Stanford research: 9.5% of software engineers are "pretending to be busy"! Characteristics of 3 ghost engineers exposed, IBM, Microsoft, and Google all fell

Yegor Denisov-Blanch, a productivity researcher at Stanford University, recently published the "Software Engineering Productivity Research" report, analyzing the performance data of more than 50,000 software engineers from multiple companies. Measure productivity with data from a private code repository. Data points out that about 9.5% of software engineers make almost no actual contributions but still receive full salary. The productivity of these "ghost engineers" is only one-tenth that of other engineers.

Stanford scholars estimate that 6.5% of the world's software engineers are "ghost engineers", causing a loss of 90 billion US dollars.

"Ghost engineers" usually have three characteristics, such as doing almost no work, working less than 1/10 of the median, and possibly wearing multiple hats. "Ghost engineers" may "pretend to be working" by submitting code with minimal or only minor modifications.

Research shows that 58% of engineers submit code less than three times a month, and the other 42% make some "trivial" changes, such as editing a line or a character to pretend to be working. This phenomenon has caused a huge waste of resources and also revealed loopholes in the management and performance evaluation system.

Even without extrapolating the 9.5% data pointed out by the study, and using a conservative rough estimate of 6.5%, Blanch pointed out that the total cost of the global software industry due to ghost engineers may reach as high as 90 billion US dollars.

Among them, "large technology companies" are the most affected. For example, IBM has 180,000 software engineers, including 17,100 ghost engineers, causing annual costs of up to 2.565 billion US dollars; Microsoft has 110,000 software engineers software engineers, including about 10,000 ghost engineers, with annual costs reaching US$ 1.568 billion; Oracle and Google each have 100,000 software engineers, with approximately 9,500 ghost engineers, with annual costs reaching US$ 1.425 billion.

Management is difficult, communication becomes a challenge, and remote work is more likely to give rise to "ghost engineers"

Most of these "ghost engineers" are concentrated in remote working situations. 14% of remote engineers are classified as ghost engineers. In comparison, the proportion of ghost engineers under the hybrid office system is 9%. The proportion of ghost engineers working is only 6%. However, in the remote working model, it is more likely to appear "efficient engineers" whose productivity level is five times the average.

Why are ghost engineers more common in remote working? Research speculates that in a remote working environment, it is more difficult for managers to directly observe employees' work conditions, and it is also more difficult to track work progress and results in real time. This may give some engineers an opportunity to "pretend to be working" even if they have no actual contribution.

In addition to the general lack of oversight and performance tracking mechanisms, remote communication and collaboration are also a challenge. Working remotely can make engineers more likely to feel isolated and unmotivated. If engineers are not actively participating in teamwork, they may be more likely to slack off and end up becoming "ghost engineers."

A former Apple engineer said frankly: "I submitted a lot of low-quality code and fixed it, and then I was rewarded for my ability to solve the problem. It made me feel that the incentive system was fundamentally flawed."

The existence of ghost engineers may not only cause a waste of corporate resources, but also show the challenges of remote working and management in the technology industry. Enterprises need to pay more attention to performance transparency and data-based management in order to improve talent efficiency and avoid the increasingly common situation of "pretending to be at work".

Source: fudzilla, Software Engineering Productivity Research, Yegor Denisov-Blanch X