Research Brain Offer Evidence Standards

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.