Base120 · Decomposition

Orthogonalization

DE · Decomposition

Ensure factors vary independently without correlation or interdependence

Model code
DE17
Transformation
DE (Decomposition) — Break problems into factors, layers, and independent subproblems.
Priority
17 of 20
Framework
HUMMBL Base120 — 120 mental models across 6 transformations

When to use

"Use when factors are confounded (correlated) and you cannot tell which is causing the effect. Redesign the system or experiment so factors vary independently, enabling clean measurement of each factor's impact."

Example

"A team ships a new design and a new backend simultaneously. Conversion improves, but they cannot tell which change caused it. Orthogonalization — ship the new design with the old backend, then the old design with the new backend. Now each factor's impact is measurable."

Common misuse

"Assuming factors are independent when they are not. Even after orthogonalization, hidden correlations may remain. Test for residual correlation after the redesign."

Also known as

Independence DesignDecorrelation

Related models

DE14DE2DE9