Replenishment Systems and Operating Policies

Introduction

Inventory replenishment is the set of decisions and routines used to determine when to order, how much to order, and how frequently to review inventory. These decisions translate demand and supply uncertainty into operating rules. A replenishment policy is therefore more than a formula: it is a control system linking inventory records, forecasts, suppliers, service objectives, order-processing capacity, and physical execution.

The central design choice is whether inventory is monitored continuously or at fixed intervals. Continuous-review systems respond whenever inventory reaches a defined trigger. Periodic-review systems respond only at scheduled review points. Reorder-point, order-up-to, min–max, two-bin, and base-stock policies are practical forms of these broader approaches.

A sound policy must distinguish physical inventory from inventory position, account for lead time and outstanding orders, define service expectations, and specify how parameters are maintained. Poorly maintained parameters can make an apparently sophisticated system behave worse than a simple visual control.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Distinguish continuous-review and periodic-review replenishment.
2. Calculate and interpret reorder points, order-up-to levels, min–max parameters, and base-stock quantities.
3. Select a review cadence and order trigger consistent with demand, lead time, risk, and operating cost.
4. Explain the roles of safety stock, inventory position, and demand during protection periods.
5. Design parameter-maintenance, exception-management, and governance routines.
6. Evaluate replenishment policies in operational and managerial contexts.

1. The Control Logic of Replenishment

A replenishment policy converts information into an order decision. At time t, the relevant state is not merely units on the shelf. A useful inventory-position identity is:

where:

• IP_t = inventory position in units;
• OH_t = on-hand inventory;
• OO_t = open or on-order inventory expected to arrive;
• BO_t = backorders or committed demand not yet fulfilled.

All terms must be measured in the same unit and for the same stock-keeping unit. In systems where reservations are not represented as backorders, committed allocations should be deducted through an equivalent adjustment. Using on-hand inventory alone can create duplicate orders because goods already ordered are ignored.

A policy normally has four elements:

The policy should also specify treatment of minimum order quantities, order multiples, supplier calendars, shortages, substitutions, returns, and obsolete stock.

2. Continuous-Review Systems

In a continuous-review, or Q-system, inventory position is monitored whenever a transaction occurs or through frequent system updates. An order is released when inventory position reaches or falls below a reorder point. The traditional form is:

Here, r is the reorder point and Q is the order quantity. The policy is often called an (r, Q) policy.

2.1 Reorder point under deterministic demand

If demand is constant at d units per time period and replenishment lead time is L periods, demand during lead time is dL. With safety stock SS,

The units are consistent: units/time × time = units. This expression assumes that lead time and demand are represented on compatible time scales and that the replenishment order is available at the end of lead time.

2.2 Reorder point under uncertainty

When demand during lead time is random, the reorder point is commonly written as:

where μ_LT is expected demand during lead time. If demand and lead time are sufficiently stable and demand during lead time is approximated by a normal distribution,

where σ_LT is the standard deviation of lead-time demand and z is the standard-normal service factor. The term zσ_LT is safety stock.

This formula is a modelling approximation, not a universal law. It assumes that the selected service measure corresponds to a probability of not stocking out during the replenishment period, that the distributional approximation is reasonable, and that the parameter estimates are representative. A cycle-service target of 95%, for example, uses approximately z = 1.645 under a normal assumption.

If demand per period has standard deviation σ_d and lead time is a fixed integer L, independent-period demand gives:

If both demand and lead time vary, a commonly used approximation is:

where d̄ is mean demand per period and σ_L is the standard deviation of lead time in periods. This approximation assumes independence between demand and lead time and ignores covariance. If they are correlated, a covariance term is required; omitting it can materially understate or overstate risk.

Worked example: continuous review

A distributor sells 40 units per day. Supplier lead time is five days. Daily demand standard deviation is 12 units, lead time is fixed, and the desired cycle service level is 95%.

Mean lead-time demand is:

Lead-time standard deviation is:

Using z = 1.645, safety stock is:

Thus:

The operational rule is to trigger replenishment when inventory position is approximately 244 units or less. Rounding upward is prudent when fractional units cannot be ordered and service protection is more important than a small increase in inventory.

3. Periodic-Review Systems

In a periodic-review, or P-system, inventory is reviewed at intervals of P time units. An order is not necessarily placed when inventory crosses a threshold between reviews. At each review, the system commonly raises inventory position to an order-up-to level S.

The policy is often called an (S, P) policy. The order quantity varies, while the review interval is fixed.

The protection period is the time from one order opportunity until the next replenishment can arrive:

where L is lead time. Under deterministic demand and safety stock:

Under uncertain demand:

where μ_T and σ_T describe demand over the protection period T. For fixed lead time and independent-period demand, σ_T = σ_d√(P+L).

Periodic review requires more safety stock than an equivalent continuous-review system when P > 0, because demand can occur after the last review but before the next order is placed. The benefit is administrative simplicity: orders can be consolidated, supplier visits can follow a route schedule, and buyers can process a portfolio at planned times.

Worked example: periodic review

Mean demand is 100 units per week, weekly standard deviation is 25 units, lead time is two weeks, review interval is one week, and the target cycle service level is 95%.

The protection period is three weeks. Therefore:

Safety stock is approximately:

The order-up-to level is:

At a review, if inventory position is 148 units, the order is:

This calculation presumes that all open orders are included in inventory position and that the order-up-to level has not been constrained by a minimum order quantity or an order multiple.

4. Reorder-Point, Order-Up-to, Min–Max, Two-Bin, and Base-Stock Policies

4.1 Reorder-point policy

A reorder-point policy separates the trigger from quantity. In its basic form, IP ≤ r triggers a fixed or economically determined order. It is appropriate where demand occurs continuously, stockouts are costly, and transaction-level data are available. Fixed Q may reflect economic order quantity, supplier packaging, transport economics, or a contractual minimum.

The classical economic order quantity under constant demand is:

where D is annual demand, S is ordering cost per order, and H is annual holding cost per unit. This result assumes constant demand, instantaneous replenishment, no shortages, constant unit price, and no quantity constraints. It is a cost benchmark, not a complete replenishment policy; the reorder point still requires lead-time and service analysis.

4.2 Order-up-to policy

An order-up-to system orders the difference between a target and current inventory position. It is particularly useful in periodic review, supplier schedules, and multi-item ordering. Its danger is overreaction: if open orders are omitted or forecasts change sharply, the system may place excessive orders.

4.3 Min–max policy

A min–max policy uses a lower trigger, min, and an upper target, max. At review or when the lower level is reached, the order raises inventory position toward the maximum:

The terms are often used loosely. In some organizations, “min” means a reorder point; in others, it means physical on-hand inventory. The implementation document must define which interpretation applies. A coherent design commonly sets min to expected demand through lead time plus safety stock, and max to demand through lead time plus a replenishment or review cycle plus safety stock.

4.4 Two-bin system

A two-bin system divides usable stock into two physical containers. The first bin supplies normal demand; when it is emptied, the second bin is used and the empty-bin signal initiates replenishment. The quantity in one bin is normally sized to cover expected demand during replenishment lead time plus protection against variability.

Two-bin control is transparent and robust for low-value, stable, standardized items. It is weak for expensive, highly variable, serialized, perishable, or space-constrained items. It also depends on disciplined scanning, clear bin labels, and correct container sizing.

4.5 Base-stock system

A base-stock policy attempts to maintain a target level of inventory position, often close to one unit for each unit of expected demand plus protection stock. Each demand transaction causes an equivalent replenishment signal, subject to batching and supplier constraints. In a simplified form:

and the replenishment quantity restores the position toward S_t. Base-stock logic is common in spare parts, component pipelines, and pull systems. It can create frequent orders and therefore requires transaction accuracy and supplier responsiveness.

5. Choosing Review Cadence and Parameters

Review cadence should reflect demand frequency, lead-time risk, ordering cost, operational capacity, and the value of rapid response. A shorter cadence generally reduces response delay but increases processing effort. A longer cadence supports consolidation but increases protection-period demand and usually safety stock.

Parameter design should use a hierarchy of evidence:

1. Clean historical demand, excluding known anomalies where appropriate.
2. Current lead-time observations rather than outdated supplier promises.
3. Forecasts for intermittent, seasonal, promotional, or lifecycle-dependent demand.
4. Explicit service objectives and shortage consequences.
5. Physical and commercial constraints such as minimum order quantities, shelf life, and order multiples.

Safety stock should not be used to conceal poor master data, unreliable suppliers, forecast bias, or process defects. Increasing it may improve availability temporarily while increasing working capital and obsolescence.

6. Parameter Maintenance and Governance

A replenishment policy is only as good as its maintenance process. Governance should define ownership, review frequency, approval thresholds, and audit trails.

Important controls include:

• Lead-time monitoring: estimate actual order-to-receipt time, including variability, not merely quoted lead time.
• Demand review: identify trend, seasonality, promotions, product introductions, and one-time events.
• Exception management: flag unusually high orders, negative inventory, repeated expedites, excess stock, and imminent stockouts.
• Data integrity: reconcile receipts, cancellations, returns, reservations, and open purchase orders.
• Constraint validation: enforce supplier minimums, multiples, shelf-life limits, capacity, and transportation calendars.
• Segmentation: apply different policies to critical, expensive, volatile, and routine items.
• Change control: record who changed a parameter, why, when, and with what evidence.

Useful performance measures include fill rate, cycle service level, stockout frequency, inventory turns, days of supply, excess and obsolete inventory, order frequency, forecast bias, and supplier lead-time adherence. These metrics should be interpreted together. High turns with poor service may indicate understocking; high service with excessive obsolescence may indicate overstocking.

The cycle-service probability for a period is not the same as fill rate. Cycle service asks whether any stockout occurred during a replenishment cycle. Fill rate asks what fraction of demand was supplied immediately. A policy can achieve a high cycle service level while still having a low fill rate when shortage quantities are large.

7. Integrated Illustrative Case: Northstar Medical Components

This is a fictional illustrative case.

Background

Northstar assembles diagnostic instruments. A small electronic module is purchased from a single supplier. Average demand is 60 units per week, weekly demand standard deviation is 18 units, and observed lead time averages three weeks with standard deviation 0.5 weeks. Demand and lead time are treated as independent for initial planning. The module is valuable and a shortage can stop final assembly.

Problem

The existing periodic policy reviews every four weeks and uses an outdated two-week lead-time master-data value. Buyers frequently expedite orders, while inventory is simultaneously high for several slow-moving variants.

Analysis

For variable demand and lead time, an approximate standard deviation of lead-time demand is:

Substituting L = 3, σ_d = 18, d̄ = 60, and σ_L = 0.5:

Mean lead-time demand is 180 units. At a 95% cycle-service target, safety stock is approximately 1.645 × 43.27 ≈ 71 units, producing a continuous-review reorder point of approximately 251 units.

However, the current four-week review interval creates a protection period of seven weeks, not three. If the organization retains periodic review, it must size the order-up-to level over seven weeks and use the corresponding uncertainty. That would materially increase stock. Because the item is critical and transactions are electronically visible, Northstar selects continuous review with a reorder point near 251 units. Orders are placed in the supplier’s 20-unit multiple, with a managerial review when projected demand, engineering changes, or supplier performance changes.

Outcome

The new policy reduces unnecessary review-period protection stock while improving responsiveness. The company also establishes a monthly parameter review, weekly supplier lead-time monitoring, and an exception dashboard. For low-value fasteners, it retains two-bin control rather than extending the expensive electronic policy to every item.

Learning points

1. A review interval is part of the risk exposure and cannot be ignored.
2. Actual lead-time variation can be as important as demand variation.
3. Policy selection should be segmented by item economics and criticality.
4. A calculated parameter requires governance, not automatic trust.

8. Technology, AI, and Contemporary Practice

Enterprise resource-planning and warehouse-management systems can calculate inventory position, generate planned orders, and enforce order multiples. Radio-frequency identification, barcode scanning, and Internet-of-Things sensors improve transaction timeliness. These technologies do not remove the need for policy design; they make errors propagate faster when master data are wrong.

Machine-learning forecasting can improve demand estimates for products with sufficient history and explanatory variables. It is less reliable for intermittent demand, structural breaks, new products, and highly censored demand. AI-generated recommendations should therefore be subject to human review, version control, explainability requirements, and monitoring for systematic bias. A model that recommends higher stock for customers or regions based on protected or inappropriate proxies creates governance and fairness risks.

Contemporary replenishment design increasingly considers carbon and resilience alongside cost and service. Larger, less frequent orders may reduce transport emissions but increase inventory and obsolescence. Dual sourcing and regional buffers may improve resilience but raise cost. These are multi-objective decisions rather than purely mathematical minima.

9. Limitations and Professional Judgment

No single policy is universally superior. Normal approximations may fail under intermittent or highly skewed demand. Independent-demand assumptions may fail when many items are driven by a common project. Lead times may be censored because late orders are expedited. Forecasts may be biased by stockouts, meaning observed demand understates true demand. Fixed service factors may also obscure different shortage costs across customers.

Professional judgment is required to decide whether to use empirical quantiles, simulation, negative-binomial or Croston-type methods for intermittent demand, scenario analysis for disruptions, or robust buffers for strategic components. The policy should be tested through historical back-testing and, where possible, discrete-event simulation. Evaluation should compare service, inventory, order workload, and resilience—not merely average stock.

Chapter Summary

Replenishment systems control the timing and quantity of inventory orders. Continuous-review systems trigger orders from inventory position and are effective when transactions are frequent and rapid response matters. Periodic-review systems order at scheduled intervals and raise inventory position toward an order-up-to level, but their protection period includes both review interval and lead time. Reorder-point, order-up-to, min–max, two-bin, and base-stock systems are related designs with different information, simplicity, and responsiveness requirements.

Core calculations use demand during the relevant protection period plus safety stock. Correct design requires consistent units, realistic demand and lead-time data, explicit assumptions, and recognition of commercial constraints. Parameter maintenance, exception management, segmentation, and auditability are as important as the initial formula. Technology and AI can strengthen replenishment, but neither substitutes for data governance, human accountability, and critical evaluation.

Key Terms

Inventory position: On-hand plus on-order inventory minus backorders or equivalent commitments.

Continuous review: Monitoring inventory state continuously or transaction by transaction.

Periodic review: Reviewing inventory at predetermined time intervals.

Reorder point: Inventory-position level that triggers replenishment.

Order-up-to level: Target inventory position to which a system replenishes.

Min–max policy: Policy using a lower trigger and an upper target.

Two-bin system: Physical two-container replenishment method using an empty-bin signal.

Base-stock policy: Policy that restores inventory position toward a target after demand.

Lead time: Time between placing an order and its usable receipt.

Protection period: Time during which demand must be covered before replenishment becomes available.

Safety stock: Inventory held to protect against uncertainty.

Cycle service level: Probability of experiencing no stockout during a replenishment cycle.

Fill rate: Fraction of demand supplied immediately from available stock.

Order quantity: Units released in a replenishment order.

Inventory position: The control-state quantity used to avoid ignoring open supply.

Parameter maintenance: Ongoing review and adjustment of policy inputs.

Order multiple: Required increment in which a supplier accepts orders.

Master data: System data governing items, suppliers, lead times, and constraints.

Exception management: Investigation of conditions requiring attention outside normal policy operation.

Review Questions

Short Answer Questions

1. What is the difference between on-hand inventory and inventory position?
2. Why does a periodic-review system generally require more protection stock than a continuous-review system?
3. Define a reorder point and an order-up-to level.
4. Identify two advantages and two limitations of a two-bin system.
5. Why should lead-time variability be included in replenishment design?

Long Answer Questions

1. Explain the structure, assumptions, and managerial uses of an (r,Q) continuous-review policy.
2. Compare continuous-review and periodic-review systems for a critical purchased component.
3. Discuss how safety stock should be interpreted and governed in an organization.
4. Explain why min–max and order-up-to policies can produce excessive orders when open orders are omitted.
5. Evaluate the role of AI forecasting in replenishment governance.

Analytical / Application Questions

1. Daily demand is 30 units, fixed lead time is six days, daily standard deviation is 8 units, and the target cycle service level is 95%. Calculate the reorder point using a normal approximation and explain the assumptions.
2. A firm reviews an item every two weeks and has a lead time of three weeks. Explain how changing the review interval to one week affects protection-period demand, safety stock, order workload, and responsiveness.
3. Design a replenishment policy for a low-value, stable-demand consumable and a high-value, intermittent-demand spare part. Justify the review method, trigger, quantity rule, and parameter-maintenance process for each.

Multiple Choice Questions with Answers

Question 1

Which quantity is normally used to determine whether a continuous-review replenishment trigger has been reached?

A. On-hand inventory only

B. Inventory position

C. Annual demand

D. Supplier capacity

Correct Answer: B. Inventory position

Question 2

A product has mean demand of 40 units per day and a fixed lead time of 5 days. Ignoring safety stock, what is the reorder point?

A. 8 units

B. 45 units

C. 200 units

D. 400 units

Correct Answer: C. 200 units

Question 3

In a periodic-review system, the protection period normally equals:

A. Lead time minus review interval

B. Review interval plus lead time

C. Order quantity plus lead time

D. Safety stock plus order quantity

Correct Answer: B. Review interval plus lead time

Question 4

At a periodic review, inventory position is 148 units and the order-up-to level is 372 units. What order quantity is indicated before applying constraints?

A. 148 units

B. 224 units

C. 372 units

D. 520 units

Correct Answer: B. 224 units

Question 5

Which policy uses a lower trigger and an upper target?

A. Min–max policy

B. Base-stock policy only

C. Economic order quantity only

D. Supplier-managed inventory only

Correct Answer: A. Min–max policy

Question 6

For fixed lead time and independent-period demand, lead-time demand standard deviation is commonly calculated as:

A. σ_dL

B. σ_d/L

C. σ_d√L

D. σ_d + L

Correct Answer: C. σ_d√L

Question 7

A two-bin system is generally most suitable for:

A. High-value intermittent spare parts

B. Low-value stable consumables

C. Perishable products with uncertain shelf life

D. Serialized products requiring individual traceability

Correct Answer: B. Low-value stable consumables

Question 8

Which statement best describes cycle service level?

A. The proportion of demand filled immediately

B. The probability of no stockout during a replenishment cycle

C. The percentage of suppliers delivering on time

D. The fraction of inventory that turns annually

Correct Answer: B. The probability of no stockout during a replenishment cycle

Question 9

If annual demand is 10,000 units, ordering cost is 50 currency units per order, and annual holding cost is 2 currency units per unit, the classical EOQ is approximately:

A. 500 units

B. 707 units

C. 1,000 units

D. 2,500 units

Correct Answer: B. 707 units

Question 10

Which is the strongest reason to include open purchase orders when calculating inventory position?

A. To increase reported demand

B. To avoid counting expected supply as unavailable and placing duplicate orders

C. To eliminate safety stock

D. To reduce supplier lead time

Correct Answer: B. To avoid counting expected supply as unavailable and placing duplicate orders