The six fields
| Field | What it holds | Why it matters |
|---|---|---|
| Context | The situation at the time of the call — constraints, pressures, what was known | Without it, every past decision looks arbitrary |
| Decision | The exact choice made, dated | The single source of truth for "what did we decide?" |
| Owner | Who made the call (role, not just name) | Accountability survives staff changes |
| Trade-offs | What was knowingly given up or risked | The most valuable part — it's what nobody writes down |
| Links | Bi-directional references to related decisions | Turns a list into a knowledge graph |
| Domain | Hiring · Sales · Product · Finance · Ops | Filtering, asset generation, and ownership boundaries |
A raw entry
The ledger is stored as structured Markdown plus a JSON index — readable and exportable without ADAMAS. This is the entry behind SAL-021 in the demo:
{
"id": "SAL-021",
"domain": "sales",
"date": "2025-04-14",
"title": "Decline the automotive OEM frame contract",
"context": "OEM offered a frame contract worth ~35% of annual revenue at rates 20% below card, with penalty clauses and 90-day payment terms.",
"decision": "Decline. Volume would crowd out three A-clients and concentrate revenue risk in one buyer with the weakest terms.",
"owner": { "role": "founder", "dissent": ["head-of-sales"] },
"tradeoffs": [
"Walks away from guaranteed volume in a soft quarter",
"Possible door closed at that OEM for years",
"Kept pricing power and client mix intact"
],
"links": ["FIN-016", "SAL-017", "OPS-020", "PRD-019"],
"sources": ["email:2025-04-02#thread-114", "meeting:2025-04-11#weekly-review"]
}
Note the sources array: every entry traces back to where the decision actually happened. The dissent field records disagreement — five years later, knowing who disagreed and why is gold.