Inventory Review and Replenishment Systems

Introduction

Inventory review and replenishment systems determine when an organization should place an order and how much it should order. They convert demand information, lead-time knowledge, service objectives, and supply constraints into operating decisions. A sound system must balance two competing risks: insufficient inventory, which causes stockouts and disrupted operations, and excessive inventory, which ties up capital, consumes space, increases obsolescence, and may conceal process problems.

The principal replenishment architectures are continuous-review, periodic-review, min–max, two-bin, base-stock, and order-up-to systems. These are not mutually exclusive labels. For example, a min–max policy can be operated through continuous monitoring or through periodic reviews, while an order-up-to policy is usually associated with periodic review. The appropriate choice depends on demand characteristics, item criticality, lead-time variability, transaction technology, supplier responsiveness, and the cost of monitoring and ordering.

This chapter develops the logic of each system, presents the principal design equations, and evaluates their use in different operating contexts.

Learning Objectives

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

1. Distinguish inventory position from on-hand inventory and explain why replenishment decisions use inventory position.
2. Design continuous-review and periodic-review systems using reorder points, order quantities, safety stock, and order-up-to levels.
3. Explain and apply min–max, two-bin, base-stock, and order-up-to policies.
4. Select a replenishment rule according to demand, lead time, service requirements, cost, and operational context.
5. Diagnose implementation risks, including parameter error, nervousness, data-quality problems, and inappropriate automation.
6. Evaluate replenishment policies using service, inventory, cost, and process-performance measures.

1. The Replenishment Decision

A replenishment policy specifies at least four decisions:

• Review trigger: when inventory information is assessed.
• Order trigger: the condition that causes an order or production release.
• Order quantity: the amount replenished.
• Protection level: the inventory intended to cover expected demand and uncertainty.

The central stock-status measure is inventory position:

where:

• IP = inventory position, in units;
• OH = on-hand inventory, physically available or usable;
• OO = on-order inventory, already released but not received;
• BO = backorders or committed shortages.

Inventory position is superior to on-hand inventory for replenishment control because an outstanding purchase order already provides future supply, while a backorder represents a requirement not covered by available stock. If a firm used only on-hand inventory, it could place duplicate orders while ignoring supply already in transit.

A related measure is inventory position net of allocations, depending on the organization’s information system. The exact definition must be standardized; otherwise, different sites may interpret the same policy differently.

Demand, lead time, and uncertainty

Let demand per unit of time have mean μ_D and standard deviation σ_D. Let replenishment lead time have mean μ_L and standard deviation σ_L. The demand expected during lead time is approximately:

where D_L is lead-time demand. This multiplication is valid when demand rate and lead time are represented consistently—for example, units per day multiplied by days—and when the mean demand rate is not systematically dependent on lead time.

If demand and lead time are independent, the variance of lead-time demand is:

This expression assumes that demand in successive periods is independent or that σ_D² appropriately represents the variance per unit time. It also assumes that lead time is measured in the same time units as demand. A more general formulation must include covariance when demand and lead time are dependent:

under the common random-sum approximation with independent period demand and random lead time. In practice, organizations often estimate lead-time-demand variability directly from historical observations, which can be preferable when seasonality, autocorrelation, promotions, or supplier disruption make simple assumptions unreliable.

2. Continuous-Review Systems

In a continuous-review system, inventory position is monitored after each transaction or at sufficiently frequent intervals. An order is triggered when the inventory position reaches a specified reorder point. The basic policy is written as (Q, r):

Here, Q is the order quantity and r is the reorder point. The notation describes a rule, not necessarily a fixed economic order quantity. Q may be fixed, variable, supplier-constrained, or determined by an optimization routine.

Reorder point

A common reorder-point equation is:

where SS is safety stock. The first term covers expected lead-time demand; safety stock protects against demand and lead-time uncertainty.

For a cycle-service target, a frequently used approximation is:

where z is the standard-normal quantile corresponding to the desired probability of no stockout during a replenishment cycle, and σ_DL is the standard deviation of lead-time demand. Thus:

Typical standard-normal values are approximately z = 1.28 for 90%, z = 1.645 for 95%, and z = 2.33 for 99% cycle service. These values should not be interpreted as fill-rate guarantees. Cycle service is the probability that an individual replenishment cycle experiences no stockout; fill rate is the proportion of demand supplied immediately. They are different performance measures.

The safety-stock formula assumes approximately normal lead-time demand, stable parameters, and a service objective expressed as a probability. It is less reliable for intermittent demand, very low-volume items, severe seasonality, or highly skewed demand. In those cases, empirical quantiles, Croston-type methods, simulation, or risk-based inventory models may be more suitable.

Worked example: continuous review

A distributor sells 40 units per day. Daily demand standard deviation is 8 units, lead time is 6 days, and lead time is constant. Management wants approximately 95% cycle service. Assume independent daily demand.

Expected lead-time demand is:

μ_D μ_L = 40 × 6 = 240 units.

Because lead time is constant:

σ_DL = σ_D√L = 8√6 ≈ 19.6 units.

At 95% cycle service, z = 1.645, so:

SS = 1.645 × 19.6 ≈ 32.3 units.

Rounding upward, the reorder point is:

r = 240 + 33 = 273 units.

The rule is therefore: monitor inventory position and place an order whenever IP ≤ 273. If the selected order quantity is 400 units, the policy is (Q = 400, r = 273). The order quantity does not alter the basic calculation of the reorder point, although it affects average inventory and ordering cost.

Strengths and limitations

Continuous review responds quickly to demand and is especially valuable for expensive, critical, or fast-moving items. It can achieve tight control with relatively low safety stock when transactions are captured accurately. Its disadvantages include the need for reliable real-time records, system integration, transaction discipline, and potentially frequent orders. In low-value environments, the monitoring cost may exceed the inventory benefit.

3. Periodic-Review Systems

In a periodic-review system, inventory position is reviewed at fixed intervals, such as weekly or monthly. An order is placed at each review if required. The standard order-up-to form is (P, S):

where:

• P = review period;
• S = order-up-to level;
• IP_t = inventory position at review time t;
• q_t = order quantity.

The order-up-to level must protect demand during the review period plus the replenishment lead time, because the next opportunity to order occurs only after P time units. If demand and lead time are stable:

The safety stock protects the combined protection interval P + L. If demand variability is measured per unit time and lead time is constant, then:

and:

For variable lead time, the variance formula for the protection period should reflect both review-period demand and lead-time uncertainty. A practical approximation is to calculate the distribution of demand over P + L, using an appropriate forecast-error model rather than simply adding standard deviations.

Worked example: periodic review

A wholesaler reviews a product every 7 days. Average daily demand is 30 units, daily standard deviation is 6 units, and lead time is 5 days. Lead time is constant. The target is 95% cycle service.

The protection interval is:

P + L = 7 + 5 = 12 days.

Expected demand during the protection interval is:

30 × 12 = 360 units.

The standard deviation is:

σ_P+L = 6√12 ≈ 20.8 units.

Safety stock is:

SS = 1.645 × 20.8 ≈ 34.2 units.

Thus, the order-up-to level is approximately:

S = 360 + 35 = 395 units.

If the inventory position at the review is 220 units, the order is:

q = 395 − 220 = 175 units.

The next review will occur seven days later, whether or not demand has changed. This periodic schedule is convenient for route deliveries, consolidated purchasing, and supplier agreements with fixed ordering days.

Strengths and limitations

Periodic review simplifies administration and permits order consolidation. It is useful when suppliers accept orders only on scheduled days, when a salesperson visits on a route, or when many items must be reviewed together. However, a stockout can occur between reviews, and the protection requirement is larger than in an otherwise comparable continuous-review system because it includes review-period demand. The system is also vulnerable to demand surges shortly after a review.

4. Min–Max Replenishment

A min–max policy defines a lower threshold and an upper target. When inventory position reaches or falls below the minimum, the system orders enough to restore inventory position toward the maximum:

with an order trigger:

The minimum is often equivalent to a reorder point, while the maximum is an order-up-to level. In a continuous-review implementation, the policy resembles a variable-order-quantity (s, S) system. In periodic review, the minimum is checked only at review times.

A practical design is:

• S_min: expected demand during lead time plus safety stock;
• S_max: expected demand during lead time plus review-cycle demand plus safety stock, or a specified economic replenishment level.

The difference S_max − S_min should be large enough to avoid uneconomic small orders but not so large that average inventory becomes excessive. Min–max systems are widely used in enterprise resource planning because they are easy to explain and maintain, but their effectiveness depends heavily on whether the minimum and maximum are periodically recalibrated.

5. Two-Bin Systems

A two-bin system divides stock into two physical or virtual quantities. The first bin supplies normal demand. When it is exhausted, the second bin is opened and signals replenishment. The quantity in the second bin is intended to cover demand during replenishment lead time, usually with some protection against uncertainty.

The two-bin method is a visual and operational form of a reorder-point system. It is effective for low-value, standardized consumables such as fasteners, packaging materials, maintenance supplies, and stationery. It reduces transaction effort because an empty bin is an immediate signal. Kanban cards, barcode scans, and electronic shelf labels are digital equivalents.

Its limitations are important. The method may be inaccurate where demand is highly variable, items are substituted, stock is shared across locations, or the second bin is used before replenishment is recorded. It is not inherently a two-bin “quantity formula”; bin sizes still require analysis of demand, lead time, service requirements, and container capacity.

6. Base-Stock Policies

A base-stock policy attempts to maintain a specified inventory position, commonly one unit or one replenishment cycle above expected requirements. In its simplest form, every unit consumed generates an equivalent replenishment order:

where D_t is demand or usage during period t, subject to order-size, batching, and supplier constraints.

More generally, a base-stock level B is maintained:

after each demand or review event. The base-stock level normally covers expected demand over lead time plus safety stock. Base-stock systems are particularly suitable for stable, high-service environments with short lead times, frequent replenishment, and reliable transaction data. Examples include hospital supplies, service parts, and production components supplied through electronic kanban.

A major advantage is responsiveness: inventory is replenished immediately after consumption. A major limitation is order amplification when demand is noisy or when suppliers impose minimum batches. The policy also requires dependable visibility of consumption and pipeline inventory.

7. Order-Up-To Policies and the Bullwhip Problem

An order-up-to policy replenishes inventory position to a target at every review. It is often expressed as:

when S_t ≥ IP_t, with no negative order. The target may be fixed or forecast-based. In a changing-demand environment, a useful target is:

A safer textual formulation is: the target equals forecast demand over the protection interval plus safety stock. If the forecast for period t is F_t and the protection interval contains n periods, then, under a simple constant forecast:

where SS_t is safety stock appropriate to forecast error and supply uncertainty. The target should not be calculated from raw demand alone when seasonality, promotions, or known future events exist.

Order-up-to systems can contribute to the bullwhip effect. If each supply-chain stage reacts to noisy demand forecasts, inventory deficits, and changing lead-time assumptions, orders may fluctuate more than final customer demand. Stabilizing practices include sharing point-of-sale data, using consistent forecasts, reducing lead times, limiting unnecessary forecast overrides, and distinguishing genuine demand changes from temporary inventory corrections.

8. Choosing Among Replenishment Policies

No single system is universally superior. Selection should consider the following dimensions.

The decision should also account for the total cost of ownership. Relevant costs include purchase price, ordering or transaction cost, holding cost, shortage cost, obsolescence, expediting, handling, and the value of management attention. A policy that minimizes average inventory may be poor if it creates unacceptable shortages in a critical operating process.

Service measures

A policy should be evaluated using measures that match the business objective:

• Cycle service level: percentage of replenishment cycles without a stockout.
• Fill rate: percentage of demand supplied immediately from available stock.
• Ready rate: proportion of time an item is available, often used in service parts.
• Backorder frequency and duration: operational impact of shortages.
• Inventory turns: annual usage divided by average inventory value, interpreted carefully across different item classes.
• Average inventory and days of supply: capital and coverage indicators.
• Forecast error and parameter bias: drivers of poor replenishment decisions.

Service metrics should be segmented by item criticality and customer promise. A single aggregate service percentage can hide severe shortages for strategically important items.

9. Implementation Architecture

A replenishment policy becomes effective only when supported by disciplined processes.

Data governance

Item master data should include unit of measure, pack size, supplier, minimum order quantity, order multiple, lead time, calendar, shelf life, and substitution rules. Lead time should distinguish quoted, planned, actual, and reliable lead time. If a supplier’s nominal lead time is five days but actual delivery ranges from four to twelve, using five without protection will create systematic shortages.

Parameter governance

Parameters should be reviewed according to volatility and importance. A useful segmentation combines annual usage value, demand variability, criticality, lead-time risk, and obsolescence exposure. High-risk items require more frequent review and human oversight. Low-risk items can use standardized rules.

Transaction discipline

Receipts, issues, returns, scrapping, transfers, and backorders must be recorded promptly. Cycle counting should focus on items with high value, high movement, or high operational consequence. An apparently sophisticated algorithm cannot compensate for inventory records that are materially wrong.

Exception management

Automation should generate exceptions rather than merely generate orders. Examples include an order quantity below supplier minimum, demand outside historical limits, a lead time exceeding its control threshold, a negative inventory position, or a forecast change that materially alters the target. Human review is most valuable for exceptions, not for routine transactions.

10. Technology, Analytics, and Artificial Intelligence

Enterprise resource planning, warehouse-management, and supplier-portal systems can support real-time inventory position, automated triggers, electronic purchase orders, and delivery confirmation. Radio-frequency identification, barcode scanning, and Internet-of-Things sensors improve visibility where physical movements are frequent or difficult to record manually.

Analytics can estimate demand distributions, lead-time distributions, seasonality, forecast error, and service-cost trade-offs. Machine-learning models may improve forecasts when external drivers—weather, promotions, price, project schedules, or equipment condition—are informative. However, better forecasts do not automatically produce better replenishment. The forecast must be translated into a policy with correct lead-time, service, capacity, and lot-size assumptions.

AI-based recommendations require governance. Users should be able to identify the data used, the objective optimized, and the reason for an exceptional recommendation. Models should be monitored for drift, systematic bias, data leakage, and unstable behavior during disruptions. Accountability for a stockout or excessive order should remain assigned to an accountable process owner rather than being obscured by the phrase “the algorithm decided.”

11. Illustrative Case: Meridian Medical Supplies

Background

Meridian Medical Supplies distributes sterile procedure kits to regional hospitals. The kits are expensive, have a fixed shelf life, and are critical during emergency procedures. Demand is moderately variable, suppliers deliver twice weekly, and purchase orders are transmitted electronically.

Problem

Meridian used a monthly periodic-review rule for all items. The rule reduced purchasing workload but created two problems: emergency stockouts on high-use kits and expiry on slow-moving kits. Managers also calculated replenishment from on-hand inventory rather than inventory position, so open purchase orders were sometimes duplicated.

Analysis

The organization segmented items by criticality, demand velocity, and expiry risk. High-criticality kits were assigned continuous review. Their reorder points were based on expected demand during actual supplier lead time plus safety stock, with a service target higher than that of noncritical items. Medium-use kits were assigned a twice-weekly order-up-to system aligned with supplier delivery days. Slow-moving kits used lower target levels and explicit expiry controls.

Meridian also introduced a common inventory-position definition:

IP = on-hand + confirmed on-order − backorders − allocated stock.

The additional allocation term prevented stock reserved for known hospital orders from being treated as available. A dashboard displayed stockout risk, days to expiry, overdue orders, and orders exceeding normal quantities.

Approach and outcome

The company piloted the new policies in one distribution center. Continuous review was implemented through barcode transactions at picking and receiving. The periodic system was retained for items where delivery consolidation was economically important. Supplier lead-time observations replaced quoted lead times in parameter calculations.

The pilot reduced emergency purchases and improved availability of critical kits. It also exposed a data problem: several “slow-moving” items were actually being issued under substitute item codes. Correcting item master and substitution data was necessary before policy parameters could be trusted. The case illustrates that policy design, data governance, and operating discipline are inseparable.

Learning points

1. Policy choice should reflect item criticality and physical characteristics, not administrative convenience alone.
2. Inventory position must include the supply pipeline and relevant commitments.
3. Expiry risk may justify a lower target even when a conventional service calculation suggests more stock.
4. Pilot implementation and exception dashboards reveal problems that a theoretical design may miss.

12. Challenges and Limitations

Replenishment models are representations, not exact descriptions of reality. Demand may be nonstationary, lead time may depend on order size, and supplier capacity may collapse during disruptions. Safety stock based on a normal distribution can understate risk when demand is skewed or has occasional large spikes. Service-level targets may also be politically chosen rather than economically justified.

A further limitation is policy interaction. If several supply-chain members independently use order-up-to systems, local decisions may create upstream volatility. If sales promotions are not communicated, the replenishment system may interpret a temporary surge as a permanent change. If planners frequently override parameters without recording reasons, the organization loses the ability to learn from outcomes.

Resilience requires more than additional stock. Alternatives include dual sourcing, capacity reservations, shorter and more reliable lead times, postponement, substitution, supplier collaboration, and contingency transportation. The preferred response depends on the cost and duration of disruption and the criticality of the item.

Best-Practice Checklist

• Define inventory position consistently across all sites.
• Match review frequency to value, criticality, volatility, and lead-time risk.
• Separate expected demand coverage from safety stock.
• Use actual lead-time performance, not only supplier promises.
• Distinguish cycle service from fill rate.
• Include pack sizes, minimum order quantities, and shelf life in the policy.
• Recalculate parameters after structural demand or supply changes.
• Use exceptions and pilots to govern automation.
• Measure both service and total inventory cost.
• Document overrides and investigate repeated exceptions.

Chapter Summary

Inventory replenishment systems translate demand and supply uncertainty into ordering rules. Continuous-review systems trigger orders when inventory position reaches a reorder point and are appropriate when rapid response and close control justify monitoring effort. Periodic-review systems assess stock at fixed intervals and replenish to an order-up-to level; they are convenient for scheduled deliveries but require protection over both lead time and the review interval.

Min–max policies use lower and upper thresholds, while two-bin systems provide a simple visual reorder signal. Base-stock systems replenish consumption toward a specified level and work well where usage is visible and replenishment is frequent. Order-up-to policies restore inventory position to a target and are particularly useful in scheduled review environments, although poorly coordinated targets can intensify the bullwhip effect.

Policy design requires correct inventory-position data, appropriate demand and lead-time assumptions, service definitions, lot-size constraints, and governance. Technology and AI can improve visibility and estimation, but they do not remove the need for professional judgment, exception management, and accountability. The best policy is context dependent: it balances service, cost, risk, operational effort, shelf life, and resilience.

Key Terms

Inventory position: On-hand inventory plus on-order inventory minus backorders and relevant commitments.

Continuous review: A system in which inventory position is monitored continuously or after each transaction.

Periodic review: A system in which inventory is assessed at predetermined time intervals.

Reorder point: Inventory-position threshold that triggers replenishment.

Safety stock: Inventory held to protect against demand, lead-time, or forecast uncertainty.

Lead-time demand: Demand expected during the replenishment lead time.

Cycle service level: Probability that a replenishment cycle experiences no stockout.

Fill rate: Proportion of total demand supplied immediately from available inventory.

Order quantity: Amount released in a replenishment order.

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

Min–max policy: Rule that triggers replenishment at a minimum and restores stock toward a maximum.

Two-bin system: Visual or electronic system in which depletion of one container signals replenishment while another supplies demand.

Base-stock policy: Policy that replenishes usage or restores inventory position toward a specified base level.

Protection interval: Time during which inventory must cover demand before the next effective replenishment opportunity.

Demand variability: Dispersion or uncertainty in demand over time.

Lead-time variability: Dispersion in the time between order release and receipt.

Bullwhip effect: Amplification of order variability as demand information moves upstream.

Parameter governance: Processes for setting, reviewing, documenting, and changing replenishment inputs.

Exception management: Use of alerts and human review for conditions outside normal policy parameters.

Review Questions

Short Answer Questions

1. Why is inventory position generally more useful than on-hand inventory for triggering replenishment?
2. Distinguish cycle service level from fill rate.
3. What demand period must a periodic-review order-up-to level protect?
4. Explain the operational logic of a two-bin system.
5. Identify two limitations of normal-distribution safety-stock calculations.

Long Answer Questions

1. Compare continuous-review (Q,r) and periodic-review (P,S) systems in terms of information requirements, service performance, inventory, and operating cost.
2. Explain how min–max, base-stock, and order-up-to policies are related and how they differ in implementation.
3. Discuss the role of lead-time variability in determining reorder points and safety stock.
4. Develop a policy-selection framework for a distributor managing high-value, low-value, perishable, and intermittent-demand items.
5. Evaluate how data governance and exception management affect the success of automated replenishment.

Analytical / Application Questions

1. Daily demand averages 50 units with a standard deviation of 10 units. Constant lead time is 4 days and the target cycle service is 95%. Assuming independent daily demand, calculate lead-time demand, safety stock, and the reorder point.
2. A firm reviews stock every 10 days. Average demand is 25 units per day, daily standard deviation is 5 units, and constant lead time is 6 days. Using a 95% cycle-service target, calculate an approximate order-up-to level.
3. A hospital experiences stockouts under a monthly periodic-review policy but expiry under a continuous high-stock policy. Propose a segmented replenishment design and justify the review frequency, target levels, and governance controls for critical and slow-moving items.

Multiple Choice Questions with Answers

Question 1

Which expression correctly defines inventory position?

A. On-hand inventory plus backorders

B. On-hand inventory plus on-order inventory minus backorders

C. On-order inventory minus on-hand inventory

D. Annual demand divided by average inventory

Correct Answer: B. On-hand inventory plus on-order inventory minus backorders

Question 2

In a continuous-review (Q,r) policy, an order is triggered when:

A. On-hand inventory equals zero

B. Demand exceeds the annual forecast

C. Inventory position is at or below the reorder point

D. The review calendar reaches month-end

Correct Answer: C. Inventory position is at or below the reorder point

Question 3

A product has average daily demand of 40 units and a constant six-day lead time. What is expected lead-time demand?

A. 46 units

B. 120 units

C. 240 units

D. 400 units

Correct Answer: C. 240 units

Question 4

Why does a periodic-review system generally require protection for the review period as well as lead time?

A. Because orders cannot be placed until the next review opportunity

B. Because suppliers always double the quoted lead time

C. Because safety stock is not permitted in periodic review

D. Because inventory position excludes on-order stock

Correct Answer: A. Because orders cannot be placed until the next review opportunity

Question 5

Which policy uses a lower trigger and an upper replenishment target?

A. Min–max policy

B. First-in-first-out policy

C. Economic production schedule only

D. Cross-docking policy

Correct Answer: A. Min–max policy

Question 6

A two-bin system is most suitable for which situation?

A. Highly perishable products with unpredictable seasonal demand

B. Low-value, standardized consumables with stable usage

C. Unique project equipment with no replenishment lead time

D. Products that cannot be physically counted or scanned

Correct Answer: B. Low-value, standardized consumables with stable usage

Question 7

If daily demand standard deviation is 6 units and constant lead time is 9 days, what is the approximate standard deviation of lead-time demand under independent daily demand?

A. 2 units

B. 6 units

C. 18 units

D. 54 units

Correct Answer: C. 18 units

Question 8

What is a principal risk of calculating replenishment only from on-hand inventory?

A. It eliminates all safety stock

B. It may cause duplicate orders despite existing pipeline supply

C. It guarantees excessive service levels

D. It prevents periodic review

Correct Answer: B. It may cause duplicate orders despite existing pipeline supply

Question 9

Which measure represents the proportion of demand supplied immediately from available stock?

A. Fill rate

B. Lead-time variance

C. Inventory turnover

D. Review period

Correct Answer: A. Fill rate

Question 10

Which practice can help reduce bullwhip effects associated with order-up-to systems?

A. Increasing undocumented forecast overrides

B. Ignoring point-of-sale information

C. Sharing demand data and reducing unnecessary forecast reactions

D. Placing orders only after a stockout occurs

Correct Answer: C. Sharing demand data and reducing unnecessary forecast reactions