Joint Replenishment

The fundamental cost decomposition that drives the joint replenishment logic:

• Major Ordering Cost (A): A fixed administrative, transportation, or setup cost incurred every time an order is placed, regardless of the number of line items. This is the cost being amortized.
• Minor Item Cost (a_i): An incremental cost specific to adding a particular SKU to the order, such as line-item picking, labeling, or receiving inspection.
• Holding Cost (h_i): The per-unit carrying cost that penalizes large batch sizes, creating the optimization tension against the major cost.

The algorithm minimizes the sum of these three cost components across the entire group.

A widely implemented heuristic for joint replenishment that defines two inventory thresholds per SKU:

• Must-Order Point (s_i): The standard reorder point. When any single item's inventory position drops to or below this level, it triggers a group replenishment event.
• Can-Order Point (c_i): A higher threshold. Once a group order is triggered by another item hitting its must-order point, any other item currently at or below its can-order point is opportunistically included in the batch.

This prevents the system from ordering items with healthy stock levels while capturing near-reorder items to avoid a separate future major cost.

The rigorous mathematical underpinning of joint replenishment relies on renewal theory to model the stochastic intervals between orders:

• The system is modeled as a renewal process where the inter-arrival time between group replenishment events is a random variable determined by the first item to hit its reorder point.
• The objective function minimizes the long-run average cost per unit time, expressed as the ratio of expected cost per cycle to expected cycle length.
• This framework handles the complex dependency where the demand processes of individual items are independent, but their replenishment epochs are coupled through the shared major cost.

Two distinct modeling approaches exist for coordinating replenishment:

• Deterministic Joint Replenishment (JRP): Assumes constant, known demand rates. The classic solution finds a base cycle time T, with each item i ordered every k_i * T intervals, where k_i is an integer multiplier. This creates a nested, periodic schedule.
• Stochastic Joint Replenishment: Addresses random demand variability. The can-order (s, c, S) policy is the dominant heuristic, where S is the order-up-to level. Items are raised to S when ordered, and the system state is continuously reviewed.

The stochastic variant is computationally NP-hard, making heuristic policies essential for practical deployment.

Joint replenishment creates a direct financial incentive for supplier rationalization:

• Amortization Effect: As more SKUs are sourced from a single vendor, the major ordering cost A is spread across a larger base, reducing the cost allocation per item.
• Transportation Efficiency: Grouping items enables full truckload (FTL) or full container load (FCL) shipments, converting expensive less-than-truckload (LTL) shipments into lower per-unit freight costs.
• Vendor-Managed Inventory (VMI) Enabler: The supplier can autonomously execute joint replenishment on behalf of the buyer, optimizing their own production and delivery schedules while guaranteeing service levels.

This directly contradicts the procurement instinct to multi-source every item for price leverage.

Effective joint replenishment requires pre-processing to group items into logical families:

• Common Supplier: The primary grouping criterion. Items must share the same vendor and ship-from location.
• Shared Transportation Mode: Items must be compatible with the same freight method and routing.
• Similar Demand Velocity: Grouping a fast-moving A-item with a sporadic C-item can distort the cycle, forcing frequent orders of the slow mover. ABC-XYZ classification is applied first to ensure velocity alignment.
• Product Compatibility: Hazardous materials, temperature requirements, and physical dimensions constrain which items can share a purchase order.

Poor family definition leads to dysfunctional grouping where the major cost savings are eroded by excess holding costs on artificially accelerated slow movers.

The total annual cost of holding one unit in stock, expressed as a percentage of item value. It encompasses:

• Capital cost: The opportunity cost of frozen working capital
• Storage cost: Warehouse space, utilities, and handling
• Risk cost: Obsolescence, shrinkage, and insurance In joint replenishment models, carrying cost is the penalty that prevents over-ordering minor items just to amortize the major setup cost.

A collaborative strategy where the supplier assumes responsibility for monitoring the buyer's stock levels and autonomously generating replenishment orders. Joint replenishment logic is often embedded within VMI agreements, as the supplier is uniquely positioned to identify family grouping opportunities across multiple SKUs and consolidate shipments to reduce the major ordering cost for both parties.

A two-dimensional segmentation matrix that guides which items should be included in a joint replenishment group:

• ABC: Ranks items by annual consumption value (A = high value, C = low value)
• XYZ: Ranks items by demand predictability (X = stable, Z = erratic) Joint replenishment is most effective for AX and BX items—high-value, predictable SKUs where coordinated ordering yields significant carrying cost savings without introducing excessive variability risk.

The predetermined inventory level that triggers a replenishment order, calculated as expected demand during lead time plus safety stock. In a joint replenishment context, the reorder point of one item can trigger a family-wide review, where the system evaluates whether adding other items nearing their reorder points to the same order is economically justified by the shared major setup cost savings.

A supply chain distortion where small demand variations at the retail level cause amplified order oscillations upstream. Joint replenishment can inadvertently worsen this effect if batch ordering logic creates lumpy, synchronized demand signals that mask true consumption patterns. Advanced implementations mitigate this by applying order smoothing algorithms and sharing point-of-sale data with suppliers.