Document Type: Framework
Status: Active
Version: v1.0
Authority: MWMS HeadOffice
Parent: Finance Brain Canon
Slug: finance-brain-forecast-sensitivity-framework
Last Reviewed: 2026-04-13
Purpose
The Finance Brain Forecast Sensitivity Framework defines how MWMS identifies which variables most strongly influence forecast outcomes and how changes in those variables alter planning reliability.
Forecasts are not only affected by revenue totals.
They are affected by the sensitivity of key assumptions.
A forecast may appear stable while remaining highly fragile.
Sensitivity analysis improves:
• planning realism
• capital discipline
• risk visibility
• scenario interpretation quality
• growth pacing decisions
• survivability protection
Forecast reliability depends on understanding which variables can materially change outcomes.
Scope
This framework applies to:
• forecast input sensitivity analysis
• revenue assumption stress testing
• CAC sensitivity interpretation
• retention sensitivity interpretation
• AOV sensitivity interpretation
• cohort-behaviour sensitivity interpretation
• risk-aware scenario planning
• identification of fragile forecast assumptions
This framework governs how Finance Brain interprets forecast fragility and variable impact.
It does not govern:
• final capital approval
• accounting treatment
• experiment validation authority
• paid media execution decisions
• lifecycle execution decisions
Those remain governed by Finance Brain capital rules, Accounting systems, Experimentation Brain, Ads Brain, and Ecommerce Brain.
Definition / Rules
Core Principle
Forecast accuracy depends on assumption quality.
Forecast stability depends on assumption sensitivity.
Some assumptions have small effect on total outcome.
Some assumptions have large effect on total outcome.
Sensitivity analysis identifies which assumptions most strongly influence revenue, cash flow, and growth stability.
Sensitivity awareness improves decision quality.
Forecast Sensitivity Definition
Sensitivity is the degree to which a forecast changes when a key variable changes.
Examples:
small retention change may produce large revenue effect
small CAC increase may reduce growth efficiency sharply
small AOV increase may significantly improve survivability
small repeat purchase shift may materially alter cohort value
Sensitivity must be understood before forecast confidence is accepted.
Core Sensitivity Inputs
Finance Brain should evaluate at minimum the following sensitivity categories.
Acquisition Sensitivity
Measures how forecast outcomes respond to changes in traffic acquisition conditions.
Key variables include:
• CAC
• traffic volume
• traffic quality
• conversion rate from paid traffic
• platform cost inflation
Acquisition sensitivity is high when growth depends heavily on constant new-customer inflow.
Retention Sensitivity
Measures how forecast outcomes respond to changes in repeat purchase behaviour and customer durability.
Key variables include:
• second purchase probability
• churn rate
• repeat purchase frequency
• engagement persistence
• reactivation success rate
Retention sensitivity is often one of the strongest forecast drivers.
Average Order Value Sensitivity
Measures how forecast outcomes respond to changes in transaction value.
Key variables include:
• bundle uptake
• cross-sell effectiveness
• premium mix
• product mix variation
• discount dependency
AOV sensitivity reveals how much monetisation structure influences total revenue.
Cohort Decay Sensitivity
Measures how forecast outcomes respond to changes in customer-file deterioration over time.
Key variables include:
• cohort decay speed
• active-file contraction
• time-to-inactivity
• cohort revenue tail length
High cohort decay sensitivity indicates fragile revenue durability.
Margin Sensitivity
Measures how forecast outcomes respond to changes in realised profitability, not just revenue.
Key variables include:
• gross margin variation
• discount depth
• shipping cost pressure
• refund rate changes
• platform fee changes
Revenue growth without margin stability may still weaken survivability.
Timing Sensitivity
Measures how forecast outcomes respond to delays in behaviour or revenue realisation.
Key variables include:
• slower repeat purchase timing
• delayed payback
• campaign learning delay
• slower conversion velocity
Timing shifts may materially affect cash flow even when topline forecast remains similar.
Sensitivity Classification
Sensitivity should be classified into structured impact levels.
Low Sensitivity
Small variable shifts produce minimal forecast movement.
These variables are not primary forecast risk drivers.
Moderate Sensitivity
Variable shifts materially affect planning assumptions but do not immediately destabilise the model.
These variables require monitoring.
High Sensitivity
Small changes materially alter revenue, profitability, or survivability assumptions.
These variables require active risk visibility and tighter review cadence.
Sensitivity Testing Logic
Forecast inputs should be stress-tested using controlled changes.
Examples:
• CAC +10%
• AOV -5%
• second purchase rate -10%
• cohort decay faster by one interval
• refund rate +3%
Sensitivity testing reveals which assumptions create the greatest forecast fragility.
Stress testing improves planning realism.
Relationship to Cohort Revenue Forecasting Framework
Cohort Revenue Forecasting defines the structural forecast model.
Forecast Sensitivity Framework defines how fragile that model is to assumption change.
Forecast structure and sensitivity must be interpreted together.
Forecasts without sensitivity review create false confidence.
Relationship to Percentile Scenario Forecasting Framework
Percentile Scenario Forecasting expresses ranges of likely outcomes.
Forecast Sensitivity Framework explains which variables widen or narrow those ranges.
Sensitivity analysis improves scenario interpretation discipline.
Relationship to Revenue Leakage Diagnostic Framework
Revenue leakage influences forecast sensitivity.
Examples:
• high retention leakage increases retention sensitivity
• monetisation leakage increases AOV sensitivity
• acquisition leakage increases CAC sensitivity
Leakage visibility improves sensitivity accuracy.
Relationship to LTV Signal Framework
LTV estimation is directional and uncertain.
Sensitivity analysis should identify how dependent growth projections are on LTV assumptions.
High dependence on uncertain LTV assumptions increases forecast risk.
Sensitivity testing protects against overconfidence.
Planning Use
Sensitivity analysis should be used in:
• budget planning
• scaling pacing
• capital preservation decisions
• survivability analysis
• growth target realism checks
• offer expansion decisions
Sensitivity analysis should be reviewed before aggressive expansion assumptions are adopted.
Failure Modes Prevented
This framework prevents:
• relying on forecasts without understanding key assumption fragility
• assuming stable acquisition conditions
• assuming repeat purchase behaviour is fixed
• underestimating the impact of CAC inflation
• overconfidence in average-case models
• using topline revenue forecast without profitability awareness
Forecasting must remain risk-aware.
Drift Protection
The system must prevent:
• single-point forecasts being treated as stable without variable testing
• retention assumptions being accepted without sensitivity review
• margin risk being ignored because topline appears healthy
• timing effects being excluded from planning
• sensitivity review being skipped during expansion planning
Sensitivity analysis must remain a standard part of forecast discipline.
Architectural Intent
Finance Brain Forecast Sensitivity Framework exists to make forecast fragility visible before capital is committed.
Its role is to show which assumptions matter most, where planning risk is concentrated, and how quickly seemingly small shifts can change growth outcomes.
Visible fragility improves decision quality.
Improved decision quality protects survivability.
Change Log
Version: v1.0
Date: 2026-04-13
Author: MWMS HeadOffice
Change:
Initial creation of Forecast Sensitivity Framework defining variable-impact analysis across acquisition, retention, AOV, cohort decay, margin, and timing assumptions.
END – FINANCE BRAIN FORECAST SENSITIVITY FRAMEWORK v1.0