Document Type: Framework
Status: Active
Version: v1.0
Authority: MWMS HeadOffice
Applies To: customer cohort interpretation, behavioural pattern extraction, retention diagnostics
Parent: Research Brain Canon
Last Reviewed: 2026-04-12
Purpose
The Research Brain Cohort Behaviour Analysis Framework defines how customer cohorts are analysed to identify behavioural patterns that influence retention, lifetime value, and revenue predictability.
Cohort analysis allows MWMS to understand how different groups of customers behave over time rather than relying on aggregate averages.
The purpose of this framework is to:
• identify behavioural differences across customer groups
• detect retention decay patterns
• identify high-value customer characteristics
• detect acquisition channel quality differences
• identify repeat purchase timing patterns
• identify lifecycle optimisation opportunities
• improve forecasting reliability
• improve traffic quality evaluation
• improve monetisation strategy decisions
Cohort structure improves interpretability of customer behaviour.
Improved interpretability improves optimisation precision.
Scope
This framework applies to:
• acquisition cohorts
• behavioural cohorts
• retention cohorts
• repeat purchase cohorts
• lifecycle stage cohorts
• traffic source cohorts
• product-based cohorts
• campaign-based cohorts
This framework governs how Research Brain extracts behavioural intelligence from grouped customer data.
It does not govern:
• experimentation methodology
• attribution modelling implementation
• marketing channel execution logic
Those remain governed by Experimentation Brain and Ads Brain systems.
Definition / Rules
Core Principle
Customers acquired at different times or through different conditions behave differently over time.
Aggregate averages obscure behavioural differences.
Cohort structure reveals behavioural patterns.
Behavioural patterns enable optimisation leverage.
Acquisition Cohort Structure
Customers grouped by acquisition period allow identification of:
repeat purchase decay curves
time-to-second-purchase distribution
seasonal behavioural variation
quality variation across acquisition windows
Acquisition cohort comparison reveals performance differences across marketing periods.
Behavioural Pattern Extraction
Cohorts allow identification of:
high-value customer characteristics
high-retention product combinations
high-frequency purchase behaviours
high-monetary value patterns
Behavioural patterns improve targeting decisions.
Targeting decisions improve traffic quality.
Retention Decay Analysis
Retention behaviour changes over time.
Typical pattern:
highest drop-off occurs after first purchase.
Second purchase probability strongly influences lifetime value potential.
Improving second purchase rate significantly influences revenue predictability.
Retention decay curves identify lifecycle optimisation opportunities.
Traffic Quality Interpretation
Different traffic sources produce different behavioural profiles.
Examples include:
channels producing higher repeat purchase probability.
channels producing lower long-term value customers.
channels producing promotion-sensitive customers.
Traffic quality must be evaluated based on long-term behaviour.
Short-term conversion rates may misrepresent customer value.
Product-Based Cohort Analysis
Products acquired on first purchase influence future behaviour patterns.
Examples include:
entry products associated with high repeat purchase probability.
entry products associated with low lifetime value behaviour.
Product-level cohort patterns inform merchandising prioritisation.
Merchandising structure influences long-term value outcomes.
Cohort-Based Forecasting Inputs
Cohort behaviour provides forecasting signals including:
expected repeat purchase timing patterns.
expected revenue decay curves.
expected customer lifetime contribution.
Cohort-informed forecasts improve growth predictability.
Predictability improves decision confidence.
Relationship to RFM Segmentation Frameworks
Cohort analysis complements recency, frequency, monetary segmentation structures.
Combined usage improves behavioural resolution.
Behavioural resolution improves lifecycle optimisation precision.
Relationship to CLV Interpretation
Cohort structure improves interpretation of lifetime value estimates.
CLV should be treated as directional guidance rather than fixed prediction.
Cohort variability introduces uncertainty into lifetime value estimation.
Uncertainty must be considered in acquisition investment decisions.
Relationship to Traffic Allocation Decisions
Cohort behaviour informs traffic allocation prioritisation.
Traffic sources producing high-retention customers may justify higher acquisition cost tolerance.
Traffic sources producing low-retention customers require tighter acquisition cost discipline.
Drift Protection
The system must prevent:
relying solely on aggregate performance averages
ignoring behavioural differences across acquisition periods
optimising for short-term conversion without long-term behaviour consideration
assuming all customers produce similar lifetime value
ignoring repeat purchase timing patterns
ignoring acquisition channel behavioural differences
Behavioural variability must be acknowledged.
Ignoring variability reduces optimisation precision.
Architectural Intent
Research Brain Cohort Behaviour Analysis Framework exists to provide behavioural resolution beyond aggregate metrics.
Its role is to improve decision quality by identifying behavioural patterns that influence retention, lifetime value, and revenue predictability.
Improved behavioural clarity improves optimisation accuracy.
Improved optimisation accuracy improves system stability.
Future Expansion
Cohort analysis may integrate:
behaviour-weighted segmentation models
predictive retention scoring
traffic quality weighting algorithms
product sequence modelling
lifecycle stage prediction models
churn probability estimation
Future development may improve behavioural prediction precision.
Final Rule
Customer averages hide behavioural variability.
Cohorts reveal behavioural structure.
Research Brain must prioritise behavioural interpretability discipline.
Change Log
Version: v1.0
Date: 2026-04-12
Author: MWMS HeadOffice
Change: Initial creation of Research Brain Cohort Behaviour Analysis Framework defining behavioural cohort logic, retention decay interpretation structure, drift protection requirements, and architectural intent aligned with MWMS Canon standards.
CHANGE IMPACT
Pages Created:
• Research Brain Cohort Behaviour Analysis Framework
Pages Updated:
None
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