Document Type: Specification
Status: Active
Version: v1.2
Authority: MWMS HeadOffice
Applies To: Ads Brain campaign decision logic and optimisation control
Parent: Ads Brain Canon
Last Reviewed: 2026-03-17
Purpose
The Ads Brain – Decision Engine defines the structured decision logic used to determine what action should be taken after advertising creatives and campaigns are tested.
Its purpose is to prevent reactive or emotional decision-making.
The Decision Engine converts campaign signals into controlled actions.
Ads Brain must not decide based on opinion.
Ads Brain must decide based on structured signal interpretation.
Scope
This specification applies to:
• post-test campaign decision logic
• creative and campaign action selection
• structured conversion of signal into action recommendation
• decision control after review and scoring
• governance-safe optimisation decisions inside Ads Brain
This document governs how Ads Brain should determine what action follows campaign analysis.
It does not govern:
• offer viability approval
• capital allocation approval
• final survivability authority
• Finance Brain override
• Affiliate Brain structural approval
• HeadOffice strategic override authority
Those remain governed by Affiliate Brain, Finance Brain, HeadOffice, and related MWMS systems.
Definition / Rules
Core Decision Outputs
The Ads Brain – Decision Engine may produce only the following operational outputs:
• Launch
• Continue
• Iterate
• Pause
• Scale
• Retire
• Escalate
No other decision output is valid unless approved by HeadOffice.
Decision Flow Position
The Decision Engine operates after:
Creative production
↓
Experiment launch
↓
Campaign review
↓
Creative performance scoring
↓
Decision output
The Decision Engine sits between campaign analysis and action recommendation.
Decision Inputs
The Decision Engine evaluates the following signal categories.
Creative Signals
• hook performance
• viewer retention
• engagement behaviour
• creative fatigue indicators
• message clarity
Campaign Signals
• traffic delivery
• click-through rate
• conversion rate
• cost per acquisition
• revenue per click
Audience Signals
• audience quality
• targeting response
• frequency exposure
• saturation indicators
Platform Signals
• learning-phase stability
• algorithm behaviour
• delivery consistency
• optimisation stability
Governance Signals
• experiment validity
• statistical reliability
• capital exposure constraints
• structural compliance
Decision Rules
All decisions must follow structured logic.
Launch
A Launch recommendation is valid only when:
• the opportunity has been approved for test
• the creative brief is complete
• the experiment design is defined
• the platform context is clear
• governance constraints are satisfied
Launch may not occur if any core input is incomplete.
Continue
A Continue decision is valid when:
• early campaign signals are neutral to promising
• insufficient data exists for conclusion
• delivery is stable enough to gather more evidence
• no major failure condition is present
Continue means:
Keep the campaign live and keep collecting signal.
Iterate
An Iterate decision is valid when:
• the campaign shows partial promise
• one or more weak variables are identifiable
• creative or targeting adjustments are justified
• the experiment remains worth pursuing
Common reasons for Iterate include:
• weak hook performance
• weak message clarity
• moderate CTR with weak conversion
• audience mismatch
• early fatigue indicators
Iterate means:
Modify a defined variable while preserving test discipline.
Pause
A Pause decision is valid when:
• campaign performance is weak
• further spend is not justified immediately
• more information or redesign is required
• experiment integrity has been compromised
Pause does not mean permanent rejection.
Pause means:
Stop spend while preserving the option for later re-entry.
Scale
A Scale decision is valid only when:
• the campaign demonstrates stable conversion behaviour
• the creative remains durable
• the platform shows optimisation stability
• the campaign has sufficient data volume
• Finance constraints are respected
• no major structural risk signal is active
Scale means:
The campaign may be considered for controlled expansion under Scaling Intelligence and Finance review.
Ads Brain may recommend scaling.
Ads Brain may not independently authorise capital expansion.
Retire
A Retire decision is valid when:
• campaign performance is consistently weak
• the tested structure has low recovery probability
• repeated iterations fail to improve signal
• structural viability appears too low to justify more testing
Retire means:
The campaign or creative path should be closed and recorded as a failed or non-viable testing branch.
Escalate
An Escalate decision is valid when:
• Finance approval is required
• SIT review is required
• Experimentation Brain validation is required
• structural risk exceeds Ads Brain authority
• platform behaviour is unclear or abnormal
Escalate means:
Decision authority passes upward or sideways to the relevant governing Brain.
Decision Tree
The high-level logic is:
Opportunity approved for test
↓
Creative and experiment launched
↓
Signals collected
↓
Campaign reviewed
↓
Creative scored
↓
Decision Engine applied
↓
One valid output selected
Decision Priority Order
When multiple decisions appear possible, apply this priority order:
• Governance integrity
• Capital protection
• Experiment validity
• Creative signal quality
• Scaling opportunity
If governance or capital discipline conflicts with performance optimism, governance wins.
Relationship to Campaign Review Protocol
The Campaign Review Protocol defines how campaign signals are reviewed.
The Decision Engine converts those reviewed signals into an operational recommendation.
Relationship to Creative Performance Scorecard
The Creative Performance Scorecard evaluates creative strength.
The Decision Engine uses that score to determine whether the creative should continue, iterate, scale, or retire.
Relationship to Scaling Intelligence
Scaling Intelligence defines how controlled expansion occurs.
The Decision Engine determines whether the campaign qualifies to enter scaling evaluation.
Relationship to Finance Brain
Finance Brain governs capital allocation.
The Decision Engine may recommend Scale or Escalate, but Finance determines whether capital expansion is approved.
Relationship to Experimentation Brain
Experimentation Brain governs experiment validity.
If a test lacks statistical reliability or suffers contamination, the Decision Engine must favour Pause, Continue, or Escalate rather than false confidence.
Relationship to SIT Brain
SIT Brain governs structural integrity.
If structural compliance is uncertain, the Decision Engine must defer to SIT review.
Recording Requirement
Every Decision Engine output must be recorded alongside:
• campaign name
• experiment ID
• decision output
• reason for decision
• primary signals observed
• next recommended action
This requirement ensures cumulative decision intelligence over time.
Future Expansion
The Decision Engine may eventually support:
• weighted decision scoring
• automated signal interpretation
• dashboard integration
• campaign state recommendations
• cross-platform decision logic
Final Rule
Ads Brain decisions must remain disciplined.
No campaign may be scaled on optimism.
No campaign may be retired on frustration.
No campaign may be iterated without a defined variable.
No campaign may continue indefinitely without decision review.
The Decision Engine exists to convert signal into controlled action.
Drift Protection
The system must prevent:
• campaign action being taken without a valid decision output
• optimistic scaling without governance or Finance discipline
• frustration-based retirement without structured evidence
• iteration without a defined variable change
• contaminated or weak experiments producing false confidence
• campaigns continuing indefinitely without review discipline
Decision logic must remain explicit, bounded, and auditable.
Architectural Intent
Ads Brain – Decision Engine exists to turn campaign analysis into governed action rather than reactive behaviour.
Its role is to ensure that Ads Brain converts creative, campaign, audience, platform, and governance signals into one valid operational recommendation at a time while preserving capital discipline, experiment integrity, and structural compliance.
Change Log
Version: v1.2
Date: 2026-03-17
Author: MWMS HeadOffice
Change: Standardised page metadata and naming to align with the locked MWMS standards pack. Normalised Status from “Operational Engine” to “Active”, standardised title usage to “Ads Brain – Decision Engine”, preserved the original signal-to-action logic, decision priority order, governance protections, and recording requirements, and retained the document as a Specification under the locked MWMS document taxonomy.
Version: v1.1
Date: 2026-03-14
Author: MWMS HeadOffice / Ads Brain
Change: Rebuilt page to align with MWMS document standards. Added standardised document header, introduced Purpose / Scope / Definition / Rules structure, normalised decision outputs, inputs, rules, relationships, and recording requirements, and preserved the original signal-to-action logic, decision priority order, and governance protections.
Version: v1.0
Date: 2026-03-13
Author: Ads Brain / MWMS HeadOffice
Change: Initial creation of Ads Brain – Decision Engine defining structured campaign decision outputs, input categories, action rules, decision priority order, inter-brain relationships, and recording requirements.
CHANGE IMPACT
Pages Created: None
Pages Updated: Ads Brain – Decision Engine
Pages Deprecated: None
Registries Requiring Update:
• MWMS Architecture Registry
• MWMS Brain Registry
• MWMS Brain Interaction Map
• MWMS Canon Hierarchy Map
Canon Version Update Required: No
Change Log Entry Required: Yes
END – ADS BRAIN – DECISION ENGINE v1.2