Document Type: Canon
Status: Canon
Version: v1.1
Authority: MWMS HeadOffice
Applies To: All MWMS Brains running experiments
Parent: HeadOffice
Last Reviewed: 2026-03-15
Purpose
The Experiment Registry System exists to create permanent institutional memory for all MWMS experiments.
It ensures:
• structured hypothesis tracking
• statistical discipline documentation
• capital impact visibility
• learning compounding
• audit transparency
• drift prevention
No experiment may scale without registry entry.
Scope
This canon applies to:
• all MWMS Brains running experiments
• pre-test and post-test experiment logging
• statistical planning visibility
• validation and classification tracking
• business decision traceability after experimentation
• cross-system learning preservation
This document governs the minimum registry requirements for recognising, validating, and learning from experiments inside MWMS.
It does not replace:
• Experimentation Brain statistical methodology
• SIT enforcement logic
• Finance Brain capital authority
• Brain-specific test design protocols
• experiment execution systems themselves
Those remain governed by their own authoritative documents.
Definition / Rules
Registry Requirement Rule
Every experiment must be logged:
• before launch (Pre-Test Entry)
• after completion (Post-Test Entry)
If not logged:
Classification:
NON-COMPLIANT — EXPERIMENT NOT RECOGNISED
SIT may block scaling.
Experiment Record Structure (Mandatory Fields)
Each experiment entry must contain the following sections.
A. Experiment Identification
• Experiment ID
• Brain Owner
• Responsible Operator
• Date Created
• Traffic Source
• Funnel/Page Tested
B. Hypothesis Layer
• Problem Statement
• Root Cause Assumption
• Hypothesis (If X, then Y because Z)
• Expected Direction of Change
• Primary KPI
• Secondary KPIs
C. Statistical Planning Layer (Pre-Test)
• Baseline Conversion Rate
• Target MDE (%)
• Statistical Power Target (≥ 80%)
• Confidence Target (≥ 95%)
• Required Sample Size
• Traffic Allocation
• Minimum Duration
• Stop Rules Defined (Yes/No)
D. Execution Layer
• Start Date
• End Date
• Total Visitors per Variant
• Total Conversions per Variant
• Observed Lift (%)
• P-Value
• Power Achieved
• Business Cycles Covered
E. Validation Layer (SIT Confirmation)
• Sample Size Met (Yes/No)
• Power Requirement Met (Yes/No)
• Confidence Requirement Met (Yes/No)
• Duration Requirement Met (Yes/No)
• Primary KPI Validated (Yes/No)
• Contamination Detected (Yes/No)
Final Classification:
• VALID WIN
• VALID LOSS
• INCONCLUSIVE
• INVALID — STATISTICAL BREACH
F. Business Decision Layer
• Decision Taken
• Capital Allocated (if any)
• Risk Classification (Level 1–4)
• Override Used (Yes/No)
• Override Justification (if applicable)
G. Learning & Compounding Layer
• Why did this likely happen?
• What structural insight was gained?
• Does this affect other funnels?
• Should this be replicated?
• What is the next experiment?
Registry Integrity Rule
Registry entries are read-only after completion.
Edits require:
• logged amendment
• reason for update
• version notation
Recognition Rule
An experiment is not institutionally recognised inside MWMS unless it exists in the registry with the required fields.
Running a test without registry discipline does not create valid system learning.
Strategic Purpose
The Registry transforms MWMS from:
Random experimentation
into
Controlled compounding intelligence.
It prevents repeated mistakes.
It prevents forgetting losing tests.
It prevents ego-driven scaling.
It builds statistical maturity over time.
Cross-Brain Role
The Experiment Registry System exists above any single Brain.
It provides shared experiment memory and structural consistency across MWMS so that experiments can be reviewed, validated, compared, and learned from without fragmentation.
Final Rule
No experiment may scale, compound into policy, or be treated as valid institutional learning unless it is properly registered.
Registry discipline is mandatory.
Drift Protection
The system must prevent:
• experiments running without pre-test registration
• post-test outcomes being remembered informally instead of structurally
• scaling decisions being made from unregistered tests
• losing tests being forgotten because they were never logged
• edits to completed records occurring without amendment traceability
• separate Brains using incompatible experiment memory standards
Experiment memory must remain structured, auditable, and cumulative.
Architectural Intent
The Experiment Registry System exists to make experimentation inside MWMS permanent, reviewable, and compounding.
Its role is to stop experiments from becoming isolated events and instead turn them into governed institutional memory that improves statistical maturity, reduces repeated mistakes, and supports safer decision-making across all Brains.
Change Log
Version: v1.1
Date: 2026-03-15
Author: MWMS HeadOffice
Change: Rebuilt page to align with the locked MWMS document standard for this cleanup pass. Preserved the original purpose, registry requirement rule, mandatory experiment record structure, registry integrity rule, and strategic purpose. Added Document Type, Parent, Scope, Definition / Rules structure, Recognition Rule, Cross-Brain Role, Final Rule, Drift Protection, and Architectural Intent sections.
Version: v1.0
Date: 2026-02-20
Author: MWMS HeadOffice
Change: Initial creation of Experiment Registry System defining the permanent institutional memory requirements for MWMS experiments.
END – EXPERIMENT REGISTRY SYSTEM v1.1