Predicting Developer Attrition
The factors that influence attrition include opportunities, culture, leadership support, and burnout.
This is the latest issue of Engineering Enablement, a weekly newsletter covering the data behind world-class engineering organizations. To get articles like this in your inbox every Friday, subscribe:
This week I read Predicting Attrition among Software Professionals: Antecedents and Consequences of Burnout and Engagement. In this study, researchers investigated the factors that influence burnout and engagement among software developers, and whether these factors can predict attrition.
My summary of the paper
Prior studies have examined burnout and engagement, however few studies have investigated their consequences. Even fewer have done so within software engineering. The researchers here sought to fill that gap: they wanted to understand whether there’s a relationship between three factors—leadership support, organizational culture, and opportunities to learn—and burnout, engagement, or developers’ intention to stay.
This study was conducted through a large-scale survey with over 13,000 responses at a globally distributed organization. The researchers then conducted multiple different types of analyses, including Partial-Least Squares Structural Equation Modeling, Multi-Group Analysis, and Importance-Performance Map Analysis to answer their different research questions.
Key definitions:
In this paper, the authors define ‘burnout’ as an individual’s experiences of exhaustion on physical, emotional, and cognitive levels.
Engagement is defined as a positive, fulfilling, work-related state of mind that is characterized by vigor, dedication, and absorption. Rather than merely the opposite of burnout, or a momentary and specific state, engagement is a positive affective-motivational state of fulfillment in employees, characterized by individual perceptions of energy, effectiveness, and motivation at work.
Here are the key findings from the study.
1. More leadership support reduces burnout, but it doesn’t have as much of an impact on engagement as expected. Surprisingly, of three factors studied, leadership support was the only one where there wasn’t a statistically significant association with a higher level of engagement. However, higher levels of leadership support was found to reduce burnout. It was also associated with having more opportunities to learn.
The researchers also explored whether leadership support impacted engagement differently for different groups. They looked at gender, tenure, and age. They found two groups where leadership support was more impactful to engagement: the group with 1-3 years of experience working at the organization, and those within the age bracket of 25-34.
2. Culture influences burnout and engagement. Generative organizations are defined as being performance-oriented, with good information flow, high levels of cooperation and trust, and bridging between teams. The results from this study found that generative organizational cultures are strongly associated with higher engagement, learning, and reduced burnout.
3. Engagement and burnout are both predictive of attrition. The researchers compared survey responses to whether the respondents stayed or left voluntarily within 90 days of taking the survey. From this, they found that engagement is positively associated with having the intention to stay, while burnout is negatively associated with having the intention to stay.
After segmenting the results, the researchers found a significantly stronger effect for those with 1-3 years tenure than those with over 5 years tenure. Meaning, engagement and burnout are stronger predictors of attrition for newer employees than more tenured ones.
Another interesting insight: developers can experience burnout despite the best efforts of an organization. “People do not quit companies, they quit teams.” Whatever initiatives an organization might create at the executive level to support their employees, it may be challenging for such initiatives to trickle down throughout the organization at the operational level.
4. Burnout prevention is a key strategy for reducing attrition. The researchers used an Importance-Performance Map Analysis (IPMA) where results were mapped to a quadrant. Those using the quadrant can think of Quadrant 1 (top right) as needing little or no intervention, Quadrant 2 as being “possible overkill,” Quadrant 3 being “low priority,” and Quadrant 4 is “concentrate here.” Of the factors studied, leadership support and opportunities to learn fell within Quadrant 2, “possible overkill”. Burnout prevention fell within Quadrant 4. This suggests that decision makers should direct their focus to burnout prevention.
5. Opportunities to learn is a top predictor of attrition. Data was captured for each respondent about their employment status 90 days after the initial survey. The researchers used machine learning to train and evaluate which variables in the study predict attrition. They found that opportunities to learn was the most predictive of attrition, followed by organizational culture, leadership support, and then burnout.
Final thoughts
This paper offers useful insights for engineering leaders about what influences attrition. What stood out most to me were the findings that engineering managers and teams play a huge role ("people don't quit companies, they quit teams") and that leadership support didn’t have as much of an impact on engagement as expected.
Who’s hiring right now
Here is a roundup of Developer Productivity job openings. Find more open roles here.
SiriusXM is hiring a Staff Software Engineer - Platform Observability | US
Vidyard is hiring a Senior Software Developer - Platform | Canada
Auditboard is hiring a Sr Engineering Manager - Infrastructure | Remote
Roku is hiring a Senior Manager - Developer Tools | Cambridge
Thanks for subscribing to Engineering Enablement. This post is free so feel free to share it.
-Abi