“Ch 15 — People and Organizations: Data Modeling is a Full-Contact Sport” — @practicaldatamodeling

Why this is in the vault

Joe Reis frames data modeling as fundamentally sociological, not technical, and gives a 30-day situational-awareness checklist for entering a new data engagement. RDCO drops into client orgs cold; this is the operating manual for the first month. Joe’s own one-liner: “If I am going to pick one chapter to give to anyone working with data, this is it.”

The core argument

“The major problems of our work are not so much technological as sociological in nature.” — DeMarco/Lister, Peopleware

Reis argues most data modeling failures are political, not technical. He uses Target Canada’s $7B failure and a “Customer 360” project as case studies, then walks through the operational frameworks for surviving them.

Key frameworks:

Conway’s Law and the Silo Tax — your data architecture will mirror the org chart. The Inverse Conway Maneuver (re-org the team to ship the architecture you want) is one lever; tolerating silos with a clear “silo tax” is another.
Organizational archetypes — Traditional hierarchy (“data breadline”), hub-and-spoke (centralize nouns/decentralize verbs), decentralized/flat (agility vs. data anarchy), hybrid shadow (shadow IT + AI + ungoverned data).
Enterpriseland vs. Productland — the load-bearing dichotomy. Enterpriseland treats data as “business exhaust” and plays defense (efficiency, risk mitigation). Productland treats data AS the product and plays offense (growth, user impact). The Mixed Model Artist learns to operate in both and translate between them.
The 6 Core Questions as a requirements bridge — Who/Why/What/When/Where/How. Reis adds an “in the age of AI agents” overlay to each, plus an explicit “out of scope” declaration.
Power-Interest Grid for stakeholder mapping (high-power/high-interest = Manage Closely; low/low = Monitor; etc.). Tactics for handling resistance.
The CDO Dilemma — high accountability vs. “toothless” authority. What to do when your sponsor leaves.
Selling data modeling — different pitches for Productland (growth + user impact) vs Enterpriseland (efficiency + risk). A specific pitch for “data modeling for AI success.”
Anti-patterns — what NOT to do.
Authority vs. Liability — ownership without power is a trap.
Danger signs — red flags your initiative is about to derail.

The deliverable: a Week 1-4 checklist for “Your First 30 Days” — Map the Terrain (stakeholders, archetype, Enterpriseland vs Productland), Understand Communication Structures (Conway’s Law audit, who defines truth, what shadow IT exists), Assess the Environment (psychological safety, change management, who’s been burned), Plan Your Approach (launch partner, scope, first visible win).

Mapping against Ray Data Co

Strong. Three direct hits:

The 30-day checklist is RDCO’s missing onboarding SOP. When RDCO drops into a client engagement, the first month is currently ad-hoc. Reis’s Week 1-4 structure is a ready-made template — should be lifted into 02-sops/ as the “First 30 Days at a Client” SOP, attribution to Reis. The Power-Interest Grid + Enterpriseland/Productland diagnosis becomes the deliverable from week one.
Enterpriseland vs Productland clarifies RDCO’s positioning. RDCO sells into both. The pitch is different: for Enterpriseland clients the COO-as-Claude story is “your operations get cheaper and lower-risk”; for Productland it’s “your data product ships faster and your moat compounds.” We should have two distinct pitch decks, not one.
Authority vs. Liability is the COO-as-Claude trap. If RDCO is positioned as “we’re your COO” but the founder retains all decision authority, RDCO has liability without authority — exactly the trap Reis warns against. The engagement contract should explicitly grant decision-rights for the categories RDCO is being held accountable for, or RDCO should opt out of those categories.

Secondary: Conway’s Law applied to RDCO itself — we’re a 1.5-person team (founder + Claude). Our “shipped architecture” is whatever a one-person ops layer can hold in its head. That’s a real constraint on the products we can build.

Mapping note

This is a candidate concept-article extraction:

“Enterpriseland vs Productland” → vault concept doc, 00-vault-concepts/enterpriseland-vs-productland.md
“30-day situational-awareness checklist” → SOP, 02-sops/2026-04-20-first-30-days-at-a-client.md

Both deferred to a follow-up /compile-vault pass — not auto-creating without founder review.

2026-04-20-data-engineering-weekly-issue-266 — same week, also covers org-design-vs-data-architecture themes
../00-vault-concepts/agent-deployer-positioning — RDCO as agent deployer; Reis’s “data modeling for AI success” section is direct support
../01-projects/coo-engagement-template — onboarding template that should absorb the 30-day checklist

Copyright note