Core Insight
An aggregate number can appear small, stable, or acceptable while hiding large, offsetting forces underneath. The total is not the story — the components are.
Where This Appears
| Domain | Example |
|---|---|
| Variance analysis (FMGT Ch10) | Cyprus Company’s total manufacturing variance was 40,000 U labour efficiency offset by 30,000 F volume. Massive problems hidden by massive wins. |
| Economics (AD-AS) | GDP growth of 2% could mask a booming sector (+8%) and a collapsing sector (−6%). The aggregate hides structural change. |
| Personal finance | Net worth unchanged month-to-month could mean nothing happened — or could mean assets grew 5,000. |
| Data analysis | An average customer satisfaction score of 7/10 could mean everyone rates you 7, or half rate you 10 and half rate you 4. Same mean, completely different stories. |
The Diagnostic Question
When you see a total or an average: “What could be offsetting inside this number?”
If you can’t answer that question, you don’t understand the number well enough to act on it.
Connection to Ceteris Paribus
Decomposing totals into components IS ceteris paribus — you isolate each variable’s contribution to see its individual effect. Variance analysis splits total cost differences into price and quantity components. The split is the insight; the total is the starting point.
North: Where this comes from
- Simpson’s Paradox (statistics) — aggregate trends reverse when data is disaggregated
- Ceteris Paribus — isolating variables to see individual effects
East: What opposes this?
- Sometimes the total IS the right level of analysis — over-decomposing can create noise
- Parsimony — not every total needs to be split
South: Where this leads
- Variance analysis — the entire Ch10 system exists because cost totals hide stories
- Management by exception — investigate components, not totals
- Data visualization — dashboards that show only totals can mislead
West: What’s similar?
- Ecological fallacy — assuming group-level patterns apply to individuals
- Composition fallacy — assuming what’s true of parts is true of the whole