In a groundbreaking and quite possibly perplexing study coming out of Stanford University, software engineering productivity specialist Yegor Denisov-Blanch suggests that a surprisingly large segment of developer teams consists of so-called ghost engineers who contribute virtually nothing to the productivity of their teams. These individuals are ranked for performance at less than one-tenth the productivity of a median software engineer. Denisov-Blanch offers a definition of ghost engineering as being representative of employees who may juggle multiple jobs, leading to their minimal contributions in code development.
The essence of Denisov-Blanch’s argument is that around 9.5% of software engineers do virtually no work, regardless of whether they work remotely or in traditional office settings. His claims echo earlier criticisms by high-profile figures such as Elon Musk, who has similarly called out what he perceives to be shirkers or deadweight employees. Denisov-Blanch bases his statements on an analysis of technology companies that have granted his team access to their internal code repositories. Utilizing an algorithm, they assessed developers’ code activity to measure their productivity. Unlike traditional methods that count the number of commits, Denisov-Blanch’s algorithm evaluates the substance of each commit to determine its impact on the overall codebase.
Ghost Engineers Prevalence and Performance Analysis
Denisov-Blanch asserts that approximately 10% of software developers do little to no productive work, a significant claim supported by the analysis of over 50,000 engineers from hundreds of companies. This finding is particularly alarming as it suggests a substantial portion of the workforce is not contributing effectively to their teams.
The study also reveals interesting insights into the performance of remote versus in-office employees. While remote workers have more high-performing outliers, in-office employees exhibit a higher average performance overall. This dichotomy highlights the complexity of measuring productivity and the various factors that can influence an engineer’s output. Remote work offers flexibility and the potential for exceptional individual performance, but traditional office settings might provide better average results due to factors like immediate team interaction and supervision.
Economic Principles and Non-Code Contributions
Denisov-Blanch’s findings align with established economic principles such as the Pareto principle and Price’s Law. The Pareto principle, also known as the 80/20 rule, suggests that 80% of outcomes are driven by 20% of causes. Similarly, Price’s Law indicates that the square root of the total number of contributors accounts for 50% of the results, meaning that in a team of 100 developers, only 10 would account for half of the productivity. This highlights a significant imbalance in contributions among team members, which could remain unnoticed without precise productivity assessments.
Despite the compelling analysis, the study acknowledges potential biases. It was conducted on code repositories from companies open to such scrutiny, which could imply concerns about productivity gaps. Additionally, the algorithm might have overlooked non-code contributions such as localization materials and documentation. However, Denisov-Blanch assures that non-code contributions have been balanced in the algorithm’s weighting. Confirmations from participating organizations reveal a consensus that identified ghost engineers indeed are minimally contributing to code development, uninvolved in other activities such as mentoring or sales.
Origins of Ghost Engineers
Denisov-Blanch dismisses the idea that automation and AI directly contribute to the phenomenon of ghost engineers. Instead, he believes that high performers will be those who leverage AI tools effectively. The study identifies disengagement as a critical factor leading to minimal productivity. Frustration or loss of motivation in the work environment primarily contributes to this disengagement, causing some engineers to reduce their efforts significantly.
Over time, these engineers may test how little effort they can exert without facing consequences, ultimately resulting in ghost-like productivity. This phenomenon raises concerns about the work culture within software development teams and the necessity of addressing the root causes of disengagement. Effective management and supportive work environments are crucial to tackling these issues and promoting higher productivity across teams.
Management and Organizational Challenges
According to Denisov-Blanch, project managers and senior leaders are often removed from day-to-day operations, relying on flawed metrics or the judgment of middle management. This distance and organizational politics might discourage reducing team sizes, leading to inefficiencies. Managers, facing conflicting incentives, may downplay issues to protect their reputations, contributing further to the prevalence of ghost engineers.
Denisov-Blanch urges for transparency and data-driven decision-making to help managers build high-performing teams and address the systemic issues that lead to disengagement among developers. By equipping managers with the right tools and promoting meaningful strategies by senior leaders, a more efficient and supportive working environment can be achieved. Reducing the gap between upper management and day-to-day operations is key to ensuring more accurate performance assessments and fostering a culture of accountability.
Industry Perspectives on Measuring Developer Productivity
The study’s insights have sparked discussions across the industry. Jon Collins, VP of research at GigaOm, recognizes the complexity of measuring developer productivity. Traditional metrics like lines of code are poor indicators since much development work doesn’t involve coding. Collins emphasizes the importance of delivering focused and simplified solutions collaboratively worked out with business users.
Sergey Katsev, VP of engineering at Catchpoint, concurs with the findings due to the intricate relationship between software engineering and motivation. He highlights that larger teams often allow some members to “fall through the cracks” concerning performance. Katsev contends that ensuring developer engagement through periodic team conversations and mutual accountability is crucial. Addressing these issues can help create a more cohesive and effective working environment, reducing the prevalence of ghost engineers.
Challenges with Existing Productivity Metrics
The study sheds light on the problem of current metrics misrepresenting reality. Metrics like lines of code and commits fall short in accurately reflecting productivity because they fail to consider payload size. Story points are too subjective and vary among teams, and self-assessment surveys, while useful for gauging developer experience, don’t measure productivity effectively. DORA metrics, which assess DevOps performance, are not suitable for productivity assessment due to variations in deployment sizes.
This dependence on flawed metrics forces engineering leaders to choose between making potentially incorrect data-driven decisions or relying on intuition, both of which carry the risk of poor decision quality. This issue highlights the urgent need for better tools to measure productivity that can provide an accurate analysis of contributions without misrepresenting the actual work done.
Denisov-Blanch and his team concluded that increased transparency and data-driven decision-making could enhance developer experience and team dynamics. Their study underscored the need for improved productivity measurement tools that accurately analyze contributions. Identifying ‘ghost engineers’ could help companies address disengagement and foster a healthier work culture.
Overall, the discussion around ghost engineers emphasized the importance of transparency, better metrics, and understanding team dynamics to boost productivity and engagement among software developers. Enhanced metrics, ongoing engagement, and effective management strategies could help tackle the issue of ghost engineers, leading to a more motivated and efficient workforce.
