Forecasting Process and Demand Planning

Introduction

Forecasting is not simply the production of a number called “expected demand.” It is a coordinated managerial process through which an organization converts information about markets, customers, products, operations, and external conditions into decisions about capacity, inventory, procurement, labour, finance, and customer service. In operations management, the quality of a forecast is therefore judged not only by statistical accuracy but also by whether it is timely, actionable, transparent, and appropriately connected to decisions.

Demand planning extends forecasting by embedding demand estimates in a cross-functional planning system. It addresses questions such as: What demand is the organization trying to serve? At what level of product, location, and time should demand be forecast? Which information should be treated as a genuine demand signal? Who owns the forecast? How should commercial judgment and alternative scenarios be incorporated? How should the agreed plan feed into Sales and Operations Planning (S&OP) or Integrated Business Planning (IBP)?

A mature process recognizes that uncertainty cannot be eliminated. Its purpose is to make uncertainty visible, quantify its consequences, align functions around a common set of assumptions, and enable timely decisions.

Learning Objectives

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

1. Explain the purpose, scope, and organizational role of demand forecasting.
2. Select suitable forecasting horizons and levels of aggregation.
3. Design a data-selection and forecast-governance process.
4. Evaluate statistical, judgmental, and scenario-based forecasting methods.
5. Explain forecast ownership, consensus planning, and exception management.
6. Integrate demand planning with S&OP and IBP.
7. Assess forecast performance using appropriate measures and recognize common sources of bias.
8. Critically evaluate technology, AI, ethical, and governance issues in demand planning.

1. Forecasting as a Decision Process

A forecast is an estimate of a future quantity conditional on information available at the time it is produced. It is not a promise, target, budget, or sales quota, although these may be compared with it. Confusing these concepts creates dysfunctional behaviour. For example, if sales incentives depend on lowering the forecast, sales managers may deliberately understate expected demand; if a forecast is treated as a fixed target, planners may conceal uncertainty rather than communicate it.

Forecasting supports decisions at several levels:

The horizon should be related to the lead time of the decision. A factory expansion cannot be based on a two-week forecast, while tomorrow’s dispatch plan should not depend primarily on a five-year market projection. Forecast uncertainty generally increases as the horizon lengthens, although aggregation and stable structural relationships can reduce apparent volatility.

A useful distinction is between independent demand and dependent demand. Independent demand arises from external customers and must usually be forecast. Dependent demand is derived from a planned production quantity through a bill of materials; it is normally calculated using material requirements planning rather than forecast separately. Forecasting a component independently when its usage is determined by a production plan can create inconsistency and double counting.

2. Defining the Forecast Before Choosing a Method

A forecasting project should begin with a clear definition of the forecast object. The following decisions are fundamental.

2.1 Purpose and decision linkage

The organization should state what decision the forecast will support. A forecast used for inventory replenishment may require daily item-location estimates, whereas a forecast for capacity investment may require monthly product-family or market estimates. The purpose determines the required horizon, granularity, accuracy standard, data latency, and ownership model.

A forecast without a decision owner often becomes a reporting exercise. Every forecast should therefore be connected to a decision such as purchase authorization, production release, workforce planning, budget revision, or customer allocation.

2.2 Forecast horizon

The horizon should cover the relevant decision lead time plus a planning buffer. For example, if a supplier requires eight weeks’ notice and the organization reviews procurement monthly, the procurement forecast should extend beyond eight weeks and should be refreshed at a frequency that allows action before the commitment point.

Forecasts are often organized into three overlapping horizons:

• Near term: constrained by orders, current inventory, confirmed promotions, and known operational events.
• Medium term: influenced by commercial plans, product transitions, supplier capacity, and aggregate market conditions.
• Long term: concerned with strategic scenarios, structural growth, technology, regulation, and network decisions.

The farther the horizon, the more important it becomes to use ranges, scenarios, and assumptions rather than a single apparently precise number.

2.3 Granularity and the hierarchy of demand

Demand can be represented across multiple dimensions: product, customer, channel, geography, facility, and time. A forecast at one level may be reconciled with forecasts at other levels. For example, the sum of regional forecasts should be consistent with the corporate total, and product-family plans should reconcile with the product portfolio.

Highly granular forecasts offer operational specificity but are often noisy and difficult to govern. Highly aggregated forecasts are more stable but may hide local constraints and product differences. The appropriate level is the lowest level at which the signal is sufficiently reliable and the decision can be executed.

Aggregation has an important statistical implication. If demand for several items is not perfectly correlated, the variability of their total may be proportionally lower than the variability of individual items. This is one reason aggregate forecasts are often more accurate than forecasts for every individual stock-keeping unit (SKU). However, aggregation must not conceal substitution, cannibalization, or capacity bottlenecks.

2.4 Forecast frequency and freezing rules

A process should specify how often forecasts are updated and when the plan is frozen. Excessive updating can create nervousness: orders, schedules, and supplier commitments change before previous decisions have produced results. A common design uses a rolling horizon with different rules by time bucket: an execution zone that is frozen, a tactical zone that can be changed with approval, and a strategic zone that remains flexible.

3. Data Selection and Demand Signal Design

The quality of a forecast depends on the quality and meaning of its inputs. Historical shipments are not always equivalent to demand. A stockout may produce low shipments despite high unmet demand; a forward-buy promotion may produce unusually high orders that do not represent normal consumption; and a customer cancellation may distort the observed series.

3.1 Core data categories

A demand-planning data set commonly includes:

• historical orders, shipments, consumption, returns, and cancellations;
• inventory availability and stockout records;
• prices, discounts, promotions, and trade allowances;
• product launches, discontinuations, substitutions, and pack changes;
• customer, channel, geography, and market-segment attributes;
• calendar effects, holidays, working days, and weather where relevant;
• competitor actions, macroeconomic indicators, and industry information;
• confirmed projects, tenders, contracts, and customer commitments.

Data should be selected because it has a defensible causal or predictive relationship with the demand decision, not merely because it is available.

3.2 Cleaning and demand reconstruction

Before modelling, planners should distinguish ordinary observations from exceptional events. A data-quality workflow may include validation of units, duplicate removal, missing-value treatment, product-code mapping, calendar alignment, and identification of censored demand caused by stockouts.

Where possible, demand should be reconstructed. If available inventory was zero for a period, observed sales represent fulfilled demand, not necessarily market demand. Reconstruction may use lost-sales estimates, customer backorders, substitution records, or comparable periods. Such adjustments must be documented because an estimated demand history is not an observed fact.

Outliers should not be removed automatically. An unusual order may signal a repeatable event, a structural market change, or a one-time anomaly. The correct treatment depends on the forecasting purpose. A promotion should be represented as a promotion if similar promotions will recur; it may be adjusted out when estimating baseline demand.

3.3 Baseline and causal demand

A baseline forecast estimates demand under normal conditions, excluding exceptional commercial interventions. A causal uplift represents the incremental impact of a price change, promotion, distribution expansion, or other identifiable driver. The total plan can be expressed conceptually as:

The identity is a planning decomposition rather than a universal statistical law. Each term must be defined consistently, and double counting must be avoided. For instance, if historical promotional sales are already included in a model’s baseline, adding the promotion uplift again will overstate demand.

4. Forecasting Methods

4.1 Statistical and quantitative methods

Quantitative methods use historical data and, in some cases, explanatory variables. Common approaches include naïve forecasts, moving averages, exponential smoothing, regression, seasonal models, and machine-learning methods.

A naïve forecast uses the most recent observation as the next-period estimate:

where Y_t is actual demand in period t and F_t+1 is the forecast for period t+1. It is a useful benchmark, particularly for series without stable level or seasonality.

A simple moving average of order k is:

It assumes that the recent k observations are relevant and equally weighted, and is most suitable where demand has a relatively stable level and limited trend. A larger k smooths more strongly but responds more slowly to change.

Simple exponential smoothing is:

where 0 < α ≤ 1 is the smoothing constant, Y_t is actual demand, and F_t is the forecast for period t. It is appropriate for a series with a changing level but no systematic trend or seasonality. The terms have the same units as demand, and α is dimensionless.

For a series with trend, Holt’s method estimates level and trend. One common formulation is:

Here ℓ_t is the estimated level, b_t is the estimated trend per period, h is the forecast horizon, and α and β are smoothing parameters between 0 and 1. The method assumes that a local trend is informative and may perform poorly when trends are temporary or structural breaks occur.

Regression models are useful when demand is related to explanatory variables. A basic model may be written as:

where Y_t is demand, X_jt are predictors, β_j are estimated coefficients, and ε_t is the unexplained error. The model is useful only when the predictors are available or can themselves be forecast for the required horizon, relationships are sufficiently stable, and assumptions concerning error behaviour are examined. Correlation alone does not establish causation.

4.2 Judgmental forecasting

Judgment is necessary when historical data are sparse, products are new, markets are changing, or known future events are not represented in the past. Methods include expert panels, sales-force composite estimates, customer-intention surveys, analogy, executive judgment, and structured workshops.

Judgment should be structured rather than informal. A robust process asks contributors to state the assumption, evidence, expected timing, magnitude, confidence, and owner. It separates the statistical baseline from the judgmental adjustment so that overrides can later be evaluated. This creates accountability without assuming that all judgment is wrong.

Important cognitive risks include anchoring on the previous forecast, optimism bias, availability bias, groupthink, escalation of commitment, and deliberate sandbagging. Independent initial estimates, evidence templates, anonymous input before meetings, and explicit challenge roles can reduce these risks.

4.3 Combining forecasts

Combining methods can improve robustness when methods capture different information. A simple combination is a weighted average:

where 0 ≤ w ≤ 1. The weights should have a documented rationale and may be informed by historical performance, although past performance is not a guarantee under structural change. Combining should not become a way to average incompatible assumptions without understanding them.

5. Scenarios, Uncertainty, and Forecast Accuracy

A single-point forecast can create false precision. Scenario planning represents plausible states of the world, such as base, upside, and downside cases. Each scenario should specify drivers, not merely a percentage adjustment. For example, an upside scenario may assume increased distribution, a successful product launch, and stable supply; a downside scenario may assume competitor entry, reduced consumer spending, and delayed customer projects.

Scenarios should be internally coherent. A high-demand scenario may also require additional labour, constrained supplier capacity, higher working capital, and different service risks. The purpose is not to predict every outcome but to prepare decisions and triggers.

Forecast accuracy should be measured using more than one metric. For actual demand Y_t and forecast F_t, error may be defined as:

Mean absolute error (MAE) is:

where n is the number of observations. MAE is expressed in demand units and is easy to interpret.

Mean absolute percentage error (MAPE) is:

It is unsuitable when actual demand is zero and can become unstable when actual demand is close to zero. A scale-free alternative is the weighted absolute percentage error:

WAPE is interpretable at aggregate level but can be dominated by high-volume items. Bias can be monitored using mean error:

Positive mean error under this definition indicates systematic under-forecasting, because actual demand exceeds forecast on average. Metrics should be calculated on comparable horizons, products, and demand states. A forecast that appears inaccurate because a promotion was cancelled should not be judged identically to a forecast that ignored a known promotion.

Accuracy is not the only performance criterion. Organizations should also assess forecast value added, inventory effects, service level, obsolescence, planner workload, stability, and decision quality. Improving forecast accuracy at the cost of excessive inventory may destroy value.

6. Forecast Ownership and Governance

Forecast ownership means accountability for the integrity and use of the forecast, not that one person possesses all relevant knowledge. A typical governance design assigns:

• Demand planner: owns the process, data quality, baseline preparation, exception analysis, and documentation.
• Sales and marketing: provide customer intelligence, promotion plans, launch assumptions, and market evidence.
• Supply planning: tests feasibility against capacity, materials, labour, and lead times.
• Finance: connects assumptions with revenue, margin, cash, and business-plan implications.
• Executive or S&OP leader: resolves cross-functional trade-offs and approves the consensus plan.

A RACI matrix can clarify who is Responsible, Accountable, Consulted, and Informed for each activity. Governance should define forecast versions, approval dates, override thresholds, escalation rules, data ownership, and audit trails.

Exception management is more efficient than treating every item identically. Exceptions may be triggered by unusually large forecast changes, persistent bias, low data quality, high value, supply risk, or a gap between demand and capacity. The thresholds should be calibrated: too many alerts create alert fatigue; too few allow important risks to pass unnoticed.

7. Consensus Planning and the S&OP/IBP Cycle

Consensus planning creates one agreed demand plan from statistical analysis, commercial intelligence, customer commitments, and documented assumptions. It does not mean that every participant obtains the number they prefer. It means that differences are surfaced and resolved using an agreed decision process.

A typical monthly S&OP cycle contains the following stages:

1. Data and portfolio review: validate actuals, master data, product changes, and history.
2. Statistical baseline: produce the unconstrained demand forecast and identify exceptions.
3. Demand review: incorporate promotions, launches, customer intelligence, and market events.
4. Supply review: test capacity, materials, labour, logistics, and inventory feasibility.
5. Pre-S&OP reconciliation: evaluate gaps, scenarios, financial consequences, and recommendations.
6. Executive S&OP: decide priorities, resource allocation, risks, and actions.
7. Execution and learning: publish the plan, monitor assumptions, and review forecast performance.

In traditional S&OP, the focus may be product volume and operational balance. IBP broadens the process to include financial plans, strategic portfolios, new-product decisions, profitability, risk, and enterprise value. The distinction is not universal across organizations; IBP should be understood as an expanded integration ambition rather than a guaranteed maturity label.

The demand plan should remain distinct from the supply plan. Demand planning asks what customers are likely to require; supply planning asks what can be produced, procured, or delivered; the executive process decides how to manage any gap. Replacing an unconstrained demand forecast with a constrained supply number may hide lost-sales risk and weaken commercial decisions.

8. Worked Example: From Baseline to Consensus Plan

Consider a distributor forecasting monthly demand for a product family. The last three months of normal demand were 1,000, 1,100, and 1,200 units. A three-month moving-average baseline for the next month is:

Marketing has approved a promotion expected to create an incremental 250 units. Sales reports a customer project likely to require 150 units, with a confidence assessment of 60%. If the planning convention uses expected value for uncertain opportunities, the expected project demand is:

The preliminary unconstrained demand plan is therefore:

Supply planning identifies a temporary capacity limit of 1,300 units. The consensus process should not simply replace 1,440 with 1,300. It should record 1,440 as expected demand, 1,300 as available supply, and 140 units as an identified gap. The executive decision may be to add overtime, allocate scarce stock to strategic customers, accept delayed delivery, or revise the promotion. The resulting decision and assumption should be recorded for later evaluation.

This example illustrates three principles: the baseline should be visible, judgmental additions should be evidenced, and demand–supply gaps should be managed explicitly rather than concealed in the forecast.

9. Illustrative Case: Meridian Home Appliances

Background

Meridian Home Appliances sells small kitchen appliances through retailers, e-commerce, and institutional buyers. Its forecasting process had evolved independently within sales, finance, and supply chain. Sales submitted optimistic monthly estimates, finance used annual budget phasing, and supply planning relied on shipments from the previous year.

Problem

The organization experienced frequent stockouts on promoted products, excess inventory after promotions, expedited freight, and conflict between functions. Performance reports showed acceptable aggregate MAPE, but the metric excluded stockout periods and did not reveal systematic under-forecasting during launches.

Analysis

A cross-functional diagnostic found four root causes. First, shipments were used as a proxy for demand despite recurrent stockouts. Second, promotions were not coded consistently. Third, there was no distinction between a statistical baseline and commercial overrides. Fourth, the consensus meeting changed numbers without recording reasons or owners.

The team also found that forecast accuracy was measured at a highly aggregated level, masking errors for important SKU–channel combinations. Forecast ownership was ambiguous: everyone influenced the number, but no one owned the process.

Approach and solution

Meridian introduced a rolling 18-month process with weekly operational visibility and monthly S&OP governance. The first three months were treated as an execution zone, months four to twelve as a tactical planning zone, and the remaining months as a scenario zone. A baseline exponential-smoothing forecast was generated for stable items, while new products used analogues and structured commercial assumptions.

Promotions were registered in a common calendar with start date, duration, price change, expected uplift, cannibalization risk, and responsible manager. Stockout periods were flagged and reviewed rather than automatically treated as low demand. Demand planners owned the baseline and data quality; sales owned customer and promotion assumptions; supply planning owned feasibility analysis; finance translated scenarios into revenue and margin implications.

The monthly meeting began with exceptions rather than line-by-line negotiation. Every override required a reason code, expected impact, confidence level, and review date. Base, upside, and downside scenarios were maintained for products with significant launch or market risk.

Outcome

After two planning cycles, Meridian had not eliminated forecast error, but decisions improved. Promotion-related capacity gaps were identified earlier, inventory commitments became more selective, and disagreements were expressed as explicit assumptions rather than unexplained number changes. Management also discovered that one product family had structurally declining demand and stopped using historical growth assumptions in its capacity plan.

Learning points

The case demonstrates that process design can create value even without sophisticated algorithms. Forecast accuracy is affected by data definitions, incentives, granularity, ownership, and decision discipline. Technology cannot compensate for an undocumented promotion or a forecast that confuses constrained supply with expected demand.

10. Technology, AI, and Contemporary Developments

Modern demand-planning platforms can automate data ingestion, hierarchy management, model selection, anomaly detection, scenario analysis, workflow approval, and performance dashboards. Machine-learning methods may identify nonlinear relationships and interactions across many variables. They are especially useful where demand is affected by rich, high-frequency data and where sufficient history exists.

However, algorithmic sophistication does not guarantee better decisions. Models can reproduce historical bias, fail under structural breaks, and create false confidence through opaque outputs. Generative AI may assist with explaining forecast changes, summarizing market information, or querying planning data, but generated explanations must be validated. Human accountability remains necessary for commercial commitments, customer allocation, safety-critical supplies, and regulatory decisions.

Good practice includes model version control, reproducible data pipelines, access controls, monitoring for drift, explainability appropriate to the decision, and a clear override policy. Personal customer information should be minimized and protected. Data obtained from customers, employees, or third parties should be used lawfully and ethically, with attention to consent, confidentiality, discrimination, and contractual restrictions.

Contemporary developments also include probabilistic forecasting, which represents a distribution or prediction interval rather than only a point estimate; demand sensing using near-real-time signals; collaborative planning with suppliers and customers; and digital twins for testing capacity and inventory scenarios. These developments are valuable when they improve a decision, not merely when they increase technological complexity.

11. Challenges and Limitations

Several limitations recur in professional forecasting:

• Structural breaks: pandemics, regulation, technology shifts, or competitor entry can invalidate historical relationships.
• Intermittent demand: many zero observations make conventional percentage metrics and standard models unreliable.
• New products: there is little or no direct history, so analogies and assumptions dominate.
• Promotional distortion: forward buying and cannibalization make observed sales difficult to interpret.
• Bullwhip effects: order variability may exceed consumer-demand variability because of batching, incentives, and shortage gaming.
• Hierarchy inconsistency: bottom-up and top-down forecasts may disagree unless reconciliation rules are defined.
• Organizational politics: forecasts may be manipulated to obtain inventory, capacity, or favourable performance evaluations.
• Overfitting: complex models may fit historical noise and fail in future periods.
• Metric misuse: one aggregate accuracy metric can hide bias, service failures, or high-value errors.

A mature organization responds through segmentation, scenario analysis, review of assumptions, independent challenge, and continuous learning. It does not promise perfect accuracy.

Best-Practice Design Checklist

An effective end-to-end process should:

1. Begin with the decision and required lead time.
2. Define the forecast object, horizon, frequency, and hierarchy.
3. Separate independent demand from dependent requirements.
4. Establish trusted data definitions and document adjustments.
5. Generate a transparent statistical baseline.
6. Capture commercial judgment through structured assumptions.
7. Keep unconstrained demand separate from constrained supply.
8. Use scenarios where uncertainty is material.
9. Assign clear ownership, approval rights, and escalation rules.
10. Measure accuracy, bias, stability, service, inventory, and forecast value added.
11. Reconcile plans across product, location, time, and financial hierarchies.
12. Connect the process to S&OP or IBP and review outcomes continuously.

Chapter Summary

Forecasting and demand planning form an end-to-end decision process rather than a purely statistical exercise. The process begins by defining purpose, horizon, frequency, granularity, and the decision that the forecast will support. Data must be interpreted carefully because shipments may be constrained by stockouts, promotions may create temporary distortions, and new products may lack historical observations.

Quantitative methods provide disciplined baselines, while judgment is necessary for launches, customer intelligence, market change, and events not represented in history. Judgment should be structured, documented, and evaluated rather than inserted invisibly. Scenarios communicate uncertainty and expose the operational and financial consequences of alternative assumptions.

Forecast ownership requires clear roles for demand planning, sales, marketing, supply, finance, and executive decision-makers. Consensus planning should reconcile these perspectives without concealing the difference between expected demand and available supply. S&OP and IBP provide the governance mechanisms through which demand, supply, finance, risk, and strategy are aligned.

Forecast metrics such as MAE, MAPE, WAPE, and mean error are useful but must be selected and interpreted in context. Technology and AI can improve scale, speed, and pattern recognition, but they require data governance, model monitoring, ethical safeguards, and human accountability. The central managerial objective is not perfect prediction; it is better decisions under uncertainty.

Key Terms

Forecast: An estimate of a future quantity based on available information.

Demand planning: The coordinated process of estimating, interpreting, and governing future customer requirements.

Forecast horizon: The future time span covered by a forecast.

Granularity: The level of detail at which a forecast is prepared.

Independent demand: Demand originating from external customers.

Dependent demand: Demand derived from another planned item or production quantity.

Baseline forecast: Expected demand under normal conditions before specified interventions.

Causal uplift: Incremental demand attributed to an identifiable driver such as a promotion.

Judgmental forecast: An estimate incorporating structured expert or commercial knowledge.

Consensus forecast: A cross-functional forecast agreed through a defined governance process.

Scenario: A coherent set of assumptions describing a plausible future state.

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

MAE: Mean absolute error, expressed in the units of demand.

MAPE: Mean absolute percentage error, a percentage metric sensitive to zero or near-zero actuals.

WAPE: Weighted absolute percentage error based on total absolute error divided by total actual demand.

S&OP: Sales and Operations Planning, a cross-functional process for balancing demand and supply.

IBP: Integrated Business Planning, an expanded planning process linking operations with finance, portfolio, and strategy.

Override: A deliberate adjustment to a statistical forecast.

Exception management: Focusing planning attention on material deviations, risks, or data problems.

Forecast value added: The incremental benefit created by each forecasting-process intervention.

Review Questions

Short Answer Questions

1. Why should a forecast be linked explicitly to a decision and its lead time?
2. Distinguish independent demand from dependent demand.
3. What is the purpose of separating a baseline forecast from promotional uplift?
4. State one limitation of MAPE.
5. Why should unconstrained demand be kept distinct from constrained supply?

Long Answer Questions

1. Explain how horizon, granularity, frequency, and hierarchy should be selected in an end-to-end demand-planning process.
2. Compare quantitative and judgmental forecasting methods, including their strengths, limitations, and governance requirements.
3. Describe a complete monthly S&OP cycle and explain how IBP extends its scope.
4. Discuss the causes of forecast bias and recommend organizational mechanisms for reducing it.
5. Evaluate the opportunities and risks associated with AI-enabled demand planning.

Analytical / Application Questions

1. A product has actual demand of 800, 950, and 1,050 units in three consecutive months. Calculate the three-month moving-average forecast for the next month and explain one reason why the result may be unsuitable if a promotion is planned.
2. A sales team proposes 2,000 units for a customer opportunity with a 40% probability of conversion. Explain how expected-value reasoning would treat the opportunity and identify at least two qualitative factors that should also be reviewed.
3. A supply plan can provide 5,000 units while the unconstrained demand plan is 5,800 units. Design the information that should be presented to the executive S&OP meeting and recommend two possible responses.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes a forecast from a target?

A. A forecast is an estimate of likely future demand, whereas a target is a desired outcome.

B. A forecast is always lower than a target.

C. A target is calculated only from historical data.

D. A forecast is a contractual commitment to supply.

Correct Answer: A. A forecast is an estimate of likely future demand, whereas a target is a desired outcome.

Question 2

A three-month moving average uses demands of 900, 1,000, and 1,100 units. What is the next-period forecast?

A. 950 units

B. 1,000 units

C. 1,050 units

D. 1,100 units

Correct Answer: B. 1,000 units

Question 3

Which situation most clearly indicates dependent demand?

A. Retail customers buying finished bicycles

B. Consumers purchasing coffee from a supermarket

C. Steel required according to a planned production schedule

D. A customer placing an uncertain service request

Correct Answer: C. Steel required according to a planned production schedule

Question 4

Using the chapter’s error definition e_t = Y_t − F_t, a positive mean error generally indicates:

A. Systematic over-forecasting

B. Systematic under-forecasting

C. Zero demand variability

D. Perfect forecast accuracy

Correct Answer: B. Systematic under-forecasting

Question 5

What is a principal weakness of MAPE?

A. It cannot be expressed as a percentage

B. It ignores the forecast horizon completely

C. It is unstable or undefined when actual demand is zero or near zero

D. It measures only positive errors

Correct Answer: C. It is unstable or undefined when actual demand is zero or near zero

Question 6

Which activity belongs primarily to the demand review stage of S&OP?

A. Testing machine capacity against the plan

B. Incorporating approved promotions and customer intelligence

C. Authorizing supplier invoices

D. Sequencing jobs on a production line

Correct Answer: B. Incorporating approved promotions and customer intelligence

Question 7

Why should stockout periods be flagged during demand-data preparation?

A. Shipments during stockouts may understate true customer demand

B. Stockouts always prove that the forecast was accurate

C. Stockouts eliminate the need for a baseline forecast

D. Stockouts make all historical data unusable

Correct Answer: A. Shipments during stockouts may understate true customer demand

Question 8

What is the main purpose of a scenario in demand planning?

A. To guarantee the exact future outcome

B. To replace all statistical forecasts with opinions

C. To represent coherent alternative assumptions and their consequences

D. To conceal supply constraints from executives

Correct Answer: C. To represent coherent alternative assumptions and their consequences

Question 9

Who typically owns the baseline process, data quality, and exception analysis?

A. Demand planner

B. External customer only

C. Warehouse security officer

D. Financial auditor exclusively

Correct Answer: A. Demand planner

Question 10

If unconstrained demand is 1,440 units and available supply is 1,300 units, what gap should be explicitly considered?

A. 40 units

B. 100 units

C. 140 units

D. 2,740 units

Correct Answer: C. 140 units