Use this when

The same fact lives in several places that drift apart, and people or systems act on different versions.

The pattern

        ┌─────────────────────┐
        │  SOURCE OF TRUTH    │  one place, owned, maintained
        └─────────┬───────────┘
                  │ read / sync
      ┌───────────┼───────────┐
      ↓           ↓           ↓
  copy (view)  report     workflow reads here

One place holds the authoritative value for a fact. Everything else either reads from it live, or is kept in sync with it deliberately. Nothing edits a copy directly.

Why it works

Most data inconsistency comes from having more than one place that can be edited. When only one place is authoritative and everything else follows it, the disagreement cannot happen. People and AI read the same current value, so they reach the same answer.

How to apply it

  1. For each important fact, choose one source of truth, usually the maintained system of record.
  2. Turn other copies into read-only views, or sync them from the source automatically.
  3. Move all edits back to the source.
  4. Give the source an owner who keeps it correct.

Good fit

  • Facts used across several teams or systems: prices, addresses, statuses.
  • Any workflow where acting on a stale value would cause real harm.

Poor fit

  • Genuinely local, short-lived working data that never needs to agree with anything else.

What this does not solve

A single source removes disagreement, not inaccuracy. The one source still needs an owner and a way to correct errors. It also does not remove the need to keep copies genuinely in sync; a broken sync recreates the problem silently.