Document Type: Framework
Status: Structural
Version: v1.1
Authority: HeadOffice
Applies To: Data Brain, Experimentation Brain, Ads Brain, Affiliate Brain, Engineering Layer
Parent: Data Brain
Last Reviewed: 2026-04-22
Purpose
The Data Layer Architecture Framework defines how behavioural data is structured and transmitted from digital environments into MWMS measurement systems.
It ensures signals are:
stable
reliable
interpretable
implementation-independent
resilient to site changes
suitable for experimentation
consistent across environments
A structured data layer reduces measurement fragility and increases signal continuity.
The framework prevents:
unstable tracking logic
dependence on page structure scraping
hidden signal breakage
inconsistent event transmission
implementation-specific signal distortions
duplicate signal pathways
ambiguous event parameter meaning
Data must be delivered through stable structural pathways.
Stable signal architecture improves decision reliability.
Reliable signal transmission improves experiment validity.
Reliable signal continuity improves optimisation accuracy.
Scope
This framework governs:
event data transmission structure
parameter transmission structure
data schema design
data structure consistency
implementation independence
signal reliability infrastructure
tracking stability principles
multi-environment signal consistency
schema expansion discipline
signal routing clarity
This framework applies to:
web environments
landing pages
funnels
applications
advertising landing environments
server-side tracking environments
client-side tracking environments
tag manager environments
This framework applies regardless of tracking platform.
Signal meaning must remain consistent across implementation tools.
Core Principle
Signals should be transmitted using structured data architecture rather than inferred from page structure.
Behavioural data must be intentionally provided.
Signals should not rely on fragile extraction methods.
Stable measurement requires structured data delivery.
Structured signal transmission improves:
interpretability
maintainability
implementation flexibility
long-term measurement reliability
Definition of Data Layer
A data layer is a structured mechanism for transmitting behavioural event information from digital environments to measurement systems.
A data layer provides:
event definitions
parameter values
contextual behavioural information
structured signal delivery
The data layer separates measurement logic from page presentation logic.
This separation increases reliability and flexibility.
Decoupling measurement from interface structure improves signal stability.
Structured Data Transmission Principle
Measurement signals should be transmitted through structured data objects rather than extracted through interface interpretation.
Structured data provides:
clear signal meaning
explicit behavioural definitions
stable event triggering
consistent parameter structure
reduced dependency on visual page structure
Reduced dependency on page structure improves signal stability.
Signal stability improves experiment reliability.
Signal stability improves decision confidence.
Instability of Interface-Based Tracking
Tracking approaches that rely on page structure interpretation are inherently unstable.
Interface structures change frequently.
Examples of unstable dependencies include:
HTML structure changes
CSS class name changes
layout changes
text changes
component structure changes
DOM restructuring
visual redesign updates
When tracking logic depends on these structures, signals may silently fail.
Unstable tracking produces unreliable data.
Unreliable data produces unreliable decisions.
Structured event delivery reduces silent measurement failure risk.
Data Layer Stability Advantage
Structured data layers improve:
tracking reliability
signal consistency
interpretation clarity
implementation maintainability
future adaptability
signal continuity across redesigns
schema continuity across implementations
Signal structure remains stable even when page structure changes.
Structured transmission reduces implementation fragility.
Stable structure improves long-term measurement continuity.
Event-Based Data Structure
Data layers typically transmit information as structured event objects.
Each event object contains:
event name
parameter values
contextual attributes
event timing information
Example conceptual structure:
event name
parameter key-value pairs
contextual metadata
Event-based structure allows clear mapping between behaviour and measurement.
Event structure must align with Signal Design Specification Framework.
Event structure must align with Measurement Strategy Framework.
Data Schema Requirement
Data layers require defined schema structures.
A schema defines:
event naming structure
parameter naming structure
parameter value formats
allowed value types
required parameters
optional parameters
schema relationships between events
Schema clarity improves implementation consistency.
Schema clarity improves validation reliability.
Schema clarity improves signal interpretability.
Schema clarity improves experiment comparability.
Data Schema Design Principles
Schemas should:
align with signal design framework
align with decision requirements
align with experimentation requirements
avoid unnecessary complexity
avoid excessive nesting
maintain interpretability
remain extendable
maintain parameter consistency
support reusable parameter structures
Schema design must balance flexibility and clarity.
Clear schema improves implementation stability.
Clear schema improves cross-environment comparability.
Parameter Value Clarity Requirement
Parameters must contain clearly interpretable values.
Ambiguous values reduce signal usefulness.
Examples of potential ambiguity:
multiple identifiers representing same concept
unclear classification naming
inconsistent categorical values
inconsistent parameter casing
multiple synonyms representing same concept
Parameter values must:
remain consistent
remain interpretable
remain reusable
remain stable across environments
Clear values improve analytical clarity.
Clear parameter logic improves experiment interpretation reliability.
Multi-Location Event Consistency
Events may be triggered from multiple interface locations.
Examples:
add_to_cart triggered from:
product page
product list
quick view modal
checkout summary
cart adjustment interface
Event meaning must remain consistent regardless of trigger location.
Consistent parameter structure improves behavioural comparability.
Inconsistent trigger structure reduces funnel clarity.
Funnel Continuity Signal Support
Data layer structure must support behavioural progression visibility.
Example funnel sequence:
view_item
add_to_cart
begin_checkout
purchase
Signal continuity improves funnel interpretability.
Funnel continuity improves diagnostic clarity.
Funnel continuity improves optimisation direction confidence.
Data Layer Implementation Independence
Signal meaning must remain independent from specific tools.
Signal definitions should not rely on:
specific analytics platform naming limitations
specific tag manager constraints
specific implementation tools
specific vendor schema limitations
Signal meaning should remain consistent even if implementation technology changes.
Measurement continuity depends on implementation independence.
Implementation independence improves long-term adaptability.
Validation Readiness Principle
Data layer structures must support validation.
Validation ensures:
correct event triggering
correct parameter population
correct value formatting
correct signal timing
correct routing destination
correct event sequence continuity
Validation may occur through:
debug environments
preview modes
logging tools
structured test scenarios
network inspection tools
Validation readiness reduces implementation risk.
Validation readiness improves signal trustworthiness.
Incremental Implementation Principle
Data layers may be implemented progressively.
Signals may be added incrementally.
Incremental implementation allows:
faster deployment
earlier validation
reduced implementation bottlenecks
reduced project delays
Progressive signal implementation must preserve structural consistency.
Incremental expansion must maintain schema continuity.
Relationship to Signal Design Framework
Data layer architecture operationalises signal design.
Signal Design Framework defines:
what signals exist
Data Layer Framework defines:
how signals are transmitted
Both frameworks must remain aligned.
Signal definitions must map cleanly into data schema structures.
Signal structure must remain consistent across environments.
Relationship to Experimentation Brain
Experimentation requires stable measurement inputs.
Unstable data creates false experiment results.
Reliable data architecture improves:
confidence in test results
interpretation reliability
decision accuracy
experiment repeatability
Experiment validity depends on data reliability.
Relationship to Ads Brain
Advertising optimisation requires reliable event tracking.
Conversion signals must be transmitted accurately.
Click signals must remain consistent.
Funnel progression signals must remain stable.
Data layer reliability improves optimisation speed.
Signal continuity improves scaling confidence.
Relationship to Affiliate Brain
Offer evaluation depends on consistent funnel signals.
Conversion tracking must remain stable.
Engagement signals must remain interpretable.
Data reliability improves offer decision accuracy.
Signal continuity improves opportunity evaluation confidence.
Data Layer Governance Considerations
Data schema changes must follow structured review.
Changes to parameter meaning must be documented.
Changes to event naming must preserve interpretability.
Breaking structural continuity reduces historical comparability.
Signal continuity improves learning continuity.
Schema stability improves experiment comparability.
Drift Prevention
Data drift occurs when:
event definitions change without documentation
parameter structures change inconsistently
values change meaning across time
schema structures change unpredictably
event hierarchy changes silently
trigger logic changes without schema update
Drift reduces signal comparability.
Drift reduces decision reliability.
Drift must be controlled through disciplined schema management.
Stable schema improves long-term interpretability.
Minimum Viable Data Layer Principle
Initial implementations should prioritise:
core decision signals
core conversion signals
core behavioural signals
core funnel continuity signals
Data structures may expand as required.
Minimal viable data architecture enables faster deployment and faster learning cycles.
Minimal viable architecture reduces implementation complexity risk.
Minimal viable structure improves early signal reliability.
Change Log
Version: v1.1
Date: 2026-04-22
Author: HeadOffice
Change:
Expanded framework to include:
multi-trigger signal consistency logic
funnel continuity structure support
schema expansion discipline
validation readiness structure
alignment with:
Signal Design Specification Framework
Measurement Strategy Framework
Signal Integrity Framework
Debugging and Validation Framework