Back to all disciplines
DATUM FIELD CHAPTER · DEMAND

Discipline: Demand

Ten canonical operational blueprints cataloged under the Demand field domain.

#002
Forecast Diagnostics

Forecasts are consistently too high or too low

Are we repeatedly leaning in one direction?

#002 CUSUM Bias Detection+23 Over-forecast
Actual Demand
Forecast Baseline
CUSUM OFFSET:-0 u
SYSTEM BIAS:+23 Shift
💡Small errors in one direction create compounding operational waste.
Blueprint
#003
Demand Analytics

Promotions make demand unpredictable

How much of the spike is promotion rather than normal demand?

#003 Promo Uplift & Base+90% Uplift
TRUE BASELINE (100u)PEAK: 190u
Discount: 28% OFFCannibalization: -25%
Promo Volume Lift:+90% Units
Post-Promo Dip:-25% Forward-Buy
5% Discount50% Heavy Discount
Verified ProofBlueprint
#004
Demand Analytics

New products have no historical data

How do we forecast something that did not exist before?

#004 Analogue LaunchSpeed: 1.5x
ANALOGUE PRIOR
Benchmark PriorPeak: 100k Units
Adoption Velocity:1.5x S-Curve
Cold-Start Confidence:88% Bayesian
Slow RampFast S-Curve
Verified ProofBlueprint
#005
Demand Analytics

Seasonal demand keeps surprising us

Which patterns repeat and when?

#005 Fourier SeasonalityK = 3 Orders
Annual + Weekly + Payday95% R² Captured
Fourier Harmonics:K = 3 Orders
Seasonal Fit:95% Model Fit
1 Cycle (Simple)4 Cycles (Complex STL)
Verified ProofBlueprint
#006
Demand Analytics

We cannot forecast at SKU × location level reliably

How does granularity change forecast quality?

#006 MinT Reconciliation-26% Trace Variance
TOTALEASTWEST
Coherence: 100% Math ReconciledMinT Optimal
Reconciliation:MinT Optimal
Variance Drop:-26% Error
Bottom-UpMinT Trace Optimal
Verified ProofBlueprint
#007
Demand Analytics

Stockouts are corrupting our demand history

Are recorded sales hiding demand we could not fulfill?

#007 Censored Demand+38% Hidden
STOCKOUT CEILING (CAP)TRUE UNCONSTRAINED DEMAND
Observed Sales: TruncatedEM Algorithm Recovery
Censored Sales Gap:+38% Units
Stockout Days:14 / 30 Days
Naive Sales DataEM Unconstrained Math
Verified ProofBlueprint
#008
Demand Analytics

Different teams produce different forecasts

Why do Sales, Finance and Supply Chain disagree?

#008 Demand Model68% Calibrated
Simulation Active68% Parameter Match
Model Calibration:68% Optimization
Risk Margin:High Confidence
Min BoundOptimal Frontier
Verified ProofBlueprint
#009
Demand Analytics

Forecast accuracy improves but the business does not

Why does a better forecast not automatically improve inventory or service?

#009 Demand Model68% Calibrated
Simulation Active68% Parameter Match
Model Calibration:68% Optimization
Risk Margin:High Confidence
Min BoundOptimal Frontier
Verified ProofBlueprint
#010
Demand Analytics

We don't know how uncertain the forecast is

What range of outcomes should we plan for?

#010 Demand Model68% Calibrated
Simulation Active68% Parameter Match
Model Calibration:68% Optimization
Risk Margin:High Confidence
Min BoundOptimal Frontier
Verified ProofBlueprint