When to call
Use this review before replacing a vendor, combining feeds, migrating a normalized store or relying on a new source for a production decision. It is useful when two outputs look comparable but their coverage, session logic or instrument definitions may differ.
Who and what to bring
Typical sponsors are data, research, platform or product leads. The starting materials are sample extracts or approved access, data dictionaries, current mapping rules, expected use cases and the decision that depends on the comparison.
What gets tested
- Population coverage, sessions and source availability
- Symbols, contracts, strikes, option side and expiration
- Timestamps, trade date, duplicate, cancel and correction behavior
- Multi-leg representation and an unmatched taxonomy
- Field-level parity and downstream analytical impact
How the work runs
The review agrees a comparison schema, preserves every match state, tests representative edge cases and returns a written record. The deliverables are a scope, field map, evidence log, defect register, unmatched taxonomy and regression-control recommendation.
What is excluded
This is not a vendor certification, trading recommendation or blanket data-quality warranty. It does not publish licensed source records or reduce a migration to a single headline rate.