Document Type: Standard
Status: Active Standard
Version: v1.0
Authority: Research Brain (Subordinate to MWMS HeadOffice)
Applies To: All offer-level and opportunity-level research conducted inside Research Brain
Parent: Research Brain Architecture
Linked Systems:
Research Brain Canon
Research Brain — Offer / Opportunity Research Task Specification
Research Intelligence Database
Affiliate Brain
Finance Brain
MWMS Decision Authority Matrix
Last Reviewed: 2026-03-26
Purpose
This standard defines what counts as acceptable evidence when researching an offer or opportunity inside Research Brain.
The purpose of this standard is to prevent weak, promotional, duplicated, or unverified information from being treated as decision-grade intelligence.
Research Brain exists to improve decision quality through information quality.
That requires evidence discipline.
This standard defines:
• what evidence types are allowed
• how evidence quality is classified
• how evidence strength affects confidence
• what evidence must not be overstated
• how factual integrity is preserved
Scope
This standard applies to all Research Brain work involving:
• offer research
• opportunity research
• market research
• funnel review
• competitor review
• compliance risk observation
• signal validation
• downstream research handoff
This standard applies whether the request originates from:
• Research Brain
• Affiliate Brain
• Ads Brain
• PPL Brain
• AIBS
• HeadOffice
• another subordinate MWMS system
This standard governs evidence quality.
It does not govern:
• capital decisions
• ROI decisions
• campaign launch decisions
• scaling decisions
• budget sizing
• final execution authority
Core Principle
Not all information is evidence.
Not all evidence is equal.
Research Brain must distinguish between:
• observable fact
• supported inference
• weak signal
• promotional claim
• speculation
Research Brain must not allow repeated weak claims to become false certainty.
Research Brain must not treat volume of noise as validation.
Evidence Integrity Rule
Evidence quality must always be assessed before conclusions are formed.
The strength of a research conclusion must never exceed the strength of the evidence supporting it.
Where evidence is weak, confidence must remain weak.
Where evidence is incomplete, uncertainty must remain visible.
Where evidence is contradictory, contradiction must remain visible.
Evidence Source Classes
Research Brain may use evidence from the following classes.
1. Observable Page Evidence
Evidence directly visible on the offer page, landing page, funnel page, pricing page, checkout page, advertorial, or connected public funnel asset.
Examples:
• product promise
• page structure
• pricing display
• headline language
• CTA structure
• visible proof elements
• risk language
• positioning language
This is valid evidence of what is being presented.
It is not automatically evidence that the claims are true.
2. Observable Market Evidence
Evidence visible from competitor pages, comparable offers, public ad examples, category patterns, or repeated structural observations across a niche.
Examples:
• repeated angle patterns
• repeated positioning structures
• repeated funnel formats
• common proof styles
• common pricing patterns
This is valid evidence of market behaviour patterns.
It is not automatically evidence of profitability.
3. Internal MWMS Evidence
Evidence produced from internal MWMS systems.
Examples:
• experiment results
• campaign outcomes
• testing logs
• performance history
• offer graveyard history
• route outcomes
• prior verdict calibration
This is the highest-value evidence class when relevant and recent.
4. Analytical / Interpretive Evidence
Evidence derived from expert commentary, educational material, industry analysis, or informed external interpretation.
Examples:
• strategic analysis
• framework-based reviews
• research commentary
• educational material
• market interpretation
This may support understanding.
It must not override stronger direct evidence.
5. Vendor / Promotional Evidence
Evidence originating from the seller, advertiser, vendor, affiliate manager, promotional material, or self-interested source.
Examples:
• vendor claims
• “best converting” claims
• earnings claims
• product uniqueness claims
• hype-based sales language
• self-reported credibility statements
This is the weakest evidence class for factual validation.
It may be used to understand positioning.
It must not be used as standalone proof of truth.
Evidence Source Tier
Research Brain must classify evidence using source tiers.
Tier 1 — Internal / Experimental
Highest strength.
Examples:
• internal test results
• internal performance data
• direct internal measurement
• verified controlled observations
Use for:
validated behavioural understanding
real-world confirmation
high-confidence support
Tier 2 — External Observable
Strong but not final.
Examples:
• competitor pages
• public offer pages
• visible funnel structures
• market pattern observation
• public ad examples
Use for:
structural interpretation
market comparison
funnel classification
positioning analysis
Tier 3 — Analytical / Educational
Moderate strength.
Examples:
• expert commentary
• educational content
• interpretive frameworks
• market commentary
Use for:
context
explanation
supporting interpretation
Do not treat as direct proof.
Tier 4 — Promotional / Unverified
Weakest strength.
Examples:
• vendor claims
• hype language
• social chatter
• unsupported testimonials
• forum speculation
• copied promotional assertions
Use for:
signal awareness only
positioning understanding only
Do not treat as validated truth.
Tier 4 evidence cannot become validated merely because it is repeated.
Evidence Density
Research Brain must classify Evidence_Density.
Low Evidence Density
Very little observable material.
Examples:
• one page only
• no competitor context
• no pricing clarity
• no visible funnel continuation
• few evidence points
Use low confidence.
Medium Evidence Density
Several useful observations but incomplete validation.
Examples:
• multiple funnel pages visible
• some competitor comparison
• some structural clarity
• some risk observations
Use moderate confidence with declared limits.
High Evidence Density
Broad and consistent evidence base.
Examples:
• multiple pages reviewed
• competitor comparison available
• structure is observable
• evidence supports consistent interpretation
• internal or strong comparative support exists
High confidence still depends on source tier quality.
High density does not fix poor source quality.
Confidence Rule
Confidence_Level must reflect both:
• Evidence_Source_Tier
and
• Evidence_Density
Confidence must not be derived from narrative fluency.
Confidence must not be derived from how convincing a page sounds.
Confidence must not be derived from repetition of vendor claims.
A polished page is not the same as strong evidence.
Observation vs Inference Rule
Research Brain must separate:
Observation
What is directly visible.
Examples:
• page uses advertorial structure
• headline promises rapid relief
• checkout step is visible
• pricing is monthly
• funnel uses quiz mechanism
Inference
What is reasonably interpreted from observations.
Examples:
• niche appears crowded
• angle appears common
• offer may struggle with differentiation
• structure suggests lead qualification intent
Inference is allowed.
Inference must be marked as inference.
Inference must not be presented as direct fact.
Prohibited Evidence Failures
Research Brain must not do the following:
Treat promotional claims as validated truth
Example:
“Vendor says it converts well, therefore promising.”
Not allowed.
Treat copied market chatter as evidence
Example:
“People online say this niche is hot, therefore high opportunity.”
Not allowed.
Treat one visible page as full-funnel truth
Example:
“One page is weak, therefore the whole offer is weak.”
Not allowed.
Treat polished design as commercial proof
Example:
“Looks premium, therefore must perform.”
Not allowed.
Inflate uncertainty into confidence
Example:
“Could be good” written as “is promising.”
Not allowed.
Required Evidence Declaration
Each research task must declare:
Evidence_Source(s)
Evidence_Source_Tier
Evidence_Density
Confidence_Level
Where useful, the task should also declare:
Known Gaps
Unverified Assumptions
Missing Evidence
Conflicting Signals
Evidence Use by Section
Product / Offer Overview
Use:
observable page evidence
Avoid:
performance assumptions
Offer Model
Use:
visible pricing, offer structure, billing cues, lead-form behaviour
Avoid:
assuming backend monetisation
Funnel / Page Type
Use:
observable structure and step logic
Avoid:
inventing unseen steps
Market / Niche Read
Use:
observable positioning, competitor comparison, repeated niche patterns
Avoid:
claiming demand certainty without stronger support
Competitor / Angle Observations
Use:
public examples, visible angles, repeated framing
Avoid:
declaring uniqueness without comparison
Compliance / Platform Risk
Use:
visible claim language and known platform-sensitive patterns
Avoid:
declaring formal approval or formal violation status without authority
Strengths
Use:
evidence-linked structural positives
Avoid:
subjective praise without support
Weaknesses / Concerns
Use:
observable gaps, uncertainty, structural issues
Avoid:
overstating negative assumptions as settled fact
Research Verdict Support
Use:
the weighted evidence picture
Avoid:
single-source conclusions
Recommended Next Step
Use:
evidence-supported advisory judgement
Avoid:
execution authority language
Relationship to Affiliate Brain
Affiliate Brain may use Research Brain outputs to decide whether an opportunity deserves testing attention.
Research Brain must therefore provide:
• factual structure
• evidence-aware interpretation
• clear uncertainty
• clear next-step framing
Research Brain must not attempt to replace Affiliate Brain testing judgement.
Relationship to Finance Brain
Finance Brain may later evaluate the economics of opportunities that pass early-stage research screening.
Research Brain must therefore avoid pretending to know:
• ROI
• margin durability
• spend tolerance
• scaling economics
• capital efficiency
Those remain outside Research Brain scope.
Implementation Guidance
The Research Brain UI should allow evidence-aware research capture.
Where possible, the interface should support:
• source logging
• confidence declaration
• evidence-tier declaration
• structured section outputs
• separation of observations from interpretation
Initial implementation may use existing storage fields provided the evidence structure is preserved in output format.
Drift Protection
Research Brain must not become:
• a dumping ground for random notes
• a copywriting layer
• a hype amplifier
• a pseudo-finance layer
• a false-certainty engine
Research Brain must remain:
• evidence-disciplined
• structured
• traceable
• bounded by authority
• clear about uncertainty
Architectural Intent
This standard exists to ensure that opportunity research inside MWMS is governed by evidence quality rather than by narrative persuasiveness.
The intent is to make Research Brain:
• useful to Affiliate Brain
• compatible with later Finance Brain review
• reusable across future research task types
• resistant to weak-source drift
• reliable enough to improve system-level decision quality
Final Rule
Research Brain may interpret evidence.
Research Brain may not invent evidence.
Research Brain may support decision quality.
Research Brain may not simulate certainty beyond what evidence can support.
Change Log entry
Use this in the Research Brain Change Log after saving:
2026-03-26 — Added Offer Evidence Standards v1.0
Change Type: Structural Extension
Authority: Research Brain
Scope Impact: Defines evidence quality rules for offer and opportunity research
Parent Architecture Impact: None
Decision Authority Impact: None
Backward Compatibility: Maintained
Summary
Added new subordinate standard:
Research Brain — Offer Evidence Standards v1.0
This standard defines acceptable evidence classes, source tiers, evidence density handling, confidence limits, observation-vs-inference rules, and prohibited evidence failures for offer-level research inside Research Brain.
Reason for Change
Offer research required stronger evidence discipline to prevent weak internet information, vendor hype, copied claims, and unverified signals from being treated as decision-grade intelligence.
Architectural Intent
Improve factual integrity of Research Brain outputs.
Support stronger Affiliate Brain screening.
Prepare cleaner downstream handoff to Finance Brain without crossing authority boundaries.
Migration Requirement
None.
Applies to new research tasks and future structured template builds.
The page list you showed confirms Research Brain currently has its Canon, Architecture, Employee Registry, Database Schema, Intelligence Database, and the new Offer / Opportunity Research Task Specification, but it is still comparatively thin versus more built-out brains like Affiliate and Ads, so adding these operational standards is the right move.
Next best page after this is either Research Brain — Research Verdict Framework or Research Brain — Offer Source Validation Framework.