Single-Period, Perishable and Lifecycle Inventory

Introduction and context

Inventory decisions differ fundamentally according to whether unsold stock can be carried forward. In a continuous-review system, excess units may remain useful next period. In a single-period system, however, the decision-maker commits before demand is known and faces a finite selling or operating horizon. Examples include fresh food, newspapers, event capacity, airline seats, seasonal apparel, vaccines with short shelf lives, fashion products, and spare capacity for a specific service date.

The central managerial problem is not simply to minimize purchase cost. It is to balance two asymmetric errors: ordering too little and losing profitable sales, or ordering too much and being left with obsolete, discounted, returned, or discarded stock. The newsvendor model provides a disciplined framework for this trade-off. Its logic can be extended to perishability, substitution, markdowns, returns, seasonal demand, product life cycles, and obsolescence.

Learning objectives

After studying this chapter, the reader should be able to:

1. Explain the economic logic of single-period inventory decisions.
2. Formulate and solve the newsvendor model using underage and overage costs.
3. Interpret the critical fractile and conduct sensitivity analysis.
4. Adapt the model to salvage value, returns, markdowns, substitution, and perishability.
5. Distinguish short physical shelf life from commercial obsolescence.
6. Develop lifecycle inventory policies for seasonal and technology products.
7. Evaluate the limitations, governance issues, and implementation requirements of these models.

1. The economics of a single-period decision

Let a firm choose an order quantity before random demand is observed. If demand exceeds available inventory, the firm incurs an underage cost: lost contribution, emergency replenishment, lost goodwill, or a service failure. If inventory exceeds demand, it incurs an overage cost: holding, markdown, disposal, return processing, or capital tied up in stock.

The decision is therefore made under uncertainty, but it is not arbitrary. A rational quantity balances the marginal expected cost of one additional unit against its marginal expected benefit.

Important terminology includes:

• Single-period inventory: Inventory purchased or positioned for one selling or operating opportunity.
• Perishable inventory: Inventory whose physical usability or value declines with time.
• Obsolescence: Loss of commercial value because of fashion, technology, regulation, or changing customer preference.
• Underage: Demand that cannot be served because the chosen quantity is too small.
• Overage: Inventory remaining after the relevant demand period.
• Salvage value: Net value recovered from an unsold unit.
• Critical fractile: The target probability that demand will not exceed the chosen quantity.
• Markdown: A planned price reduction used to clear remaining inventory.

2. The basic newsvendor model

Assume a single product, one order opportunity, uncertain demand, and no replenishment during the period. Let:

The standard assumptions are that demand is known probabilistically, the selling price and costs are known, the order is placed before demand is observed, and leftover units have a common net salvage value. Demand may be discrete or continuous.

When the product is sold at price p and an unsold unit has salvage value v, with c > v, the costs are:

These expressions measure contribution lost from insufficient stock and net loss from excess stock, respectively. They assume that a shortage loses the sale completely and that salvage is received net of recovery costs. If a shortage causes goodwill loss, substitution, or expedited purchasing, those effects must be added to Cᵤ.

The critical fractile is:

Equivalently, using p, c, and v:

The optimal quantity Q is the demand quantile corresponding to this probability. Thus, the firm should choose the quantity for which the probability of demand being no greater than Q equals the critical fractile.

Derivation through marginal expected value

Consider increasing the order quantity by one unit. The additional unit is useful when demand reaches that unit, which occurs with probability approximately 1 − F(Q). Its expected benefit is therefore Cᵤ[1 − F(Q)]. If demand does not reach it, which occurs with probability F(Q), it creates overage cost CₒF(Q). At the optimum, these expected marginal effects are equal:

Rearranging gives the critical-fractile equation. The result is valid when the demand distribution is correctly specified and the cost consequences are represented adequately.

For a continuous distribution, if D has known mean μ and standard deviation σ, and follows a normal distribution, then:

where z is the standard normal quantile satisfying Φ(z) = Cᵤ ÷ (Cᵤ + Cₒ). This normal approximation is inappropriate if demand is strongly skewed, bounded, intermittent, or affected by stockout censoring.

Worked example 1: basic newsvendor decision

A seasonal retailer buys insulated bottles for $18 each and sells them for $30. Unsold bottles can be cleared at a net value of $6. Weekly seasonal demand is normally distributed with mean 1,000 units and standard deviation 250 units.

Underage cost:

Overage cost:

The critical fractile is:

The 50th percentile of a normal distribution is its mean, so:

Although expected demand is also 1,000, this equality is not universal. It occurs here because underage and overage costs are equal. If the salvage value fell, overage cost would rise and the optimal quantity would be below the mean.

3. Discrete demand and expected-profit analysis

Demand is often forecast in discrete scenarios rather than by a continuous distribution. Suppose demand can be 80, 100, or 120 units, with probabilities 0.20, 0.50, and 0.30. A quantity should be selected from feasible order quantities, and expected profit can be computed directly:

Here p(d) is the probability of demand level d and π(Q,d) is profit for order quantity Q under that demand. The critical fractile remains conceptually useful, but the exact optimum may be an interval or may require comparison of neighboring quantities because the distribution is discrete.

For a selling price p, unit cost c, and salvage value v, profit in a realized demand state is:

The expression assumes all demand served is sold at p and all leftover inventory receives v. It should be modified for partial fulfillment, returns, multiple prices, or capacity constraints.

4. Sensitivity and managerial interpretation

The critical fractile rises when underage cost rises and falls when overage cost rises. Managers should order more when:

• the selling contribution is high;
• stockouts damage customer relationships;
• replenishment is impossible or expensive;
• salvage or return value is strong;
• demand uncertainty is lower, provided the distribution is centered appropriately.

Managers should order less when:

• markdown or disposal costs are high;
• demand is highly uncertain and negatively skewed;
• demand can be served through substitution;
• replenishment is fast;
• the product becomes obsolete rapidly.

A useful distinction is between forecast uncertainty and economic asymmetry. Higher uncertainty does not automatically imply a higher order quantity. If the mean remains constant, greater dispersion increases both shortage and excess risks; the direction of the optimal quantity depends on the shape of the distribution and the critical fractile. A high critical fractile places the firm on the upper tail of demand; a low fractile places it below the mean for many symmetric distributions.

5. Perishability and time-dependent value

Perishability has at least three forms:

1. Physical perishability: The item deteriorates or becomes unsafe, as with food and biological materials.
2. Functional perishability: The item no longer performs adequately, as with batteries or reagents.
3. Economic perishability: The item remains usable but loses market value, as with fashion or technology.

For a product with several time periods, a static newsvendor calculation may be insufficient. Inventory can be sold at different prices, transferred between locations, used as an input, or discarded. A dynamic model must include the value of carrying inventory into future periods. The decision is then based on a continuation value, not merely current salvage value.

If an item remaining at the end of period t has expected future value Vₜ₊₁, the effective terminal value in a one-step decision can be represented by Vₜ₊₁ net of deterioration and handling costs. The basic logic becomes:

This substitution is valid only when Vₜ₊₁ is a reliable, risk-adjusted value and does not depend on the quantity carried in a way that changes the demand or price environment.

Service-level interpretation

The critical fractile is sometimes called a service level, but the terms should not be confused. The newsvendor fractile is the probability that demand is fully covered by the chosen quantity in the modeled period. It is not necessarily a fill rate, cycle service level, or customer-service target used in a multi-period system.

For food and medicine, the decision also involves safety, legal compliance, traceability, temperature control, and disposal rules. A mathematically optimal quantity is unacceptable if it violates expiry or public-health requirements.

6. Markdown pricing and salvage channels

Unsold inventory can be recovered through several channels: a later full-price period, a markdown sale, an outlet, donation, recycling, supplier return, or disposal. These channels should be evaluated by net recovery value, including handling, transport, repackaging, commissions, taxes, and reputational effects.

If a unit is sold at a markdown price m after the primary period, and the additional markdown and processing cost is k, the net salvage value is approximately:

This approach assumes the markdown does not substantially cannibalize full-price sales. If customers strategically wait for markdowns, the primary-period demand distribution changes and the simple model becomes biased.

A lifecycle pricing policy often follows this sequence:

1. establish an initial price and order quantity;
2. monitor demand, inventory age, and remaining selling time;
3. update the demand forecast;
4. reduce price when the expected value of waiting is lower than the value of clearing stock;
5. reserve scarce units for high-value demand when capacity or inventory is constrained.

Markdowns are not merely an operational response. They are a commitment that may influence customer reference prices, brand positioning, and future demand. Therefore, the recovery value must include strategic effects where material.

7. Substitution and assortment effects

Substitution changes the underage cost. If a customer whose preferred product is unavailable buys an alternative, the firm may retain part of the margin. If the alternative is a competitor’s product, the loss is greater. If several products share capacity or customers, each item cannot be optimized independently.

For product i, a simple adjusted underage cost may be written as:

This is a managerial approximation, not a universal law. The recovered contribution depends on substitution probability, alternative margin, service effects, and whether the substitute itself becomes scarce.

In a portfolio, the firm may face a common purchasing budget, shelf-space constraint, production capacity, or component supply. The resulting problem is a multi-product newsvendor or assortment optimization model. Correlated demand is especially important: ordering separately using marginal distributions can produce excessive total risk when products experience the same seasonal shock.

8. Returns, buyback agreements, and vendor-managed risk

Returns alter the economics of overage. If the supplier accepts unsold units at credit r, the effective salvage value is the credit less return freight and processing costs. A buyback contract can increase the retailer’s order quantity by reducing Cₒ, while a revenue-sharing contract can reduce the retailer’s effective acquisition cost and alter both critical fractile and channel incentives.

Returns are not free. High return rates create reverse-logistics cost, inspection labor, waste, fraud exposure, and environmental burden. A contract should specify:

• eligibility and return deadlines;
• condition and packaging requirements;
• credit calculation;
• ownership during transit;
• data sharing and reconciliation;
• treatment of damaged or counterfeit goods;
• disposal and sustainability responsibilities.

The appropriate objective is total supply-chain value rather than the retailer’s isolated expected profit. Contract design should also avoid transferring risk in ways that encourage over-ordering, manipulation of forecasts, or unnecessary waste.

9. Seasonal products and demand learning

Seasonal inventory decisions combine a short selling window with limited opportunities to correct a forecast. Forecasts should be segmented by location, channel, customer type, and product variant. Historical demand must be adjusted for stockouts, promotions, weather, holidays, competitor actions, and changes in distribution.

A practical process is:

1. estimate a demand distribution rather than a point forecast;
2. estimate underage and overage costs transparently;
3. calculate a preliminary critical fractile quantity;
4. test the quantity against capacity, budget, and supplier constraints;
5. define replenishment, transfer, markdown, and disposal triggers;
6. update the forecast using observed sales and remaining time;
7. conduct a post-season forecast and bias review.

Observed sales during a stockout are censored: recorded sales may be lower than true demand. Treating them as ordinary low-demand observations systematically underestimates future requirements.

10. Product lifecycle and obsolescence

A product’s commercial life commonly passes through introduction, growth, maturity, and decline. Inventory policy should change across these stages.

Technology products create a special risk: a unit may have positive physical salvage but negative economic value after a new generation launches. Lifecycle planning should therefore include product roadmap information, engineering-change notices, regulatory deadlines, component lead times, and end-of-life commitments.

A final-buy decision should compare the expected value of purchasing now with the cost of being unable to serve demand after supply ends. It must consider demand decay, service obligations, repair requirements, component commonality, and disposal exposure. A final buy based solely on the last observed sales rate is often unsafe because it ignores decline dynamics and demand substitution.

11. Advanced decision extensions

Risk aversion and service constraints

Expected profit maximization treats a dollar of gain and a dollar of loss symmetrically in expectation. A risk-averse organization may impose a probability constraint, such as requiring a specified chance of meeting demand, or may use a utility function that penalizes losses more strongly than gains. Such a policy should be disclosed because it changes the objective from expected profit to risk-adjusted value.

Capacity and resource constraints

When a common resource is limited, the marginal value of capacity should be compared across products. A product with a high unit margin may not deserve capacity if it has high overage risk or displaces a product with greater expected contribution per constrained unit. This is a portfolio allocation problem, not a set of independent newsvendor calculations.

Bayesian updating and analytics

Forecasts should be updated as new sales, search activity, reservations, weather information, and competitor signals arrive. Bayesian methods can represent uncertainty about the demand distribution itself, rather than treating its parameters as known. Machine-learning models can improve conditional forecasts, but they do not determine the correct order quantity without economic costs and constraints.

AI systems should be governed through data-quality checks, forecast-drift monitoring, explainability for material decisions, human approval thresholds, and audit trails. Automation should not conceal a choice of underage and overage costs that is actually a managerial or ethical judgment.

12. Illustrative case: seasonal medical-device kit

The following is a fictional illustrative case.

Background

MedSupply prepares a disposable diagnostic kit for a six-week winter screening campaign. Each kit costs $42 to assemble and is reimbursed at an effective price of $70. After the campaign, unopened kits can be returned to the supplier for a $20 credit, but return processing costs $3 per kit. A stockout may cause a clinic to use a competitor’s kit, creating an estimated contribution loss of $28 per unserved kit. Demand is approximately normal with mean 8,000 and standard deviation 1,500 kits.

Problem

The operations manager must choose an initial quantity. The decision is complicated by returns, expiry, uncertain demand, and the clinical importance of availability.

Analysis

The net salvage value is:

The ordinary contribution from a served kit is $70 − $42 = $28. Thus:

The overage cost is:

The critical fractile is:

The standard normal quantile for 0.5283 is approximately z = 0.071. Therefore:

The preliminary quantity is approximately 8,107 kits. However, the manager should not treat this as an automatic release. Clinical service requirements may impose a higher target, while supplier return capacity, budget, expiry rules, and regional demand correlation may impose restrictions.

Approach and solution

MedSupply should release approximately 8,100 kits as the economic baseline, then apply a governance review. It should reserve a documented emergency allocation, monitor daily demand, and use regional transfers before new production. The forecast should be updated for campaign participation and weather. Any decision to exceed the baseline should identify whether it reflects a higher service requirement, a revised underage cost, or new demand evidence.

Outcome and learning points

The case demonstrates that the return agreement reduces overage cost and supports a quantity slightly above mean demand. It also shows that the calculated answer is a starting point, not a substitute for clinical, contractual, and operational judgment. In regulated products, service failure and patient risk may be more important than the narrow financial estimate.

13. Limitations, governance, and best practice

The basic model can mislead when demand is endogenous to price, supply is unreliable, lead times are uncertain, replenishment is possible, demand observations are censored, multiple products compete, or costs are nonlinear. It also assumes that probability estimates are credible. A precise calculation based on a poor distribution is not a precise decision.

Best practice includes:

• estimate distributions from cleaned, relevant, and sufficiently recent data;
• separate forecast error from demand variability;
• make underage and overage assumptions auditable;
• include net rather than gross salvage value;
• model substitutions and channel interactions;
• account for capacity, cash, shelf life, and regulatory constraints;
• use scenario and sensitivity analysis rather than one forecast;
• distinguish economic service levels from safety and legal requirements;
• review forecast bias and post-season outcomes;
• document human overrides and their rationale.

Ethically, inventory decisions can affect access to essential medicines, food waste, worker safety, and environmental burdens. Firms should not deliberately create artificial scarcity, manipulate return programs, or optimize disposal costs while externalizing harm. For regulated or safety-critical products, legal compliance and duty of care override simple profit maximization.

Chapter Summary

Single-period inventory decisions balance the cost of being short against the cost of being long. The newsvendor model expresses this balance through the critical fractile, F(Q*) = Cᵤ ÷ (Cᵤ + Cₒ), where underage and overage costs must be defined in economic, not merely accounting, terms. Salvage, returns, markdowns, substitution, and future use change the overage or underage economics. Perishability may be physical, functional, or commercial, and lifecycle products require policies that evolve from introduction through decline. Seasonal and technology products demand forecast updating, explicit end-of-life planning, and attention to censored demand. The model is valuable because it makes assumptions visible, but it must be supplemented by capacity analysis, risk constraints, contracts, governance, and ethical judgment.

Key Terms

Single-period inventory: Inventory committed for one finite selling or operating opportunity.

Newsvendor model: A model that balances underage and overage costs under uncertain demand.

Underage cost: Economic cost of having insufficient inventory.

Overage cost: Economic cost of leftover inventory.

Critical fractile: Target cumulative probability used to determine the optimal quantity.

Perishable inventory: Inventory whose usability or value declines with time.

Economic obsolescence: Loss of value caused by market, technology, or preference changes.

Salvage value: Net recovery value of unsold inventory.

Markdown: Planned reduction in selling price to clear inventory.

Substitution: Fulfillment of demand with an alternative product or source.

Censored demand: Demand that is not fully observed because stock was unavailable.

Demand distribution: Probability description of possible demand outcomes.

Service level: A measure of demand coverage or availability, whose precise meaning depends on the system.

Buyback agreement: Contract allowing a buyer to return eligible unsold stock for credit.

Lifecycle inventory: Inventory policy adapted to a product’s commercial stage.

Final buy: Last planned purchase before supply ends.

Continuation value: Expected future value of inventory carried into another period.

Demand correlation: Statistical dependence among demands for products, locations, or periods.

Forecast bias: Systematic tendency to over- or under-forecast.

Reverse logistics: Movement and processing of returned or recovered products.

Review Questions

Short Answer Questions

1. What are underage and overage costs, and why are they central to the newsvendor model?
2. Define the critical fractile and explain its managerial meaning.
3. Distinguish physical perishability from economic obsolescence.
4. How can a return agreement affect the optimal order quantity?
5. Why are sales observed during a stockout not necessarily equal to true demand?

Long Answer Questions

1. Derive the critical-fractile rule from the marginal expected cost and benefit of one additional unit.
2. Explain how markdowns, salvage, disposal, and returns should be incorporated into a single-period model.
3. Discuss the limitations of independently applying the newsvendor model to a multi-product assortment.
4. Develop an inventory policy for a product progressing from introduction to decline.
5. Evaluate the ethical, regulatory, and governance issues involved in optimizing inventory for medical or food products.

Analytical / Application Questions

1. A product sells for $50, costs $30, and has a net salvage value of $10. Demand is normally distributed with mean 2,000 and standard deviation 400. Calculate the critical fractile and identify the appropriate normal quantile direction for the order decision.
2. A retailer can substitute 40% of unavailable demand with an alternative generating $8 contribution, while the preferred product would generate $20 contribution. Explain how substitution changes the effective underage cost and the likely order quantity.
3. A technology product has falling demand, a possible redesign in four months, and a supplier final-buy deadline. Design a decision process incorporating forecast updating, obsolescence risk, service commitments, and disposal exposure.

Multiple Choice Questions with Answers

Question 1

If the underage cost is $12 and the overage cost is $8, the critical fractile is:

A. 0.40

B. 0.60

C. 0.67

D. 1.50

Correct Answer: C. 0.67

Question 2

A product costs $20, sells for $35, and has a net salvage value of $5. Its overage cost is:

A. $5

B. $15

C. $20

D. $30

Correct Answer: B. $15

Question 3

In the basic newsvendor model, a higher salvage value generally:

A. Raises overage cost and lowers the order quantity

B. Lowers overage cost and raises the order quantity

C. Raises underage cost and lowers the order quantity

D. Has no effect on the order quantity

Correct Answer: B. Lowers overage cost and raises the order quantity

Question 4

Which condition most directly makes the standard single-period model inappropriate?

A. Demand is uncertain

B. There is one selling period

C. Replenishment is possible during the period

D. Unsold inventory has salvage value

Correct Answer: C. Replenishment is possible during the period

Question 5

Observed sales during a stockout are called censored demand because they:

A. Include only returned units

B. May be lower than the demand that would have occurred with stock available

C. Are always recorded incorrectly

D. Exclude all online sales

Correct Answer: B. May be lower than the demand that would have occurred with stock available

Question 6

If demand is normally distributed and the critical fractile is 0.50, the optimal quantity is generally:

A. One standard deviation below the mean

B. The mean demand

C. One standard deviation above the mean

D. Zero

Correct Answer: B. The mean demand

Question 7

Which action usually reduces the overage cost of a seasonal product?

A. Eliminating all markdown channels

B. Increasing net salvage or return value

C. Increasing disposal fees

D. Reducing the selling price before demand occurs without analysis

Correct Answer: B. Increasing net salvage or return value

Question 8

Economic obsolescence refers to:

A. Physical damage during transportation only

B. Loss of value caused by changing technology, preferences, or markets

C. A legal prohibition on all returns

D. An increase in product shelf life

Correct Answer: B. Loss of value caused by changing technology, preferences, or markets

Question 9

A buyback agreement generally encourages a retailer to order more because it:

A. Increases the acquisition cost

B. Reduces the effective overage cost

C. Eliminates demand uncertainty

D. Guarantees that all demand will be observed

Correct Answer: B. Reduces the effective overage cost

Question 10

In the illustrative medical-device case, the net salvage value per kit was:

A. $3

B. $17

C. $20

D. $42

Correct Answer: B. $17