The DecisionOps Guide
The Growth Engine
Price Signals
AcademyDecisionOps Guide2. Mental Model
Chapter 2
The Learning Nodes Mental Model
Decisions as first-class objects, the Q-I-E Triad, and why assumptions must be explicit and versioned.
⏱ ~12 min read
2.1 Decisions as First-Class Objects
In software engineering, a "First-Class Object" is an entity that can be created, stored, passed as an argument, and returned by a function. In Learning Nodes, we apply this same concept to Decisions.
From Vapor to Logic
Usually, a "decision" is something that happens in a meeting room. It is vapor. It leaves no trace except perhaps a line item in a spreadsheet or a Slack message.
In Learning Nodes, a Decision is a structured record in the database—with an ID, a state, and a history.
🗂️ Decision Object Properties
By making the Decision a concrete object, we can track it, audit it, and improve it over time.
2.2 The Core Triad: Questions, Interpretations, Evidence
The Learning Nodes Ontology is built on three pillars. These form the "Atomic Unit" of DecisionOps.
❓
1. The Question
The starting point. Phrased in business terms, not data terms.
✅ "Should we lower prices in Brazil?"
❌ "SELECT * FROM sales WHERE country = 'BR'"
✅ "Should we lower prices in Brazil?"
❌ "SELECT * FROM sales WHERE country = 'BR'"
🔍
2. The Interpretation
The translation layer. This is where the Analyst adds value.
"To answer 'Should we lower prices', we interpret 'Success' as an increase in total revenue."
"To answer 'Should we lower prices', we interpret 'Success' as an increase in total revenue."
📊
3. The Evidence
Data assets used to validate the Interpretation.
Evidence is not "Truth". It is support for a hypothesis with a Strength score.
Evidence is not "Truth". It is support for a hypothesis with a Strength score.
2.3 Assumptions as Explicit, Versioned Entities
Every model, every query, and every dashboard has assumptions buried inside it.
💭 Hidden Assumptions
"We assume the currency exchange rate is stable.""We assume the user drop-off is due to the UI bug."
Usually, these stay in the analyst's head.
The Assassin of Logic
💣 Hidden Assumptions as Landmines
If you assumed "Inflation will stay under 3%" and it hits 8%, your decision is invalid.But if that assumption wasn't recorded, you can't verify why the decision failed. You just see "Bad Result".
Assumptions in Learning Nodes
In Learning Nodes, an Assumption is a required field. You cannot commit a decision without listing the assumptions that support it.
| Explicit | "We assume competitor X will not lower prices." |
| Versioned | If this changes, we branch the decision logic. |
| Challengeable | Other users can flag an assumption as "Weak" or "Invalid", triggering a review. |
2.4 Artifacts vs Decisions vs Data Assets
It is crucial to distinguish between three types of objects in the Learning Nodes ecosystem.
⛽
Data Assets (The Fuel)
Raw material in the Marketplace. "Potential Energy".
Example: "US Census Data 2024", "Web Traffic Log"
Example: "US Census Data 2024", "Web Traffic Log"
⚙️
The Decision (The Engine)
The logic engine combining Q + I + E into a conclusion.
Example: "Expand to Texas based on Census Data + Growth Assumption"
Example: "Expand to Texas based on Census Data + Growth Assumption"
📄
The Artifact (The Exhaust)
Output to communicate the decision.
Example: PowerBI Dashboard, PDF Slide Deck, Excel Export
Example: PowerBI Dashboard, PDF Slide Deck, Excel Export
✅ Strategic Shift
Do not confuse the Artifact with the Decision. You can throw away the PDF, but you must keep the Decision record.2.5 Organizational Learning vs Institutional Amnesia
Why do organizations get smarter? Because people learn.
Why do organizations get dumber? Because people leave.
Why do organizations get dumber? Because people leave.
Institutional Amnesia
🧠 When Knowledge Walks Out the Door
When a senior analyst leaves, they take their context with them.The dashboard remains, but the knowledge of why it was built that way—why "Japan" was excluded, why "churn" was defined that way—vanishes.
The Learning Node
This is the origin of our name: Learning Nodes. Each Decision object is a "Node" of learning. It connects to previous decisions.
When a new analyst joins, they don't just see the code. They see the lineage of Questions, Interpretations, and Assumptions. They can see that "we tried this strategy in 2022 and it failed because Assumption X was wrong."
🌐 The Result
The organization learns, independent of the individuals within it.
