sub-modules
Demand Planning
Stop arguing about the forecast — version it, test it, and let accuracy decide.
Forecast meetings run on the same theater as project status: the loudest model wins until reality audits everyone. Demand Planning replaces theater with versioning — statistical baselines, human overrides tracked against them, accuracy reviews both sides can see. The forecast that earns trust is the one that survived comparison.
Planning cycle
Each stage below moves signals toward executable supply — sensed, planned, and committed with the evidence to defend every number.
data Viz
data View
section: stages
Sense
Orders, forecasts, and service demand fuse into signals with overrides versioned against statistical baselines. Bias stays visible instead of hiding.
Solve
Deterministic solvers generate replenishment, transport, and labor plans with execution profiles. Infeasible stretches return certificates, not partials.
Commit
Accepted plans convert to purchase, transfer, and roster actions with owner gates. Commitments carry their input snapshots for later review.
Demand guarantees
- Signal lineage
- Orders, forecasts, and service demand cited as plan inputs with versions
- Replayable runs
- Identical inputs reproduce identical plans with execution profiles
- Honest infeasibility
- Binding constraints named with sensitivity instead of silent partials
- section
- capabilities
Demand by industry
Demand sensing bends per industry — same signal lineage, different volatility stakes.
data Viz
data View
section: industries
Retail
SKU-level forecasts with seasonality, promotions, and store-cluster patterns. Replenishment reacts to sell-through daily instead of weekly buyers' hunches.
Manufacturing
Program and order-driven demand with BOM explosion into component plans. Engineering changes replan affected branches with the revision cited.
Food and Agriculture
Seasonal and weather-influenced demand with harvest windows constraining supply. Perishability horizons shape lot sizing before waste happens.
Demand Planning questions
Section 1
What happens when a plan is infeasible?
The solver returns a certificate naming the binding constraints with sensitivity analysis, instead of a silent partial plan. Planners adjust constraints and rerun.
Section 2
How do forecast overrides work?
Overrides version against the statistical baseline with author, reason, and effective dates — both lines stay visible so accuracy reviews stay honest.
Section 3
Sales always overrides the forecast. How do we stop the gaming?
You don't stop it — you score it. Override accuracy reviews show whose judgment beats the model and whose doesn't, and the conversation shifts from politics to evidence.
data View
section: faqs
content section
value: Planning truth feeds operations: demand plans drive replenishment, labor plans staff shifts, and transport plans book capacity — all from the same signals, all replayable from the same sealed inputs.
Section 6
Demand sensing from orders, forecasts, and service signals
Section 7
Replenishment and transportation plan generation
Section 8
Deterministic solver runs with execution profiles
Section 9
Infeasibility certificates with sensitivity analysis