Inferensys

Glossary

Order Reservation

Order reservation is the act of creating a hard or soft link between a specific quantity of on-hand or inbound inventory and a customer order to guarantee its availability for that demand.
ML engineer developing custom LLM, model architecture diagrams on screens, technical deep work environment.
INVENTORY COMMITMENT

What is Order Reservation?

The foundational mechanism that guarantees a specific unit of supply is exclusively allocated to a specific customer demand, preventing double-selling.

Order Reservation is the transactional act of creating a hard or soft link between a specific quantity of on-hand or inbound inventory and a customer order, effectively removing that supply from the available-to-promise pool. This process guarantees that the allocated stock is exclusively set aside for that specific demand, preventing it from being sold to another customer during the fulfillment window.

The reservation can be a hard reservation, which physically prevents the allocation of that inventory to any other demand, or a soft reservation, which serves as a tentative hold that may be overridden by higher-priority orders based on predefined business rules. This mechanism is the critical execution step that follows the Available-to-Promise (ATP) check, converting a theoretical promise into a concrete, traceable commitment within the Order Promising Engine.

INVENTORY COMMITMENT MECHANISMS

Key Characteristics of Order Reservation

Order reservation is the foundational act of creating a hard or soft link between a specific quantity of on-hand or inbound inventory and a customer order, guaranteeing its availability for that demand.

01

Hard vs. Soft Reservation

A hard reservation creates a firm, irrevocable link between a specific inventory lot and a sales order, physically preventing it from being allocated elsewhere. A soft reservation creates a tentative claim that can be overridden by higher-priority demands.

  • Hard Reservation: Used for VIP customers, export orders, or regulated industries
  • Soft Reservation: Used for standard B2B orders where reallocation is acceptable
  • The distinction is critical for Available-to-Promise (ATP) calculations, as hard reservations immediately reduce the ATP quantity
02

Reservation Against Inbound Supply

Reservations are not limited to on-hand stock. Modern systems allow linking an order to in-transit inventory, purchase orders, or planned production receipts.

  • A reservation against an Advanced Shipment Notice (ASN) guarantees the goods upon arrival
  • This enables back-to-back fulfillment where a sales order is directly pegged to a purchase order
  • The reservation remains in a pending state until the inbound receipt is physically posted, at which point it converts to a hard reservation against on-hand stock
03

Reservation Hierarchy and Precedence

When supply is constrained, a reservation hierarchy determines which demands get priority. This is a configurable set of rules evaluated in sequence.

  • Customer Tier: Platinum accounts reserve before Silver
  • Order Type: Service parts orders may outrank standard wholesale
  • Commitment Date: Earliest requested delivery date often wins
  • The hierarchy prevents overselling and ensures allocation management policies are enforced automatically during the reservation process
04

Reservation Lifecycle and Expiry

A reservation is not permanent. It has a defined lifecycle with configurable expiry rules to prevent stale inventory from being locked indefinitely.

  • Reservation Time Fence: A maximum number of days a soft reservation can exist before auto-releasing
  • Order Hold Integration: If an order is placed on credit hold, the reservation may be downgraded from hard to soft
  • Cancellation Workflow: When an order is cancelled, the reservation is immediately released and the inventory is returned to the ATP pool for re-promising
05

Batch and Serial-Level Granularity

In industries like pharmaceuticals and electronics, reservations operate at the batch or serial number level, not just the SKU level.

  • Shelf-Life ATP requires reserving a specific batch with sufficient remaining freshness for the customer
  • Serialized reservation ensures a specific, traceable unit is assigned to a specific order for warranty or regulatory compliance
  • This granularity feeds directly into demand pegging, creating an unbroken chain of custody from supply receipt to customer delivery
06

Integration with Warehouse Execution

The reservation is a logical commitment; it must be physically enforced on the warehouse floor. Integration with Warehouse Management Systems (WMS) is essential.

  • A hard reservation can trigger a directed putaway to a reserved location
  • During picking, the WMS validates that the operator picks the reserved lot or serial number
  • If a picker attempts to pick reserved inventory for a different order, the system generates a pick exception and blocks the transaction, maintaining the integrity of the original commitment
ORDER RESERVATION

Frequently Asked Questions

Clear, technical answers to the most common questions about the mechanics of inventory reservation, soft vs. hard allocation, and its role in the order-to-cash cycle.

An order reservation is the act of creating a hard or soft link between a specific quantity of on-hand or inbound inventory and a customer order to guarantee its availability for that demand. The process works by decrementing the Available-to-Promise (ATP) quantity in real-time. When an order is entered, the Order Promising Engine executes a reservation algorithm that first checks sourcing rules to identify the optimal warehouse. It then performs an ATP Netting calculation, subtracting the reserved quantity from the projected available balance. A hard reservation physically blocks the stock from being allocated to other orders, while a soft reservation marks it as tentatively assigned, allowing an override if a higher-priority order requires it. This link is maintained in the ERP system until the goods are picked and shipped, ensuring deterministic fulfillment.

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.