Demand Forecasting and Materials Planning

Introduction

Materials management connects market requirements with the resources needed to satisfy them. A materials manager must decide what materials are required, in what quantities, and at what time. These decisions are difficult because demand is uncertain, suppliers have variable lead times, inventory is costly, and production capacity is limited.

Demand forecasting provides an informed estimate of future requirements. Materials planning converts that estimate into purchasing, inventory, production, and capacity actions. Forecasts are not predictions of certainty; they are structured estimates that support decisions under uncertainty. A professional materials plan therefore combines forecast information with actual orders, inventory records, supplier performance, lead-time data, and capacity constraints.

This chapter develops qualitative and quantitative forecasting methods, aggregate planning, lead-time analysis, and capacity-based materials planning. It also explains how these elements are integrated through the master production schedule and materials requirements planning.

Learning Objectives

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

1. Explain the purpose, role, and limitations of demand forecasting.
2. Distinguish qualitative and quantitative forecasting methods.
3. Calculate and interpret moving averages, weighted averages, exponential smoothing, and forecast-error measures.
4. Explain aggregate planning and compare level, chase, and hybrid strategies.
5. Analyse lead time, safety stock, reorder points, and capacity constraints.
6. Translate demand information into a practical materials plan.
7. Evaluate the managerial, technological, ethical, and operational implications of forecasting decisions.

1. The Role of Forecasting in Materials Management

A forecast is an estimate of future demand for a product, service, component, or material over a specified period. Forecasts may concern:

• independent demand, such as customer demand for finished goods; or
• dependent demand, such as demand for components calculated from a production schedule and bill of materials.

The distinction is important. Independent demand is usually forecast from market information. Dependent demand is normally calculated after the required quantity of finished products is known.

Forecasts support decisions at several planning levels:

A forecast should be evaluated according to its purpose. A yearly forecast may be suitable for capacity expansion but unsuitable for weekly purchasing. Forecast accuracy generally decreases as the time horizon increases and as the product becomes more novel or volatile.

Characteristics of a useful forecast

A useful forecast should be:

• timely, so that action can be taken before demand occurs;
• sufficiently accurate for the decision involved;
• expressed in practical units, such as units, tonnes, litres, or monetary value;
• accompanied by an estimate of uncertainty;
• consistent with known business events, such as promotions or product launches; and
• monitored through forecast-error measures.

Forecast accuracy does not eliminate uncertainty. Its purpose is to reduce avoidable uncertainty and make the remaining uncertainty visible.

2. Demand Patterns and Forecasting Data

Before selecting a forecasting method, the planner should examine the demand pattern.

Level is the average magnitude around which demand varies. Trend is a persistent upward or downward movement. Seasonality is a repeating pattern associated with a calendar period, such as month, quarter, or week. Cyclical variation is a longer-term fluctuation associated with economic or industry conditions. Random variation is irregular movement that cannot be explained reliably from available information.

Demand data should be cleaned before analysis. The planner should investigate missing values, duplicate transactions, returns, cancelled orders, stock-out periods, unusual promotions, and changes in product coding. A recorded demand of zero may mean no customer wanted the product, or it may mean the product was unavailable. These situations have different planning implications.

A time series can be represented conceptually as the combination of systematic and irregular components. In practice, the components are not always separable, and judgement is needed when market conditions change.

3. Qualitative Forecasting Methods

Qualitative methods are based mainly on informed judgement rather than historical numerical patterns. They are particularly useful when historical data are limited or when the future differs substantially from the past.

Expert opinion

Managers, engineers, sales specialists, buyers, and industry experts provide estimates based on experience. This method is fast and inexpensive, but it can be affected by optimism, anchoring, hierarchy, and personal incentives.

Delphi method

The Delphi method gathers anonymous opinions from a panel through several rounds. A coordinator summarises the responses and returns them to the panel until a reasoned convergence is reached. Anonymity reduces the influence of dominant personalities.

Sales-force composite

Sales personnel estimate demand in their territories or customer segments. These estimates may contain useful local knowledge, but sales staff may overstate demand to obtain inventory or understate it to reduce performance targets. Clear definitions and review procedures are therefore necessary.

Market research and customer surveys

Research may include interviews, surveys, test marketing, and analysis of customer intentions. Stated intention is not always the same as actual purchase behaviour, so survey results should be interpreted cautiously.

Scenario analysis

The planner develops alternative futures, such as base, high-demand, and low-demand scenarios. Scenario analysis is valuable for capacity and supply-risk decisions because it recognises that a single-point forecast may be misleading.

Qualitative methods are not inferior simply because they are not mathematical. They are often essential for new products, structural market changes, regulatory changes, and disruptions. Their weakness is that assumptions may be difficult to reproduce or audit unless they are documented.

4. Quantitative Forecasting Methods

Quantitative methods use historical data and mathematical relationships. Their results should be reviewed against operational knowledge rather than accepted automatically.

4.1 Naïve forecasting

The simplest forecast for the next period is the actual demand in the current period:

Here, F_t+1 is the forecast for period t+1, and A_t is actual demand in period t. The method is easy to use and can be effective when demand follows a random walk or when a benchmark is required. It does not explicitly handle trend or seasonality.

4.2 Simple moving average

A moving average uses the most recent n observations:

The variables are actual demands A, forecast F, and number of periods n. All observations must use the same time unit and measurement unit. A larger n produces a smoother forecast but responds more slowly to changes.

Worked example. Demand over the last three months was 820, 900, and 880 units. The three-month moving-average forecast for the next month is:

The planner may use 867 units if whole units are required, while retaining the unrounded value for analysis.

4.3 Weighted moving average

A weighted moving average gives different importance to observations:

The weights w_i must be non-negative and sum to one. The most recent observation commonly receives the highest weight, but this is a planning choice rather than a mathematical requirement.

If the last three demands were 880, 900, and 820 units and weights of 0.5, 0.3, and 0.2 are assigned respectively, then:

4.4 Exponential smoothing

Single exponential smoothing updates the previous forecast using the latest forecast error:

Here, α is the smoothing constant, where 0 < α ≤ 1. A high value reacts rapidly to new demand; a low value produces greater smoothing. The method assumes that demand has no strong trend or seasonality unless extensions are added.

Suppose F_t = 850 units, A_t = 900 units, and α = 0.3. Then:

The initial forecast can influence early results, so the starting value should be documented.

4.5 Trend and seasonal methods

When demand has a trend, methods such as trend-adjusted exponential smoothing may be used. When demand has stable seasonal patterns, seasonal indices or decomposition methods may be appropriate. These methods require more data and careful validation. A seasonal index is usually interpreted as the ratio of demand in a period to average demand for the corresponding cycle. Seasonal indices should be normalised consistently so that their average over a complete cycle is approximately one.

Regression analysis can relate demand to explanatory variables such as price, advertising expenditure, temperature, or construction activity. A simple linear model is:

Here, Y is demand, X is an explanatory variable, a is the intercept, b is the estimated change in demand per unit change in X, and ε is unexplained error. Regression should not be interpreted as proof of causation merely because two variables are correlated.

5. Measuring Forecast Accuracy

Forecast evaluation should use historical forecasts compared with actual demand. A forecast error may be defined as:

Positive error means actual demand exceeded the forecast under this convention.

Mean absolute deviation

MAD is expressed in demand units and is easy to explain. It gives equal weight to errors of different signs.

Mean squared error

MSE is in squared demand units and penalises large errors more heavily.

Mean absolute percentage error

MAPE is a percentage, but it is unsuitable when actual demand is zero and can be unstable when actual demand is very small. In such cases, MAD, MSE, or an alternative scale-free measure may be preferable.

Forecast accuracy should not be judged by one measure alone. A low average error may hide systematic over-forecasting, under-forecasting, or serious errors during critical periods.

6. From Forecast to Aggregate Plan

Aggregate planning converts medium-term demand into a feasible overall production and resource plan. It normally uses product families rather than individual stock-keeping units. The plan balances demand, workforce, inventory, overtime, subcontracting, and backorders.

The principal strategies are:

An aggregate plan should consider total relevant cost, including regular production, hiring, layoffs, overtime, inventory holding, shortage, subcontracting, and sometimes lost sales. A simple planning identity is:

All terms must refer to the same period and product-family unit. If backorders are permitted, ending inventory may be negative, but the organisation must define how shortages are measured and controlled.

A practical sequence is to establish the demand scenario, determine available regular capacity, evaluate inventory and workforce policies, test alternatives, and then disaggregate the selected plan into a master production schedule.

7. Lead-Time Analysis and Inventory Protection

Lead time is the elapsed time between initiating an activity and completing it. Materials planning may involve supplier lead time, transport time, inspection time, internal processing time, queue time, and information-release time. A quoted supplier lead time is not necessarily the total replenishment lead time.

Lead-time data should be separated into planned lead time and actual lead time. The planner should monitor average lead time, variability, late-delivery frequency, minimum order quantities, and supplier reliability.

Reorder point

For continuous review of an item with approximately constant demand and lead time:

Here, d is average demand per time unit, L is lead time in the same time unit, and SS is safety stock in units. Unit consistency is essential.

If demand and lead-time uncertainty are approximately independent and normally distributed, a commonly used safety-stock expression is:

where z is the desired service-factor value and σ_DL is the standard deviation of demand during lead time. If demand per period has standard deviation σ_d and lead time is fixed at L periods, then:

This relationship assumes independent demand periods and a stable distribution. If lead time is also variable, a more general variance expression is needed, and the planner should use validated data rather than applying the fixed-lead-time formula blindly.

Worked example. Average weekly demand is 200 units, lead time is three weeks, and safety stock is 120 units. Therefore:

A replenishment order is triggered when the inventory position reaches approximately 720 units. Inventory position includes on-hand inventory plus on-order inventory minus valid backorders; it is not simply the physical quantity on the shelf.

8. Capacity Information and Materials Planning

A materials plan is feasible only if the required work can be performed with available resources. Capacity may be measured in units, labour hours, machine hours, weight, volume, or another meaningful constraint.

For a work centre:

Available effective capacity can be approximated as:

The exact interpretation of these factors must be defined by the organisation. Rated capacity is theoretical output; effective capacity recognises losses such as downtime, speed reduction, changeovers, absenteeism, and quality failures.

Example. A work centre has 10 machines, each available for 7 hours per day over 5 days. If availability is 90%, performance is 95%, and quality is 98%, effective weekly hours are:

If the weekly plan requires 3,100 hours, the shortfall is approximately 169.69 hours. Management must revise the plan, add overtime, use another work centre, subcontract, improve availability, or change the due-date promise. Releasing material to an overloaded work centre does not create capacity; it creates queues and work-in-process.

9. Translating Demand into a Materials Plan

The integrated planning flow is:

1. Collect customer orders, historical demand, market intelligence, and inventory data.
2. Produce a baseline forecast and document assumptions.
3. Review the forecast with sales, operations, finance, engineering, and procurement.
4. Create an aggregate plan for product families.
5. Develop a master production schedule for end items.
6. Use the bill of materials to calculate dependent component requirements.
7. Net requirements against on-hand inventory, scheduled receipts, and safety stock.
8. Apply lot-sizing, minimum-order, and supplier constraints.
9. Offset planned receipts by lead time to determine planned order releases.
10. Check supplier, labour, machine, warehouse, transport, and cash capacity.
11. Release, monitor, and revise orders using actual demand and supply information.

Materials Requirements Planning uses the following conceptual relationship:

Available supply must be defined carefully and may include projected inventory, scheduled receipts, and other valid supplies. A bill of materials must have correct quantities, units of measure, revision status, scrap assumptions, and effectivity dates. Errors in these master-data elements can produce apparently precise but incorrect purchase recommendations.

Lot-for-lot and fixed-order quantities

Lot-for-lot ordering plans an order approximately equal to the net requirement, reducing inventory but potentially increasing order frequency. Fixed-order quantities use a standard quantity, often because of supplier packaging, economic considerations, or process constraints. The selected rule should reflect total cost and operational practicality rather than convenience alone.

10. Illustrative Case: Sunrise Electric Kettle Components

This is an illustrative case. Sunrise Appliances assembles electric kettles. Demand for the finished kettle is forecast monthly, but the purchasing department must plan heating elements, switches, plastic housings, cartons, and fasteners.

Background and problem

The company used last year’s demand without adjustment. During the previous peak season, heating elements arrived late, while plastic housings accumulated in excess. Production schedules were repeatedly changed, suppliers received urgent orders, and customer deliveries were delayed.

Analysis

The planning team identified three causes:

• the forecast ignored a planned retailer promotion;
• supplier lead time for heating elements varied between two and five weeks;
• the bill of materials and inventory records were not updated after a design revision.

The team prepared a baseline forecast using a weighted moving average, then added documented promotional information. It also created low, base, and high scenarios. The aggregate plan showed that normal weekly capacity was sufficient for the base scenario but not for the high scenario during weeks 8–10.

For heating elements, average demand was 1,000 units per week and the planned lead time was four weeks. The team established 500 units of safety stock. The reorder point was therefore:

Because actual lead time sometimes reached five weeks, procurement began collecting supplier lead-time data and negotiated a service-level commitment. The capacity review showed that assembly could meet the base scenario with limited overtime, while injection moulding required an additional shift during the promotion.

Approach and solution

The company introduced a monthly cross-functional demand review, a weekly supply review, and formal approval of forecast overrides. The master production schedule was frozen for the immediate two weeks and allowed controlled changes in later periods. MRP records were corrected, and component requirements were checked against the engineering revision.

For the high-demand scenario, the company arranged temporary subcontracting for plastic housings rather than building permanent capacity for a short promotion. Safety stock was increased only for the heating element because its supply risk and customer impact justified the working capital.

Outcome

In the next promotion, production met the base demand plan and covered most of the high scenario without emergency freight. Excess housing inventory fell because purchases were tied to the revised schedule and valid inventory records. The company still retained contingency actions for an exceptionally high scenario rather than treating the forecast as certain.

Learning points

The case demonstrates that forecasting, lead-time analysis, capacity planning, and master-data governance must operate together. A mathematically accurate forecast cannot compensate for an incorrect bill of materials, unreliable lead-time assumptions, or unavailable capacity. It also shows why scenario planning and selective safety stock are preferable to indiscriminate inventory increases.

11. Technology, Governance, and Professional Practice

Enterprise resource planning and MRP systems integrate forecasts, sales orders, bills of materials, inventory transactions, purchase orders, and capacity data. Advanced planning systems may use optimisation, probabilistic forecasting, and machine learning. These tools are useful when data are sufficiently complete and stable, but they do not remove the need for professional judgement.

Artificial intelligence can identify patterns, classify demand, detect anomalies, and generate forecast alternatives. However, users should examine data quality, explainability, bias, model drift, and the consequences of automated recommendations. A model trained on stock-constrained sales may learn historical shortages rather than true customer demand.

Governance practices should include:

• clear ownership of forecast assumptions and overrides;
• version control for bills of materials and planning parameters;
• audit trails for changes to forecasts and orders;
• separation of demand estimates from performance incentives where possible;
• protection of commercially sensitive customer and supplier information; and
• regular review of supplier, labour, and environmental risks.

Ethically, planners should avoid manipulating forecasts to obtain favourable targets, hide shortages, or shift responsibility to suppliers. Sustainability considerations may include transport emissions, obsolete inventory, responsible sourcing, packaging waste, and the social effects of abrupt workforce changes.

12. Challenges and Limitations

Forecasting and materials planning face several recurring limitations:

• structural breaks caused by new competitors, regulation, technology, or crises;
• intermittent demand with many zero-demand periods;
• product proliferation and short life cycles;
• promotional distortion and forward buying by customers;
• unreliable inventory balances and inaccurate bills of materials;
• supplier capacity constraints and variable transport times;
• conflict between financial targets and service objectives; and
• excessive confidence in automated outputs.

A strong planning process responds through shorter review cycles, scenario analysis, exception management, parameter review, supplier collaboration, and post-period learning. It is usually better to maintain a transparent imperfect plan than an apparently precise plan based on hidden assumptions.

Best-Practice Checklist

Before approving a materials plan, the planner should ask:

1. What demand data and assumptions produced the forecast?
2. Are stock-outs, promotions, returns, and cancellations treated correctly?
3. Is the planning horizon suitable for the decision?
4. Are lead times measured from actual performance rather than quotations alone?
5. Are safety stocks linked to service objectives and uncertainty?
6. Are inventory records, bills of materials, and engineering revisions accurate?
7. Can suppliers, labour, machines, warehouses, and transport support the plan?
8. What happens under low-demand and high-demand scenarios?
9. Who approved overrides and changes?
10. How will forecast and plan performance be reviewed?

Chapter Summary

Demand forecasting provides structured estimates of future requirements, while materials planning converts those estimates into feasible purchasing, production, inventory, and capacity actions. Qualitative methods are important when data are scarce or the future differs from the past. Quantitative methods such as naïve forecasting, moving averages, weighted averages, and exponential smoothing provide repeatable baselines, but their assumptions and limitations must be understood.

Forecast errors can be measured using MAD, MSE, and MAPE, with care where actual demand is zero or very small. Aggregate planning balances demand with production, workforce, inventory, overtime, subcontracting, and shortages through level, chase, or hybrid strategies. Lead-time analysis supports reorder points and safety stock, while capacity analysis determines whether a plan can actually be executed.

The practical planning sequence moves from forecast to aggregate plan, master production schedule, dependent component requirements, netting, lot sizing, time-phased order releases, and capacity review. Reliable master data, cross-functional governance, ethical judgement, scenario analysis, and continuous feedback are essential. Technology and AI can improve visibility and analysis, but they cannot replace sound data, accountable decisions, and operational knowledge.

Key Terms

Forecast: An estimate of future demand over a specified period.

Independent demand: Demand determined primarily by customers or the external market.

Dependent demand: Demand calculated from the requirements of another item or production plan.

Level: The average magnitude around which demand varies.

Trend: A persistent upward or downward movement in demand.

Seasonality: A repeating demand pattern associated with a calendar period.

Qualitative forecasting: Forecasting based mainly on judgement and expert information.

Quantitative forecasting: Forecasting based on historical data and mathematical methods.

Moving average: An average of recent observations used to forecast the next period.

Exponential smoothing: A forecasting method that updates a previous forecast using the latest observation.

Forecast error: The difference between actual demand and forecast demand.

MAD: Mean absolute deviation, an average absolute forecast error in demand units.

Aggregate plan: A medium-term plan for product families and major resources.

Lead time: Elapsed time between initiating and completing an activity.

Safety stock: Inventory held to protect against demand or supply uncertainty.

Reorder point: Inventory position at which replenishment is initiated.

Capacity: The amount of work an operation can perform in a specified period.

Bill of materials: A structured list of components and quantities needed to make an item.

MRP: Materials Requirements Planning, a system for calculating time-phased dependent requirements.

Inventory position: On-hand inventory plus valid on-order supply minus applicable backorders.

Review Questions

Short Answer Questions

1. Distinguish between independent and dependent demand.
2. State two advantages and two limitations of qualitative forecasting.
3. What is the effect of increasing the number of periods in a simple moving average?
4. Define lead time and identify two components that may contribute to it.
5. Why must capacity be checked before releasing a materials plan?

Long Answer Questions

1. Explain the role of forecasting at strategic, aggregate, and operational planning levels.
2. Compare level, chase, and hybrid aggregate-planning strategies.
3. Explain how a master production schedule is converted into a materials plan through MRP.
4. Discuss the assumptions and limitations of exponential smoothing.
5. Explain the importance of master-data accuracy in materials planning.

Analytical / Application Questions

1. Demand in the last three months was 1,200, 1,050, and 1,350 units. Calculate the three-month moving-average forecast and explain one limitation of using it.
2. A component has average daily demand of 80 units, a lead time of 6 working days, and safety stock of 150 units. Calculate the reorder point and state the unit-consistency assumption.
3. A work centre has 2,400 hours of effective capacity available next month, while the proposed plan requires 2,650 hours. Recommend two actions and discuss one operational risk of releasing the plan unchanged.

Multiple Choice Questions with Answers

Question 1

Which statement best describes dependent demand?

A. Demand created directly by final customers

B. Demand calculated from the requirements of another item

C. Demand that occurs only during seasonal periods

D. Demand estimated exclusively by a sales manager

Correct Answer: B. Demand calculated from the requirements of another item

Question 2

A three-period moving average uses demands of 600, 750, and 900 units. What is the forecast for the next period?

A. 700 units

B. 750 units

C. 800 units

D. 850 units

Correct Answer: B. 750 units

Question 3

In exponential smoothing, what generally happens when α is increased?

A. The forecast responds more rapidly to recent observations

B. The forecast ignores the latest observation

C. The forecast becomes independent of historical data

D. The forecast must become seasonal

Correct Answer: A. The forecast responds more rapidly to recent observations

Question 4

Which measure expresses average forecast error in the original demand units?

A. MAPE

B. MSE

C. MAD

D. Correlation coefficient

Correct Answer: C. MAD

Question 5

Average weekly demand is 300 units, lead time is 4 weeks, and safety stock is 200 units. What is the reorder point?

A. 1,000 units

B. 1,200 units

C. 1,400 units

D. 1,600 units

Correct Answer: C. 1,400 units

Question 6

Which aggregate-planning strategy attempts to keep production relatively stable?

A. Chase strategy

B. Level strategy

C. Emergency strategy

D. Zero-inventory strategy

Correct Answer: B. Level strategy

Question 7

Which item is normally included in inventory position?

A. Only physically available stock

B. On-hand inventory plus valid on-order supply minus applicable backorders

C. Only forecast demand

D. Only rejected and obsolete stock

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

Question 8

A work centre has 1,000 rated hours, 90% availability, and 80% performance. Ignoring other losses, what is effective capacity?

A. 720 hours

B. 800 hours

C. 900 hours

D. 1,070 hours

Correct Answer: A. 720 hours

Question 9

Which condition makes MAPE problematic?

A. Demand is measured in units

B. The forecast horizon is one period

C. Actual demand is zero or very close to zero

D. The forecast uses a moving average

Correct Answer: C. Actual demand is zero or very close to zero

Question 10

What is the principal purpose of a bill of materials in MRP?

A. To record employee attendance

B. To list the components and quantities required to make an item

C. To calculate supplier profit margins

D. To replace capacity analysis

Correct Answer: B. To list the components and quantities required to make an item