Cobra Effect · People and organisations
The bus factor
How many people could leave before the work stops.
7 cards, read aloud in 2:26, with a test and sources.
1994. Somebody asks what happens if Guido is hit by a bus.
The Python programming language had one author, Guido van Rossum. On a public mailing list a colleague asked the obvious question. If he stepped in front of a bus tomorrow, what happens to all of this?
The question stuck, and it acquired a number.
The bus factor is how many people would have to vanish before the work stops. One means a single person carries something nobody else can do. It is a dark joke with a serious use. No actual bus is required. People resign, fall ill, or go on leave.
Most projects score far worse than anyone guesses.
A study of 133 popular open source projects put about two thirds of them at two people or fewer. Usually it is one particular person, and everybody already knows who. They are not hoarding anything. They simply never had a reason to write it down.
2015. The world finds out who keeps its email encryption running.
Reporters revealed that the widely used GnuPG encryption software was largely maintained by one man, Werner Koch, who was nearly out of money. A great deal of the world’s secure mail rested on him. Donations arrived once people knew.
Knowledge in one head is the cheapest thing to build and the dearest to lose.
Writing it down is slow, dull and nobody thanks you for it. So it stays where it is, growing more valuable and more fragile every month. Then somebody hands in their notice and you find out what they were holding.
Raising the number is boring work that always works.
Pair on anything only one person can do, so a second person has done it once. Write the steps down while you do the job, not afterwards. Rotate the awkward tasks, even on the days when the expert would be faster.
Go and find your own one, today.
Ask who would have to be telephoned on holiday if something broke tonight. Whichever name people say first is your answer. Now go and make a second person who could have said it too.
Sources
- Bus factor, Wikipedia. The definition, the 1994 question about Guido van Rossum and Python, and the studies that measured the number across popular open source projects.
- The Cathedral and the Bazaar, Eric S. Raymond, 1999. On how open projects survive their founders, including the argument that handing a project to a competent successor is a duty rather than a favour.
- GNU Privacy Guard, Wikipedia. The encryption software behind a great deal of secure email, and the 2015 reporting that revealed how few people were keeping it alive.
Nearby ideas
- Brooks’s law. Adding people to a late project makes it later.
- The curse of knowledge. Once you know something, you cannot see what a beginner is missing.
- Conway’s law. Systems end up shaped like the teams that build them.
- The iron law of oligarchy. Even democratic groups end up run by a permanent few.
- The Shirky principle. Institutions tend to preserve the problem they exist to solve.
- Sayre’s law. The smaller the stakes, the fiercer the argument.
- Amara’s law. We overrate new technology in the short run and underrate it in the long.
- Tuckman’s stages. Why new teams have to fight before they can work well.