As retention issues continue to plague many organizations, the quest to understand why turnover is occurring and how to reduce it is ongoing. The following case study demonstrates how a large health care organization — Organization A — incorporated smarter analytics with data it already had. This approach allowed the organization to not only uncover the causes, but also create actionable plans that ultimately reduced turnover significantly and saved it substantial amounts of money.
Organization A had more than 70,000 employees dispersed throughout the United States. It was struggling with nearly 30 percent turnover among a very critical employee population — registered nurses — in its hospitals. The cost of losing roughly 3,000 registered nurses per year, at a conservative estimate of $40,000 per nurse, was $120 million annually.
Step One: Turnover Risk Analysis
At the immediate conclusion of Organization A’s annual employee engagement survey, Strategic Management Decisions conducted a turnover risk analysis. Our analysis took the items that indicated an employee’s intent to leave (e.g., “I would like to work at this organization for the next three years,” “I would like to stay with this organization”) and regressed the turnover risk items onto the remaining survey categories and items.
This analysis allowed the human resources department to:
See, statistically and objectively, which elements of the work environment measured on the employee engagement survey had the biggest direct impact on turnover risk.
Prioritize interventions to the employee experiences (survey items) that would have the biggest ROI.
Ensure that immediate, proactive action could be taken by leaders at all levels because the employees had not yet left the organization; they were, at that point, at risk for turnover.
Note: It was critical to conduct what is called an “identified survey,” where the survey vendor could track the individual survey taker behind the scenes but also guarantee their confidentiality. No one at the organization can ever see that individual’s survey responses.
Step Two: Post-Turnover Analysis
Another critical analytic approach that the organization used to directly impact turnover was to conduct a post-turnover analysis. In this approach, the organization sent SMD its raw, voluntary turnover data to identify the employee survey responses of those who had left the organization. This data approach allowed the organization to do two critical things:
Compare the survey responses of those who stayed at the organization with those who voluntarily left the organization.
Use logistic regression to understand why those voluntary terminations left the organization. Smarter analytics allow human resources to tell a powerful story to senior leaders in the organization.
Table 1.1 shows the comparison of the scores from the employees who left the organization voluntarily with those from employees who stayed with the organization. It also adds immediate validity to the survey as a predictive tool of outcomes such as turnover. In other words, the survey was a predictor (leading indicator) of something in the future (turnover) and all leaders in this organization had this information at its fingertips before the turnover occurred.
Table 1.1 Key Differences of Employed versus Termed Employees
Category Still Employed Voluntary Difference
Accountability 3.59 3.42 -0.17
Career Development 3.85 3.61 -0.24
Compensation 3.24 2.96 -0.28
Customer Focus 3.58 3.26 -0.32
Engagement 3.92 3.47 -0.45
Job Fit 4.30 4.10 -0.21
Labor Skills 3.93 3.67 -0.26
Management 3.98 3.69 -0.29
Quality 4.10 3.88 -0.21
Senior Management 3.73 3.48 -0.25
Survey Action 3.64 3.35 -0.29
Teamwork 4.31 4.16 -0.15
Work-Life Balance 3.96 3.75 -0.20
Note: Scores are on a 5-point Likert scale.
Figure 1.2 shows the three most important elements (in order of importance) that differentiate the leavers from the stayers. Customer focus, management and job fit were the three largest (and statistically significant factors) that drove employees to leave the organization when these three areas were not perceived positively by those employees. When these elements were perceived positively by employees, they were much more likely to stay with the organization.
Figure 1.2 Key Drivers of Turnover
Taking the analysis a step further, the vendor was able to uncover the specific items under each of the three key survey categories that had the biggest impact on turnover, allowing the organization to become even more focused on what to work on to reduce turnover. The specific items are shown in Table 1.3.
Table 1.3 Item-Level Drivers of Turnover
Priority Category Item
1 Customer Focus My work unit is adequately staffed.
2 Management I am involved in decisions that affect my work.
3 Management I receive useful feedback from the person to whom I report.
4 Job Fit I like the work I do.
Step Three: Actionable Recommendations
Armed with this information, the organization then crafted specific, actionable recommendations for each of these items that managers could implement with guidance and confidence. More importantly, it was done in unison, with the entire organization working in lockstep on just a few key items where there was overwhelming evidence that an impact could be made.
Step Four: Measure Long-Term Impact and Return on Investment
For this organization, the focus was turnover, and the results were quite positive. The results six months after implementing this process were as follows:
Turnover reduction of 5+ percentage points.
$9 million savings (conservative estimate).
Notice how smarter analytics provided more insights with less data and took the organization away from silver bullets and clichés. Instead of trying to work on all the elements of its employee survey, this organization could focus on just three factors and build organizationwide action plans and local management action plans to move the needle in these areas.
It’s critical to conduct analysis after taking action in order to show the actual return on investment of smarter analytics. Smarter analytics saves real dollars, and this can be proven.
This article is based on chapter 11 of the authors’ book, “Predicting Business Success: Using Smarter Analytics to Drive Results.”
Matt Betts is the director of product development at Strategic Management Decisions. Hannah Spell is director of research and analytics at SMD. To comment, email firstname.lastname@example.org.
- Cultural competency leads to meaningful connections
- Learning models in startup tech firms should be 50 percent self-learning, 50 percent social learning
- Leverage AI to foster a learn-it-all culture
- Kraft Heinz CLO Pamay Bassey is reimagining corporate learning with laughter and passion
- Listen: Discover Financial Services’ Jon Kaplan on tuition assistance programs and the importance of building trust among your learning team