Bus Factor
The number of people who would have to leave before a project stalls.
Bus factor is the number of people who would have to become unavailable — the metaphor is being hit by a bus — before a project stalls for lack of anyone who understands it. A project with a bus factor of one has exactly one person who can fix a given class of bug or explain why a piece of it works the way it does; losing that person loses not just effort but knowledge nobody else has.
The number is not headcount. A team of twelve can have a bus factor of one if eleven of them only touch code that the twelfth person wrote, reviewed, and alone understands. What it measures is the concentration of undocumented knowledge — reasoning that lives in one head rather than in a comment, a design note, or a commit message anyone can read. Written-down reasoning keeps a bus factor high even when only one person currently does the work, because anyone competent could pick it up from what is left behind.
That reframing points to what actually raises the number. Writing things down is necessary but not sufficient — Documentation Rot means a note goes stale the moment the code it describes changes underneath it — so the interventions that hold up are built into the workflow rather than treated as separate effort: Code Review that requires the reasoning behind a change to be legible to a second person, not just the diff; a history that preserves Provenance so a future maintainer can trace why something is the way it is; Code Comment and Naming chosen for a stranger's benefit rather than the author's; and rotating ownership deliberately rather than letting whoever wrote a subsystem first stay its permanent sole owner. The alternative is a codebase full of the unexplained constraint Chesterton's Fence warns against, compounding with Conway's Law: a team drawn too narrowly around a component concentrates its architecture and its knowledge in the same few people.
See also7
Documentation Rot
Documentation that has drifted from the system it describes, and is now worse than none.
Method27 connections
Code Review
A second person reading a change before it lands, and the practices that make it worth the time.
Version Control & Delivery25 connections
Provenance
The recorded origin of a change or a claim — who made it, when, and on what evidence.
Method40 connections
Conway's Law
Systems tend to reproduce the communication structure of the organization that built them.
Version Control & Delivery10 connections
Chesterton's Fence
Do not remove something whose purpose you cannot explain.
Method18 connections
Commons-Based Peer Production
Large-scale production coordinated by modular tasks and volunteer contribution rather than price or hierarchy.
Meaning & Society13 connections
Peter Principle
In a hierarchy, people get promoted based on performance in their current role until they reach one they're bad at, and stop there.
Method5 connections
Linked from6
- Blameless PostmortemMethod
Analyzing an incident on the assumption everyone involved made a reasonable decision given what they knew at the time.
- Brooks's LawMethod
Adding people to a late software project makes it later, because onboarding and coordination costs outpace the new capacity.
- Commons-Based Peer ProductionMeaning & Society
Large-scale production coordinated by modular tasks and volunteer contribution rather than price or hierarchy.
- Conway's LawVersion Control & Delivery
Systems tend to reproduce the communication structure of the organization that built them.
- Not Invented HereMethod
The bias against adopting an outside solution simply because it wasn't built by your own team.
- Peter PrincipleMethod
In a hierarchy, people get promoted based on performance in their current role until they reach one they're bad at, and stop there.