These days many project leads talk about the bus factor, the truck factor, or the brick factor, which is the number of developers, engineers on a project that can get hit by a bus/truck/brick, causing the project to stop.

To put it even more simple it is a measure of the concentration of information about a system, a code a project. The most critical bus factor is 1.

The term, of course, is not new; it appeared in business management back in 1998.

“Bus” can be anything, such as a vacation, a layoff, a maternity leave, a meditation or a retreat on Bali, or even going on a binge, it’s figurative. But it is much harder to understand where these places in the system, where the bus factor lies, to reduce it.

How to calculate the bus factor?

Let’s imagine you have 30 people on a project that does web development, it has an API, a frontend and design, no details. 5 people develop the API, 10 people develop the frontend and another 5 do the design, your bus factor is 5.

Potential signals of having a bus factor

  • You have developers whose code you don’t review. Why? He’s a cool dude, it doesn’t make sense to review him, he writes his own style. Then no one will ever understand this code base.
  • A special case of the previous one, is a “code ownership” practice, that is in general not bad, but can lead to a situation when only one preson knows a particular part of code well. Even if you stick to the “code ownership”, allow other collegues to ask questions and to challendge each other.
  • You have no uniform naming of classes, variables, module arrangement structure in the project, no requirements for documented code and you don’t check it, either manually or automatically.
  • The CI/CD on the project was set up by one person, everyone asks him to add, change, but no one goes into details.
  • You have chunks of code commented out a long time ago with no understanding of why this is the case or in what situations you can’t in any way uncomment it.
  • The documentation and knowledge base is stored in private spaces, or worse, on local machines that can get formatted.

Can it be detected automatically?

Yes, but it will only be diagnostic, the team leader will have to dig deeper manually. For example, there’s a tool like this for Git/HG repositories that lets you find files where one person is the main line-of-code contributor.