Predictive Analysis Identifies and Quantifies the Most Important Factors for Employee Satisfaction: Free Case Study
I’ve written in the past about the shortcomings endemic to conventional employee satisfaction surveys, notably:
- These tools measure satisfaction, but they usually don’t unearth the real factors driving it;
- People give inflated positives out of fear that the employer or supervisor may figure out their identity.
But these deficiencies become especially problematic when conducting employee satisfaction and engagement research in the field of healthcare, where personnel issues could potentially indirectly or even directly impact the quality of care patients receive.
So I am excited to tell you about some important work at two award-winning hospitals — Greene Memorial Hospital and the new Soin Medical Center, both part of the world renowned Kettering Health Network.
As the basis for analysis are employee survey comment data spanning now over two years at each respective facility, among staff including doctors, nurses and other personnel. Not surprisingly, the results of these HR survey data — like most ratings-based data — had to that point provided a measure (of employee satisfaction, in this case) with some directional utility, but not much in the way practical insight that could be acted upon.
Fortunately, these employee surveys also contained at least one open-ended question, allowing personnel to respond unaided and in their own words. These responses contain the why behind what is actually going on in both facilities, but the organizations, until now, hadn’t been able to uncover these insights because 1) the sheer scale of this unstructured data was unwieldy, and 2) manual coding and conventional analysis is too time-consuming, labor-intensive and expensive.
Among several unique capabilities, the OdinText technology used to analyze the data can not only analyze unstructured (text) data, but mixed data with both an unstructured and structured component. In contrast to conventional employee satisfaction tracking, using OdinText, Human Resources can therefor identify the actual factors driving satisfaction and the quantified extent to which each factor actually influences satisfaction.
In the case of these hospitals, the OdinText software is focused on a targeted key performance indicator (KPI) — Predicted Satisfaction — using responses to eight different open‐ended questions, which were merged and assessed together, resulting in 21 topics determined to be significant predictors of employee satisfaction. Within that group, OdinText identified the most powerful drivers of satisfaction (and dissatisfaction), providing a tangible foundation for a roadmap to address root issues and improve employee satisfaction at both facilities with implications across Kettering Health Network, in general.
OdinText also executed an emotional analysis both on a segment basis (SMC vs GMH) as well as for the aggregate year-over-year spanning the primary eight emotions expressed in comments and that fundamentally drive human behavior: joy, trust, fear, surprise, sadness, disgust, anger and anticipation. By identifying the prevalent emotions for each segment and any changes year over year, we have an opportunity to explore their sources and to address them.
Was this analysis worthwhile? Did it provide something meaningfully different from what the organization already knew? This quote from Jeffrey Jones, Director of Human Resources at Greene Memorial Hospital and Soin Medical Center, sums it up rather well:
“The magnitude and detail is amazing! This pinpoints exactly areas that we can really work on. Other vendors just give us material and we have to hunt and peck. For not knowing anything about our industry, this is amazing! You know atmosphere, what’s changing and what’s not…This blows me away!”
These hospitals have graciously allowed us to share a case study. I would strongly encourage anyone with a stake in employee engagement and satisfaction to download it here.
As always, please feel free to reach out to me with any questions, and I welcome your comments.