Proof
A manifest, redacted.
This is the document that goes to a lab before any commercial conversation. The supplier here is anonymised: a working education technology company, seven fiscal years, roughly seventy contributors across the period.
Manifest, anonymised supplier, as of 12 September 2026FDS-2026-0121
| Source system | Volume | Years | Contributors | De-identification |
|---|---|---|---|---|
| Mail, tenant wide | 30 mailboxes | 2022 to 2026 | 27 | Addresses tokenised, bodies scanned |
| Chat, 46 channels | 7,900 messages | 2022 to 2026 | 27 | Handles tokenised, DMs excluded |
| Application database | 140 tables | 2021 to 2026 | 70 | Identifier columns suppressed at source |
| CRM | 5,539 accounts | 2022 to 2026 | 12 | Buyer identity dropped |
| Accounting | 5 fiscal years | 2021 to 2026 | 4 | Bank and routing fields suppressed |
| Document store | 3,769 native files | 2021 to 2026 | 70 | Scanned pages redacted by pixel destruction |
Exact byte sizes per source are counted at export and stated in the manifest sent under NDA. They are withheld here because they identify the supplier.
What each column means
- Source system. The system of record the data was exported from, named specifically. Functional equivalents are fine, but we say which one it actually was.
- Volume. Counted from the source system at export, in the unit that system uses. Mailboxes, messages, tables, accounts, files, fiscal years.
- Years. The continuous period covered. A gap in the middle is stated, never smoothed over.
- Contributors. Distinct people whose work appears in that system during the period, whether on payroll or contracted.
- De-identification. The technique applied to that source before anything leaves our systems, drawn from the published standard.
What the manifest also carries under NDA
- Exact size per source, in bytes.
- The coherence map: which systems overlap in period and in people, which is what makes the corpus more than a pile.
- The full exclusion schedule for that supplier.
- Format distribution by file type and by industry segment.
- The de-identification techniques applied per field class, and the residual-risk note.