The ontology behind every tie-out, the evidence it's built on, and the direct EDGAR connection underneath it all.
A checklist tells you where something sits. An ontology tells you what it is, why it matters, how it relates to everything else, and what evidence supports the call.
The unit we classify isn't the business event. It's the atomic transformation.
A single acquisition can trigger a perimeter change, a purchase-accounting adjustment, a run-rate adjustment, a financing adjustment and a share-count adjustment — five separate records, linked by one event ID, not forced into one bucket.
Classified by what's being transformed — not by whether it's cash, recurring, or "unusual."
What actually happened to the business, on a consistent accounting and economic basis — before any metric gets built on top of it.
Adding back D&A to build EBITDA is a formula, not proof D&A is abnormal. Kept structurally separate from the judgment classes above.
Enterprise-to-equity mechanics are downstream of normalized financials — kept out of the core normalization taxonomy on purpose.
The rules that keep twenty classes from turning into twenty inconsistent judgment calls.
A strong system returns a treatment and a rationale — not just a number.
Six levels, in order of authority, for every house conclusion.
The SEC's non-GAAP guidance is a particularly important guardrail here: a recurring operating cost doesn't become an acceptable add-back just because management labels it unusual.
Every number starts at the primary source — not a vendor's interpretation of it.
Partial view — the full source stack runs deeper than what's below.
The ontology above isn't from one textbook. It's a synthesis across disciplines that each solve part of the problem:
We'd rather tell you where this stands than round up.
v0.1 of this ontology has been stress-tested against a representative cross-industry case set — banks, insurers, REITs, pharma, software, E&P, utilities, airlines, retail, M&A. That's a strong signal. It is not yet proof of exhaustiveness, and we're not going to call it that.