Logistics Demand, Data and Service Requirements

Introduction and context

Logistics systems are designed to move products through space and time while satisfying customer requirements at an acceptable cost. The starting point for any logistics decision is therefore not the warehouse, vehicle, or information system; it is the demand that the system must serve. Demand determines how many orders are received, their timing, origins and destinations, shipment sizes, product characteristics, service expectations, and the variability that capacity must absorb.

In logistics, demand is more than a sales forecast. A forecast may estimate the number of units required, but logistics planning must translate that estimate into operational requirements: order lines, pallets, cubic metres, vehicle trips, dock appointments, labour hours, delivery routes, and inventory-position decisions. A poor translation can produce a system that has sufficient annual capacity but fails during daily or hourly peaks.

This chapter develops a structured approach to analysing freight demand and service requirements. It integrates descriptive analytics, demand segmentation, forecasting, lead-time analysis, capacity planning, uncertainty measurement, and managerial interpretation. Particular attention is given to the distinction between demand variability and operational unreliability, since both affect logistics performance but require different remedies.

Learning objectives

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

1. Define logistics demand and distinguish demand volume, frequency, profile, and variability.
2. Describe order, shipment, product, lane, and customer characteristics relevant to logistics design.
3. Select appropriate segmentation dimensions and descriptive measures.
4. Apply basic forecasting methods and evaluate their limitations.
5. Relate service levels and lead times to inventory, transport, and capacity requirements.
6. Calculate practical capacity and uncertainty measures.
7. Interpret demand data critically, including data-quality, governance, ethical, and technological issues.
8. Translate demand analysis into logistics policy and operational decisions.

1. The nature of logistics demand

1.1 Derived and dependent demand

Freight demand is often derived from commercial transactions. A customer order generates warehouse activity, packaging, transport, delivery, invoicing, and often reverse logistics. Thus, transport demand is dependent on product demand, but the relationship is not always one-to-one. Consolidation, inventory positioning, cross-docking, returns, and intermodal transfers can cause one customer order to generate several logistics activities.

Demand should be described at several levels:

• Macro demand: total freight volume for a market, region, product family, or period.
• Network demand: flows between facilities, suppliers, customers, and geographic zones.
• Operational demand: orders, order lines, units, pallets, weight, volume, stops, and handling events.
• Resource demand: labour hours, vehicle capacity, dock time, storage locations, and information-processing capacity.

The same demand can appear very differently at these levels. For example, 10,000 cases per week may be easy to transport if shipped in full truckloads, but difficult if spread across 4,000 small orders requiring individual picking and home delivery.

1.2 Demand dimensions

A rigorous demand profile includes at least the following dimensions:

1.3 Orders, shipments, and order lines

An order is a customer-level request, while a shipment is a physical movement dispatched at one time. An order can create multiple shipments, and a shipment can consolidate several orders. An order line identifies a distinct product requested within an order. This distinction is essential because warehouse labour is often driven more strongly by order lines and picks than by total units.

For example, 1,000 units in ten orders may require ten picking and dispatch transactions; the same 1,000 units in 500 orders may require 500 transactions and substantially more labour. Consequently, logistics analysts should avoid using sales units alone as a proxy for workload.

2. Data foundations for logistics analysis

2.1 Data sources

Common sources include enterprise resource planning systems, warehouse-management systems, transport-management systems, point-of-sale records, order-management systems, telematics, delivery scans, customer-service records, supplier schedules, and external calendars such as holidays or weather data.

A useful logistics data model links events through common identifiers. Essential fields may include order ID, order-line ID, SKU, quantity, order timestamp, requested delivery date, promised date, actual dispatch and delivery times, origin, destination, carrier, mode, weight, dimensions, temperature class, and exception code.

2.2 Data quality

Before modelling demand, analysts should test:

• Completeness: Are orders, cancellations, returns, and failed deliveries recorded?
• Accuracy: Are quantities, timestamps, weights, dimensions, and locations plausible?
• Consistency: Do units, time zones, product codes, and customer identifiers agree?
• Timeliness: Is data available at the point when decisions are made?
• Uniqueness: Are duplicated orders or scan events removed?
• Validity: Do values conform to permitted ranges and classifications?

A particularly important issue is the difference between observed demand and constrained demand. If a product was unavailable, observed sales may be zero even though underlying customer demand was positive. Similarly, a transport capacity shortage may suppress shipments. Forecasting observed transactions without recording stockouts, rejected orders, or capacity constraints can systematically underestimate future requirements.

2.3 Granularity and aggregation

Demand may be aggregated by SKU, product family, customer, lane, facility, hour, day, or week. Fine-grained data reveal variability but may be sparse and noisy. Aggregated data are more stable but can hide peaks and operational bottlenecks. The correct granularity depends on the decision: network design may use weekly lane volumes, whereas labour scheduling may require hourly order arrivals.

Aggregation should be reversible where possible. Analysts should retain transaction-level data and create controlled analytical views rather than overwriting the source records.

3. Descriptive analytics and demand profiles

Descriptive analytics answers: What happened, when, where, to whom, and with what variability?

3.1 Fundamental measures

For observations x₁, x₂, …, xₙ, the arithmetic mean is:

where x̄ is the average demand per period, xᵢ is demand in period i, and n is the number of periods. The units of x̄ are the same as the units of xᵢ.

The sample standard deviation is:

The denominator n − 1 is appropriate when the observations represent a sample used to estimate population variability. If the complete population is being described, a denominator of n may be used. The coefficient of variation is:

CV is dimensionless and is meaningful when x̄ is positive and not close to zero. It allows comparison of relative variability across products or lanes with different volumes.

Other useful measures include the median, quartiles, percentile demand, minimum, maximum, inter-arrival time, and fill rate. Percentiles are often more useful than averages for service design. For example, a facility sized for the 95th percentile of daily volume will have more protection against peaks than one sized for average volume, although it may have lower asset utilisation.

3.2 Seasonality, trend, and intermittency

A demand series may contain:

• Level: the typical magnitude of demand.
• Trend: persistent upward or downward movement.
• Seasonality: a recurring pattern linked to calendar or business cycles.
• Cyclical movement: longer-term variation not necessarily tied to a fixed calendar.
• Intermittency: many zero-demand periods interrupted by irregular positive demand.
• Random variation: residual fluctuation not explained by systematic patterns.

These components have different managerial implications. Seasonality may require temporary labour or fleet capacity; a trend may justify permanent capacity; intermittency may favour inventory pooling or specialised forecasting rather than simple averages.

3.3 Pareto and ABC segmentation

ABC analysis ranks items by annual consumption value, commonly calculated as annual units multiplied by unit value. High-value A items receive closer control, while C items may use simpler policies. However, value alone is insufficient for logistics. A low-value item that is bulky, hazardous, temperature-sensitive, or essential to a customer may deserve high operational attention.

A multidimensional segmentation can combine value, demand frequency, cube, criticality, and variability. For example, an item may be classified as “high volume–low variability,” “low volume–intermittent,” or “critical–short lead time.” Such segmentation is more informative for transport and warehouse design than a single ABC label.

3.4 Shipment and order profiles

Important distributions include order-size distribution, shipment-weight distribution, cube distribution, stop density, delivery-window distribution, and orders-per-route distribution. The mean shipment size can conceal a long tail of small shipments. That tail may drive vehicle trips, handling cost, and emissions.

A practical profile should answer questions such as:

• What proportion of orders are single-line orders?
• What percentage of volume moves in full loads?
• How many shipments require special handling?
• Which customers generate urgent or failed deliveries?
• What share of demand occurs during the top ten busiest hours?
• Which lanes are balanced, and which create empty backhaul capacity?

4. Demand segmentation and service differentiation

A single service policy rarely minimises total logistics cost. Customers differ in willingness to wait, economic value, delivery density, product criticality, and consequences of failure. Segmentation translates these differences into feasible service propositions.

Useful segmentation dimensions include:

1. Customer value: margin, strategic importance, lifetime value, or contractual priority.
2. Service sensitivity: cost of delay, stockout consequences, and required responsiveness.
3. Geography: urban, rural, domestic, international, or remote locations.
4. Order behaviour: stable replenishment, promotional spikes, emergency orders, or returns.
5. Product characteristics: perishability, hazard, fragility, security, and cube-to-weight ratio.
6. Channel: business-to-business, retail replenishment, e-commerce, marketplace, or direct-to-consumer.

Segmentation must not become arbitrary discrimination. Criteria should be transparent, relevant to the service being provided, and consistent with contractual and legal obligations. Service differentiation should also account for accessibility and the impact of delivery policies on vulnerable customers or communities.

5. Forecasting freight demand

5.1 Forecasting as a decision process

A forecast is a conditional estimate, not a promise. Its value depends on the decision it supports, the forecast horizon, the data available, and the cost of forecast error. Forecasting should therefore begin with a clear question: Is the forecast for fleet acquisition, labour rostering, inventory replenishment, route planning, or financial budgeting?

Forecast horizons commonly include:

• Operational: hours to days; dispatching, staffing, and appointments.
• Tactical: weeks to months; carrier contracts, seasonal labour, and inventory policy.
• Strategic: years; facilities, network structure, fleet technology, and capital investment.

5.2 Naïve and moving-average methods

A naïve forecast sets the next-period forecast equal to the latest observation:

where Fₜ₊₁ is the forecast for period t + 1 and Aₜ is actual demand in period t. It is difficult to beat for some short, stable series and provides a useful benchmark.

A k-period moving average is:

where k is the number of recent periods included. This method smooths random variation but responds slowly to sudden level changes and does not explicitly model seasonality.

A weighted moving average assigns weights wⱼ such that:

The weights are non-negative in the usual application, and larger weights may be assigned to more recent observations.

5.3 Exponential smoothing

Simple exponential smoothing is:

where 0 < α ≤ 1 is the smoothing constant. Higher α responds more rapidly to recent changes; lower α produces greater smoothing. The method is appropriate for a series with a relatively stable level and no strong trend or seasonality. Trend-adjusted and seasonal methods are required when those components are systematic.

5.4 Forecast-error measures

For forecast errors eₜ = Aₜ − Fₜ, common measures include:

and

MAE is in the original demand units. RMSE is also in original units but penalises large errors more heavily. Percentage measures can be misleading when actual demand is zero or near zero. Forecast accuracy should be assessed by horizon, segment, and business consequence—not only by one aggregate statistic.

Worked example: moving-average forecast

Suppose daily pallet demand for the last four days was 96, 104, 100, and 120 pallets. A four-day moving-average forecast for the next day is:

The result is mathematically valid, but managerial interpretation is required. If the 120-pallet observation reflects a recurring promotional event, the average may understate the next requirement. If it reflects an isolated disruption, smoothing may be appropriate. Forecast selection must therefore combine statistical performance with operational knowledge.

6. Service levels and lead times

6.1 Service-level concepts

“Service level” is not a single measure. It may refer to:

• On-time delivery: proportion delivered by the promised time.
• In-full delivery: proportion delivered in the requested quantity.
• OTIF: orders delivered on time and in full.
• Fill rate: proportion of demand supplied immediately from available stock.
• Cycle-service level: probability of no stockout during a replenishment cycle.
• Order completeness: proportion of orders shipped without partial fulfilment.
• Reliability: consistency of performance, not merely average speed.

A contract should define the numerator, denominator, time point, tolerance window, exclusions, and treatment of customer-caused delays. Otherwise, two organisations may report different service levels for the same physical performance.

6.2 Lead-time decomposition

Total order lead time can be decomposed as:

where L is total lead time; tᵣ is order-registration time; tₚ is processing time; tₚᵢ is picking and packing time; tₕ is waiting or holding time; tₜ is transport time; and tᵈ is final delivery or receipt time. All terms must use the same unit, such as hours or days.

The decomposition identifies improvement opportunities. If transport time is only 20% of total lead time but waiting accounts for 45%, purchasing faster vehicles may have little effect. Lead-time variability should also be measured, since customers often value predictable delivery more than a shorter but unreliable average.

6.3 Relationship to inventory and uncertainty

For continuous demand during lead time, expected lead-time demand is:

where d̄ is average demand per unit time and L is lead time in the same time unit.

If demand and lead time are independent and both variable, a commonly used approximation for the variance of lead-time demand is:

where L̄ is mean lead time, σ²_d is demand variance per unit time, d̄ is mean demand per unit time, and σ²_L is lead-time variance. This approximation assumes a stable process and independence between demand and lead time. If demand and lead time are correlated, a covariance term is required; ignoring correlation can understate or overstate safety requirements.

If a normal approximation is considered reasonable, safety stock may be estimated as:

where z is the standard-normal factor associated with the target cycle-service level and σ_DL is the standard deviation of lead-time demand. The normal approximation is less suitable for intermittent, highly skewed, or bounded demand, where empirical percentiles or specialised intermittent-demand methods may be preferable.

7. Capacity requirements

7.1 Capacity is multidimensional

Capacity is not simply tonnes per year. Logistics capacity includes space, weight, cube, equipment, labour, dock positions, route time, information-processing capability, and regulatory constraints. A vehicle may have unused weight capacity but no remaining cubic capacity; a warehouse may have available floor space but insufficient picking labour.

A basic workload estimate is:

where W is workload, q is the number of demand units, and r is resource time per unit. If q is 8,000 order lines per week and average handling time is 0.06 labour-hours per line, workload is 480 labour-hours per week.

Required nominal capacity is:

where u is planned utilisation, expressed as a decimal. If available productive labour is 600 hours and target utilisation is 0.80, practical capacity is 480 hours. Planning at 100% utilisation is unsafe because breaks, travel, maintenance, variability, and exceptions consume time.

7.2 Peak capacity and time buckets

Capacity should be tested in the same time buckets in which congestion occurs. Weekly capacity can appear adequate while Tuesday afternoon demand exceeds dock or labour capacity. A useful process is:

1. Forecast demand by hour, day, or shift.
2. Convert demand into resource requirements.
3. Compare requirements with available capacity.
4. Identify overload probability and duration.
5. Evaluate pooling, appointment controls, overtime, outsourcing, or demand shifting.

Worked example: labour capacity

A distribution centre expects 12,000 order lines in a week. Each line requires 0.05 direct labour-hours for picking and packing. The workload is:

If the operation plans for 85% utilisation, required paid capacity is:

If scheduled labour provides 680 hours, the shortfall is approximately 26 hours. Management might add labour, reduce processing time, smooth releases, postpone non-urgent work, or revise the service promise. Simply reporting that “680 hours are available” would conceal the utilisation constraint.

8. Managing demand uncertainty

Uncertainty has several sources:

• Aleatory uncertainty: inherent random variation in arrivals, quantities, and travel times.
• Epistemic uncertainty: lack of knowledge caused by poor data, new products, or changing behaviour.
• Structural uncertainty: changes in regulations, technology, competitors, network design, or customer channels.

The remedies differ. Statistical buffering may address random variation; better data and measurement address epistemic uncertainty; scenario planning and flexible contracts address structural uncertainty.

Useful approaches include:

• pooling demand across customers or facilities where service permits;
• postponing final configuration or dispatch decisions;
• using flexible labour and carrier capacity;
• setting capacity triggers based on demand percentiles;
• maintaining alternative suppliers and routes;
• running stress tests for promotions, disruptions, and weather;
• monitoring forecast bias rather than accuracy alone.

Forecast bias can be measured as:

A positive value indicates average over-forecasting under this definition; a negative value indicates under-forecasting. The sign convention must be stated because some organisations define error in the opposite direction.

9. Technology, AI, and analytical governance

Modern logistics platforms can combine transaction history, real-time scans, GPS data, weather, traffic, promotions, and external events. Machine-learning models may improve forecasts when relationships are nonlinear or when many explanatory variables exist. However, complexity does not guarantee accuracy. Models can reproduce stockout bias, fail under regime changes, and become difficult for managers to explain.

Good practice includes:

• establishing a clear baseline such as naïve forecasting;
• separating training, validation, and test periods;
• preventing future information from entering historical features;
• monitoring drift and forecast bias;
• documenting data lineage and model assumptions;
• retaining human review for exceptional or high-consequence decisions;
• protecting personal, commercial, and location data;
• assessing cybersecurity and access controls.

Ethical governance is important in delivery systems. Algorithms may allocate delivery windows or service priority in ways that disadvantage remote or lower-income communities. Worker-monitoring data may create intrusive surveillance. Data minimisation, purpose limitation, transparency, auditability, and compliance with applicable privacy and transport regulations should be integrated into system design.

10. Illustrative case: Northstar Fresh Distribution

Background

Northstar Fresh is an illustrative regional distributor supplying supermarkets and restaurants from one chilled distribution centre. It handles 3,200 SKUs, receives orders until 18:00, and promises next-day delivery. Management is considering whether to lease additional vehicles and expand the evening picking shift.

Problem

Annual volume has increased by 14%, but on-time-in-full performance has fallen from 96% to 91%. Average daily cases are 18,000, yet Monday and Thursday volumes are substantially higher. Management initially attributes the problem to insufficient trucks.

Analysis

The analytical team first reconciles order, shipment, warehouse, and delivery records. It finds that:

• 11% of customer orders are amended after the initial cut-off;
• 7% of orders contain at least one temperature-sensitive or short-life item;
• 18% of weekly cases are concentrated in the four hours before the cut-off;
• picking represents 42% of elapsed order-to-dispatch time, while vehicle travel represents 24%;
• 9% of late deliveries are caused by incomplete picking rather than transport delay;
• the largest customers have stable, high-volume orders, while small restaurants generate intermittent urgent orders.

The team segments customers by volume, order stability, delivery density, and criticality. It uses daily and hourly profiles rather than annual averages, calculates lane-level variability, and evaluates a baseline forecast against moving-average and seasonal models. It also measures lead-time components and distinguishes customer-requested changes from internal delays.

Approach and solution

Northstar adopts a differentiated policy. Stable supermarket orders receive earlier electronic order confirmation and scheduled delivery windows. Small restaurant orders retain flexibility but incur a transparent premium for post-cut-off urgency. The company introduces order-release smoothing, reserves a small flexible carrier capacity block for high-variability days, and changes picking waves to reflect the hourly demand profile.

Rather than immediately leasing vehicles, it adds evening labour on peak days, improves master data for case dimensions, and introduces a dashboard showing forecast, actual demand, backlog, pick completion, and carrier departure reliability. A pilot uses a 95th-percentile workload trigger for temporary labour and a weekly review of forecast bias.

Outcome

In the illustrative scenario, OTIF improves to 95% after the pilot, with the largest gain arising from reduced picking backlog. Vehicle utilisation rises because dispatch waves are better synchronised, while emergency transport purchases decline. The company later leases one additional vehicle only after confirming that residual peak demand, rather than warehouse delay, is the binding constraint.

Learning points

1. Annual growth does not identify the resource causing service failure.
2. Demand must be translated into orders, lines, cases, time buckets, and lanes.
3. Lead-time decomposition prevents investment in the wrong bottleneck.
4. Segmentation can improve service without offering the same expensive promise to every customer.
5. Flexible capacity is valuable when uncertainty is concentrated rather than permanent.
6. Data governance and operational definitions are prerequisites for credible analytics.

Challenges and limitations

Demand analysis is limited by structural breaks, missing events, promotional effects, returns, stockouts, and changing customer behaviour. Historical data may not represent a new channel or facility. Forecast accuracy can also improve while operational performance worsens if the forecast is aggregated too broadly or if the service metric is poorly defined.

Analysts should avoid false precision. A forecast of 10,247 daily cases may imply more certainty than the data support. Scenario ranges, percentile estimates, and explicit assumptions often provide better managerial guidance. Correlation should not be interpreted as causation: high demand and late deliveries may coincide because both are driven by promotions, not because demand alone causes lateness.

Best-practice checklist

• Define the decision, horizon, unit, and time bucket before analysing data.
• Distinguish orders, order lines, units, shipments, weight, and cube.
• Record constrained demand, stockouts, cancellations, and rejected orders.
• Segment by operationally meaningful characteristics, not value alone.
• Establish a simple forecast benchmark before adopting advanced models.
• Evaluate error, bias, stability, and business cost.
• Measure average and peak requirements separately.
• Decompose lead time into controllable components.
• Plan practical capacity below theoretical maximum utilisation.
• Document service-level definitions and exclusions.
• Use flexible capacity for uncertainty and permanent assets for persistent demand.
• Govern personal, commercial, worker, and geolocation data responsibly.

Chapter Summary

Logistics demand analysis converts commercial activity into operational requirements. A complete demand profile includes volume, frequency, order and shipment size, timing, geography, product characteristics, customer segment, variability, and service requirements. Orders, shipments, and order lines must be distinguished because they drive different processes and resources.

Descriptive analytics uses averages, standard deviations, coefficients of variation, percentiles, distributions, and time patterns to identify level, trend, seasonality, intermittency, and peaks. Segmentation by customer, product, geography, channel, criticality, and variability supports differentiated service policies. Forecasting methods such as naïve forecasts, moving averages, weighted averages, and exponential smoothing are useful when matched to the structure of the data and evaluated using appropriate error and bias measures.

Service level includes several distinct concepts, including on-time delivery, in-full delivery, OTIF, fill rate, and cycle service. Lead time should be decomposed into processing, handling, waiting, transport, and delivery components. Capacity must be evaluated across labour, space, weight, cube, equipment, docks, vehicles, and information systems, with attention to peak rather than average demand. Uncertainty may be random, informational, or structural, and each type requires different mitigation.

Finally, logistics analytics depends on data quality, transparent definitions, responsible technology, and ethical governance. The managerial objective is not simply to produce a forecast; it is to make better decisions about service promises, capacity, inventory, transport, and resilience.

Key Terms

Demand: The quantity and pattern of logistics activity requested or required by customers and operations.

Derived demand: Demand for logistics activities generated by demand for products or services.

Order line: A distinct product entry within a customer order.

Shipment: A physical movement dispatched as one logistics consignment.

Demand profile: A structured description of demand volume, timing, size, location, and variability.

Coefficient of variation: Standard deviation divided by the mean, expressing relative variability.

Seasonality: A recurring demand pattern associated with a calendar or business cycle.

Intermittent demand: Demand containing many zero observations and irregular positive observations.

Segmentation: Grouping customers, products, or flows according to relevant characteristics.

ABC analysis: Ranking items commonly by annual consumption value for differentiated control.

Forecast: A conditional estimate of future demand based on information and assumptions.

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

Lead time: Elapsed time between initiation of an order or activity and completion.

Service level: A defined measure of logistics performance against a customer or operational requirement.

OTIF: On-time-in-full delivery performance.

Capacity: The amount of work or flow a resource can process under stated conditions.

Utilisation: The proportion of available capacity used productively.

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

Demand uncertainty: Lack of certainty about future demand magnitude, timing, or structure.

Data lineage: The documented origin, transformation, and use of data.

Review Questions

Short Answer Questions

1. Why should logistics analysts distinguish orders, shipments, and order lines?
2. Define the coefficient of variation and explain when it is useful.
3. Name four distinct service-level measures.
4. What is the purpose of decomposing total lead time?
5. Why can average capacity be insufficient for logistics planning?

Long Answer Questions

1. Explain how demand volume, frequency, timing, geography, and variability jointly influence logistics system design.
2. Compare moving-average forecasting with exponential smoothing, including their assumptions and limitations.
3. Discuss how customer and product segmentation can support differentiated logistics service without creating unjustified discrimination.
4. Explain the relationship among lead-time variability, demand variability, safety stock, and service level.
5. Evaluate the opportunities and governance risks associated with AI-based logistics forecasting.

Analytical / Application Questions

1. A warehouse receives 9,600 order lines per week, each requiring 0.045 labour-hours. If planned utilisation is 80%, calculate required paid labour capacity and interpret the result.
2. Daily pallet demand over five days is 80, 95, 100, 85, and 140. Calculate the mean and explain why the mean alone may be inadequate for staffing the next week.
3. A distributor experiences late deliveries despite sufficient weekly vehicle capacity. Design a diagnostic analysis using demand profiles, lead-time decomposition, service metrics, and segmentation to identify the likely bottleneck.

Multiple Choice Questions with Answers

Question 1

Which measure is dimensionless and compares relative variability across demand series?

A. Mean

B. Standard deviation

C. Coefficient of variation

D. Median

Correct Answer: C. Coefficient of variation

Question 2

A four-period moving-average forecast uses demand observations of 80, 90, 100, and 110 units. What is the forecast for the next period?

A. 90 units

B. 95 units

C. 100 units

D. 105 units

Correct Answer: B. 95 units

Question 3

Which situation is most likely to produce constrained observed demand?

A. A promotion is accurately recorded

B. A product is unavailable when customers request it

C. A stable customer places a normal order

D. A carrier delivers early

Correct Answer: B. A product is unavailable when customers request it

Question 4

Which service measure requires delivery to be both on time and complete?

A. Transit time

B. OTIF

C. Vehicle utilisation

D. Order frequency

Correct Answer: B. OTIF

Question 5

If workload is 600 labour-hours and planned utilisation is 0.75, what paid capacity is required?

A. 450 hours

B. 600 hours

C. 750 hours

D. 800 hours

Correct Answer: C. 750 hours

Question 6

Simple exponential smoothing gives greater weight to recent observations when:

A. The smoothing constant α is higher

B. The smoothing constant α is zero

C. The forecast horizon is longer

D. The sample size is smaller

Correct Answer: A. The smoothing constant α is higher

Question 7

Which lead-time component represents time spent waiting before processing or transport?

A. Product value

B. Holding or waiting time

C. Demand mean

D. Order-line count

Correct Answer: B. Holding or waiting time

Question 8

Why can a vehicle have unused capacity by weight but no usable capacity remaining?

A. Demand is always seasonal

B. Cube or volume may be the binding constraint

C. Forecast error is necessarily zero

D. Service level is measured incorrectly

Correct Answer: B. Cube or volume may be the binding constraint

Question 9

Which action is most appropriate for persistent under-forecasting?

A. Ignore forecast bias

B. Review model inputs, assumptions, and bias systematically

C. Aggregate all demand until variability disappears

D. Set utilisation permanently to 100%

Correct Answer: B. Review model inputs, assumptions, and bias systematically

Question 10

A high-volume product with low variability and a low-volume product with intermittent demand should generally:

A. Receive identical control policies

B. Be analysed using differentiated policies

C. Both be excluded from forecasting

D. Both be classified only by product colour

Correct Answer: B. Be analysed using differentiated policies