06-reference

practical data modeling ch14 levels of data modeling

2026-08-07·reference·source: Practical Data Modeling·by Joe Reis
data-modelingconceptual-modellogical-modelmixed-model-artsubiquitous-language

"Ch 14 — The Levels of Data Modeling" — @practicaldatamodeling

Why this is in the vault

Joe Reis lays out the Conceptual → Logical → Physical data-modeling progression as a "town plan → blueprint → build" sequence, arguing the value of the Conceptual stage is the semantic-agreement conversation it forces, not the diagram it produces — directly usable framing for RDCO's client-onboarding discipline.

Series-order anomaly

This chapter is labeled 14 but arrives Aug 7 2026, after Ch 15 (filed 2026-04-20) and Ch 16 (filed 2026-04-25) — both already in the vault. Reis's own copy explains why: this is "the final unreleased chapter of Mixed Model Arts, Book 1," an agglomeration of older standalone articles (examples updated) published last as the book compiles for a late-Aug/early-Sept 2026 release. So the numbering reflects the book's table of contents, not the newsletter's publish order. Worth noting rather than silently filing as if chapters ran 1→16 sequentially — they didn't. A near-identical standalone treatment of this exact topic already exists in the vault at 2026-01-31-practical-data-modeling-levels-outdated-redux.md (Jan 31), which argues the CDM/LDM/PDM hierarchy is a toolkit, not a mandatory ceremony — Ch 14 appears to be the more polished, book-bound version of that same argument.

Source fidelity note

The plaintext body is paywall-truncated — roughly the first 40% renders (intro + full Conceptual Data Model section + partial Logical Data Model section), cutting off mid-sentence in the LDM section. The HTML body confirms a hard paywall block ("Continue reading this post for free in the Substack app / Claim my free post / Or upgrade your subscription"). Per the founder's paid-sub-lapse (2026-04-24) — status not re-verified this run — the Physical Data Model section, the promised e-commerce worked example, and the conclusion are not visible here. Filed on the rendered 40% per the Not Boring precedent for partial paywall content rather than treated as unprocessed.

The core argument

Reis restates the classic three-level data-modeling progression using the MMA book's boxing/MMA framing (identity → game plan → execution):

Mapping against Ray Data Co

Medium. The direct hit is that RDCO's client-discovery phase is exactly the CDM conversation Reis describes — and we don't currently force it as a distinct, named step before touching a schema. Reis's claim that the CDM's value is semantic agreement, not the diagram, is a useful check against RDCO jumping straight to a logical/physical design during a fast-moving engagement (a real failure mode given RDCO's speed bias). This reinforces rather than adds to the stronger mapping already logged from Ch 15 (30-day onboarding checklist, Enterpriseland vs Productland) — Ch 14 supplies the specific vocabulary (CDM/LDM/PDM) that the Ch 15 SOP candidate should use when it names "map the terrain" activities. Downgraded from strong to medium because the paywall cuts off before the worked example and the PDM section, which is where a genuinely new tactical lift would likely live.

Related

Copyright note

Quotes ≤15 words, paraphrase otherwise. Source: Practical Data Modeling, Aug 7 2026 — view at https://practicaldatamodeling.substack.com/p/ch-14-the-levels-of-data-modeling (paywalled past the free preview).