This is a known fact that companies track your GitHub commits and lines changed / removed and stack rank it agains peers. I know it is just to find the outliers (at least how it should be used in my opinion), to see if someone started slacking and course-correct them earlier. Asking managers here: - what do companies measure exactly? total amount of code written? What about remove lines? Comments on PRs? Or maybe even jira tickets closed? Number of Docs written? - how often it gets updated? Is it averaged over weeks / months? People may go on a month vacation, this will impact their metrics. Others almost don’t take time off. No matter what, there will always be some people at the top and some at the bottom of the ranking. At which point do you jump in and try some interventions? Do you notice the performance dip as a manager before the metrics show it or metrics show it first? Have you been able to help a person come back from the bottom of the ranking? Too many question, but I am really curious about your experience with such tracking tools. TC: 280k
I know my management used it to find and PIP outliers. The guy they pipped had lots of commits but he eventually squashed and merged his PRs, making his commit count go low. The uneducated management used this metric to PIP him. He learned about this, then started creating hundreds of commits for the simplest of PRs. Serves the dumb management well.
Yep. Don’t code for efficiency, don’t code for efficacy. Code for volume.
Well, this is an example of how tools are used incorrectly. And it is bad. I am sure some managers do use it to PIP. But if you do look into lines of code and performance, I think there is still some correlation between them. Like most people that I know who got promoted were top contributors to the code base. But I had to admit their projects required a lot of work and lines of code (like a huge rewrite of the system or a completely new set of features with a lot of integrations)… modifying existing code doesn’t yield as many lines…
We had a guy who was asked to leave. He had low number of PRs / lines of code. But I think they cared about the actual results that he delivered rather than lines of code. But what do you do if you see a person is writing so little amount of code compared to their peers? How can you pay that person the same wage? Not defending the system here, firing people just based on just the commit number is bad. Performance should still be measured holistically.
This kind of evaluation tells me that management doesn't understand how to measure performance. Specifically, they don't know what is important for the business and how engineering work helps the business. So they are using BS metrics for evaluation. How about putting the management on PIP until they demonstrate they understand the business?
Jira can be tracked as well as company’s laptop
Tech Industry
Yesterday
3966
Crossed a line with my boss
Tech Industry
9h
2273
Asians - what are your thoughts on asian female white male ?
Tech Industry
6h
493
Will job market ever recover?
World Conflicts
Yesterday
575
American police seem to work only when Israel is challenged
Tech Industry
3d
40378
[Closed] I do tech screens at Google. AMA
Using lines written/commits/PRs for stack ranking is the dumbest thing I've ever heard.
So it is very consistent with American business culture. “Hey, lines of code per hour should be a good mark of productivity! By the way, what is ‘spaghetti code,’ and why do our developers keep complaining about it!” “Time spent typing on a computer” is a similar brainless metric used for other knowledge workers. The big irony is that most people who track and evaluate those bogus statistics probably don’t spend much time typing themselves.