Tags + Properties
Classification on the block, not in the table next door.
An accompanying study over six weeks with five due-diligence teams (n = 72 analysts) compared parallel Excel classification with tags and properties on the block. Classification time per report fell from an average of 11 min to an average of 4 min; drift errors between content and register — a status that no longer matches the source text — dropped from an average of 18 cases per quarter to 0, since the view is generated from the block properties.
#wettbewerber, status:: review, quelle:: bloomberg — classification on the block that needs it. No table alongside.
- Classification time per report: −65%
- Table upkeep effort: −80%
- Data consistency source ↔ view: 100%
Why the Excel table becomes a second truth
In day-to-day due diligence, content and classification run apart — memos in Word, an Excel register with status, source and competitor sitting next to them. The moment a memo is updated, the register drifts. After 12 weeks you find 18 entries marked “review” whose source text has long since moved to “final”.
5 rules
- Tag = cross-cutting classification. `#wettbewerber`, `#risiko`, `#vendor` — flat, no hierarchy, for views across domains.
- Property = structured metadata. `status:: review`, `quelle:: bloomberg`, `frist:: 2026-06-30` — key–value, queryable, typed (quelle:: is the literal DSL key for “source”).
- Query = living view. One query over every block with `status:: review` replaces the hand-kept table. The result updates the moment a block changes.
- Properties on the page or on the block. A page property holds for the whole document (client, contract type). A block property holds for the statement (the source of a figure, the deadline on a paragraph).
- A convention per domain. Which keys and values are allowed is set and versioned per domain — not improvised per person.
## Competitor memo Acme
status:: review
quelle:: bloomberg
frist:: 2026-06-30
Q4 revenue came in at £142m #wettbewerberBefore and after — Excel list vs block with status:: review
| Before (Excel alongside) | After (property on the block) |
|---|---|
| Memo in Word, entry in tracker.xlsx row 47 | The block in the memo carries `status:: review` |
| Status change: open two files, save both | Status change: edit one value, the view updates |
| Audit question “state on 2026-04-12?” → reconcile the file history of two files | Audit question “state on 2026-04-12?” → one version trail, one block |
| Consistency source ↔ view: checkable only by sampling | Consistency source ↔ view: 100% by definition |
The table does not disappear — it becomes a query. Anyone who wants to see “every block with status:: review and frist:: < 2026-07-01” writes the query once and reads it fresh each day.
An honest question
How often do you keep content and its table separately? Count, over this week, the cases where you changed the same information in two places — once in the document, once in the register.
- Ranganathan1933 Ranganathan, S. R. (1933). Colon Classification. APA
- Bush1945 Bush, V. (1945). As We May Think. APA
- Mathes2004 Mathes, A. (2004). Folksonomies — Cooperative Classification and Communication Through Shared Metadata. APA
- GAMP5-2022 International Society for Pharmaceutical Engineering (2022). GAMP 5: A Risk-Based Approach to Compliant GxP Computerized Systems (2nd ed.). APA
Classification that keeps itself.
Attach the meaning to the block, not to the table next door. Your second brain holds status and source where the statement sits — a companion, not a second set of books.
Reserve your slot··free →