Code Churn

In one sentence

Code churn measures how often lines or files are rewritten over a period—an activity signal, not a direct quality score.

Why it matters

Teams use churn to find hotspots, rework, and unstable areas. Misread as a productivity or quality grade, it punishes legitimate iteration and rewards untouched dead code.

How it works

• Common definitions count lines added and deleted in a window, sometimes excluding generated files or renames.
• File-level churn highlights paths that keep changing.
• Author or period comparisons describe activity concentration, not individual worth.
• Interpret churn with context: new feature spikes, incident fixes, and refactors all raise churn for different reasons.

Example

Rough file churn from Git (illustrative):

git log --since="30 days ago" --numstat --pretty=format: -- \
  | awk '{ add+=$1; del+=$2 } END { print add+del }'

Treat the number as “how much text moved,” then inspect which files and why before drawing conclusions.