The study was conducted to assist employers to understand their employees better through the ‘burnout’ lens. Data Science techniques were used to curate recommendations for overall employee wellbeing.
A survey of 1,000 US Full-time employees indicates 77% of respondents have experienced Burnout at their current job and quite a few have cited more than one occurrence. The survey uncovered that there exists a lack of employers’ well-being programs that address stress in the workplace.2
The World Health Organization describes the phenomenon of Burnout as a syndrome conceptualized as resulting from chronic workplace stress that has not been successfully managed. It is characterized by exhaustion, cynicism related to one’s job, and reduced professional efficacy. Burnout is referred explicitly to as a phenomenon in the occupational context.3
It is a multi-dimensional and not a unidimensional phenomenon; the effects of burnout are multifold- ranging from employee absenteeism to poor mental/ physical health, leading to high attrition rates impacting a company’s performance.
To delve more into the issue, a study was conducted to assist employers in understanding their employees better through the ‘burnout’ lens. They studied levels of Burnout prevalent across job functions, locations, etc, and tried to identify the root cause and the positive & negative areas employees feel about their workplace.
The team provided synthetic datasets of structured/ quantitative data at an ordinal level and unstructured data comprising free-form text of employee feedback for the project. Reviews of about 97k employees from four different companies were collected. To identify the latent factors, factor analysis on Ratings data was performed. Company Culture, Work Satisfaction, Work-life balance, Skill development, and salary benefits were grouped under one factor, and Career Growth and Job Security were grouped under two individual factors. It was observed that when Job Security (which had a comparatively lower correlation with other ordinal variables) was dropped, three latent factors emerged. The researchers then applied subjective judgment and grouped them against three axes of Burnout:
- Work-life balance, Company Culture, Work Satisfaction and Exhaustion
- Career Growth, Salary Benefits and De-personalization
- Skill development, Job Security, and Lack of Achievement
The recommendations were curated based on actionable insights, helping the leaders focus on overall employee well-being and growth.
Techniques used: Edmundson Summarizer, Latent Semantic Analysis Summarizer, Rule-based classifier, SUMMA Package, PytextRank, and factor analyzer module of Python.
The POC is a step towards ensuring that leaders and HR management are aware of the state of the workforce and can take necessary action.
References:
1: Viswanathan M., Banik S., Pande S., Das A., Bhavaraju A., Koppula S.R., ‘Peak Health Techniques: Using a combination of multiple facets of Health, Engagement and Productivity to estimate occupational burnout,’ 2022
2: Workplace Burnout Survey | Deloitte US
3: Burn-out an “occupational phenomenon”: International Classification of Diseases (who.int)