employees failing to participate in the programs that, on surveys, they say they want. But this is usually a flaw with the survey design or how the data is being analyzed, rather than a failure of respondents.

For this example, say a wellness coordinator needs to implement one of two programs and conducts a simple assessment survey with two questions:

  1. “Indicate on a 0-10 scale how much you want a financial wellness workshop”
  2. “Indicate (same scale) how much you want a yoga class.”

Set aside the wording and survey methodology (it’s just an example), and assume we’ve learned from previous surveys that only people who respond with a 9 or a 10 are likely to register for a program. This is based on a common scenario, simplified here for illustrative purposes. Most wellness coordinators will take the mean average score of Likert responses and choose whichever is highest. But that could be the wrong decision — overlooking important “human” sentiment.

Technically, we shouldn’t rely on mean averages of Likert-type scales. Statisticians argue that the median is more valid — but let’s not go there today.

Imagine a small majority of the 100 respondents choose 1, 2, or 3 for yoga, but many choose 9 or 10 for yoga, with a mean average score of 5.95. For financial wellness, the mean average is higher, about 6.22 (more favorable), but the answers cluster around the middle of the range with almost no 9s or 10s, meaning — despite the higher average score — hardly any of the respondents are likely to sign up for financial wellness. Simply looking at a scatter plot of the results would show this. And if I’ve lost you (sorry!), see the chart, below. No advanced statistics necessary. Only the employees represented by the dots on the top two lines are likely to register, in this scenario, regardless of what’s going on in the lines below them. In other words, these respondents either love or hate yoga, but they’re meh about financial wellness.

Scatter plot for simple scenario showing reliance on mean averages can lead to wrong decisions. Scatter plot depicting the data described above. Each dot represents a score submitted by an individual respondent. Financial wellness gets higher average scores, but more people are likely to register for yoga (based on the fact that many of them rated their interest in it 9 or 10, whereas only 2 respondents rated financial wellness above 8). Average score for yoga is 5.95, compared to 6.22 for financial wellness.

Integrate Data

Connect data from health risk assessments (more on this, in a minute), engagement surveys, turnover, disability, and performance. Measure outcomes, using your organization’s existing data, for program quality improvement and to communicate important findings to leadership (e.g. you might find that employees whose overall self-perceived health — an important proxy for wellbeing — or social-support scores improved are also more likely to improve performance appraisal ratings or work engagement scores. You’ll make a better business case — and uncover the strengths and weaknesses of your wellbeing strategy — using integrated data than you can with a touching testimonial).

If your organization isn’t ready to connect data either get the conversation started or find a workaround. If your company has a People Analytics specialist or team, recruit them into collaborating. Be a leader!

For a cool interactive example of data integration, check out the Connectivity and the Chart Generator sections of SAP’s 2018 Integrated Report.

Go Deep with Qualitative Data

Use qualitative data — especially focus groups and key informant interviews — to provide “depth” to numeric data. Wellness and HR leaders almost universally fail to collect qualitative data, or they don’t understand how to use it (e.g. they treat it more like an in-person group survey). Qualitative data is a good way to gain insight into the human story and “the why” underlying numeric data. But you need both.

For some focus group ideas, see the articles I wrote for HES: Part 1, Focusing on Focus Groups and Part 2,  How to Get the Most from Focus Groups. For focus groups and key informant interviews, I urge employers to contract, if they can, with a skilled, external facilitator. And, yes, these are services I provide in my consulting practice.

Opportunities Revealed In Filtered Data

Health risk assessments offer little value as behavioral change tools, but are a rich source of data. Sort stress scores by business function, leader, or job type (to name a few examples) to identify actionable problems among certain roles or managers (e.g. why are your call center employees twice as stressed as your sales employees?). If your HRA vendor can’t help you with this, you need a new one.

As a leader you need to have vision of the possibilities, and it’s essential to go past the self-serving reports that vendors provide. None of the examples in this post, when interpreted properly, dehumanize employees. But if we dismiss data, evangelizing about “meeting employees where they are” while failing to meet our organizations where they are, we do so at our own peril. And the humans in our organizations lose out in the end.