Skip to main content

Input data requirements

AInventory operates with three categories of data: demand history, forecast per phase and version, and per-SKU inventory parameters. Without these data loaded and kept up to date, KPI calculations, FVA, and inventory simulation do not produce valid results.

1. Demand history

The demand history is the foundation of the Collaborative Forecast module. AInventory uses it to:

  • Generate the Baseline (AI) forecast via statistical models.
  • Calculate the forecast error (D − F) once each period has elapsed.
  • Build the KPIs for each closed version.

The minimum required fields per record are:

FieldDescription
SKU identifierUnique product code within the tenant
PeriodMonth or week the demand corresponds to
Actual demand (D)Quantity sold / consumed in that period
Historical coverage

TODO: formal minimum coverage requirements (number of periods, handling of zero-demand records, seasonality thresholds) are not detailed in the available sources. Please consult the Directrix implementation team.

2. Forecast per phase and version

For the collaborative cycle to work, each phase must have its forecast values loaded before the workflow advances to the next team. The minimum required fields are:

FieldDescription
SKU identifierSame code as in the demand history
Projected periodMonth or week being forecast
Forecast value (F)Projected quantity for that period
PhaseBaseline / Sales / Marketing / Consensus
VersionPlanning cycle identifier

The version must be in open status for values to be editable. Once closed, values are locked and the system uses them to calculate KPIs and FVA as periods elapse.

3. Per-SKU inventory parameters

The Inventory Optimization module requires specific parameters for each SKU in order to run the Monte Carlo simulation. The main parameters are:

ParameterDescription
Holding costCost of keeping one unit in inventory per period (as a fraction of item value or as an absolute amount)
Shortage costCost or penalty per unit of unsatisfied demand
Lead timeReplenishment time from order placement to receipt
Initial inventoryUnits on hand at the start of the simulation horizon
In-transit inventoryUnits already ordered but not yet received
Sale priceUnit price of the SKU (used in the opportunity cost calculation)
Forecast adjustmentCorrection factor applied to the Consensus forecast before simulation

For the complete description of each parameter, its units, and recommended ranges, see the Inventory parameters reference.

Data quality

TODO: formal data quality and coverage requirements (minimum history length, treatment of new SKUs with no history, acceptable completeness thresholds) are not detailed in the available sources. This section will be updated with the validation criteria from the Directrix onboarding process.

Initial setup

If you are in the onboarding process, the Directrix implementation team will guide you through the initial load of all three data categories. Once the platform is in production, data updates are managed from the Configuration section (Administrators only) in the Sidebar.