Introduction
Demand forecasting is the process of estimating the quantity of goods or services that customers are likely to require during a future period. It is an essential activity in inventory management because organizations must make decisions about purchasing, production, staffing, warehousing, transportation, and financial resources before actual demand occurs.
Organizations rarely know exactly how much customers will purchase in the future. Instead, they use historical information, market conditions, customer behavior, economic indicators, seasonal patterns, sales information, and other relevant data to develop forecasts.
For example, a supermarket cannot wait until customers arrive before ordering products from suppliers. It must estimate how much milk, bread, vegetables, beverages, and other products customers will purchase over the coming days or weeks. If the supermarket underestimates demand, it may experience stockouts. If it significantly overestimates demand, it may hold excess inventory, resulting in additional storage costs and possible product deterioration.
Demand forecasting therefore provides a foundation for inventory planning. It helps organizations answer questions such as how much stock will be needed, when it will be needed, and where it will be needed.
Forecasting is not about predicting the future with complete certainty. Instead, it is about using available information systematically to reduce uncertainty and improve decision-making.
Meaning of Demand Forecasting
Demand forecasting involves estimating future customer demand for a product, service, or group of products.
A forecast may be prepared for different time periods.
A short-term forecast may cover several days, weeks, or months and is often used for operational decisions such as replenishment, workforce scheduling, and warehouse planning.
A medium-term forecast may cover several months to a few years and can support procurement, production planning, and capacity decisions.
A long-term forecast may cover several years and can support strategic decisions such as facility expansion, new product development, and long-term investment.
The appropriate forecasting period depends on the organization’s operating environment and the decision being supported.
Importance of Demand Forecasting
Demand forecasting is important because many business decisions depend on expected future demand.
Inventory managers use forecasts to determine how much inventory should be purchased or produced.
Procurement departments use forecasts to plan supplier orders.
Warehouse managers use forecasts to determine storage requirements.
Production managers use forecasts to schedule manufacturing activities.
Transport managers use forecasts to anticipate distribution requirements.
Finance departments use forecasts to prepare budgets and cash-flow plans.
Sales teams use forecasts to establish targets and evaluate expected revenue.
Therefore, demand forecasting is not only an inventory-management activity. It affects almost every part of the supply chain.
Demand Forecasting and Inventory Management
The relationship between demand forecasting and inventory management is particularly important.
Suppose a company expects demand for a product to be 10,000 units during the next six months. It can use this information to determine purchasing and replenishment requirements.
If the forecast is significantly below actual demand, inventory may be insufficient.
If the forecast is significantly above actual demand, the organization may purchase more stock than customers require.
Forecast accuracy therefore influences inventory levels, stockouts, excess stock, warehouse utilization, and purchasing costs.
Forecasting Methods
Demand forecasting methods can broadly be divided into qualitative methods and quantitative methods.
Qualitative methods rely heavily on expert judgment, market knowledge, customer opinions, and other non-numerical information.
Quantitative methods use historical numerical data and mathematical or statistical techniques.
The appropriate method depends on the availability and quality of data, the nature of demand, the forecasting period, and the level of uncertainty.
Qualitative Forecasting Methods
Qualitative forecasting is particularly useful when historical data is limited or when future conditions are expected to differ significantly from the past.
For example, a company launching a completely new product may have no historical sales data for that product.
Management may therefore rely on market research, expert opinions, customer surveys, sales-team estimates, and information from similar products.
Qualitative forecasting can also be useful when major changes in technology, regulation, customer preferences, or market conditions make historical patterns less reliable.
Expert Opinion
Expert opinion involves using knowledge from individuals who have substantial experience in a particular market or industry.
Sales managers, procurement specialists, marketing professionals, production managers, and industry analysts may contribute their expectations about future demand.
For example, an experienced sales manager may know that demand for a particular product usually increases after a specific marketing campaign.
Although expert judgment can be valuable, it can also be affected by personal bias.
For this reason, organizations may combine expert judgment with quantitative data.
Market Research
Market research involves collecting information about customers, competitors, market conditions, and purchasing behavior.
Organizations may use surveys, interviews, focus groups, customer feedback, and market studies to estimate future demand.
For example, a company planning to introduce a new smartphone may conduct market research to determine how many potential customers are interested in the product, which features they value, and what price they are willing to pay.
Market research is particularly useful for new products because historical sales data may not exist.
Sales Force Composite
The sales force composite method collects demand estimates from sales representatives.
Sales employees interact directly with customers and may therefore have valuable information about customer intentions and market changes.
For example, sales representatives may report that several major customers plan to increase their orders during the next quarter.
Management can combine these estimates to create an overall demand forecast.
The weakness of this method is that sales representatives may overestimate or underestimate demand due to optimism, sales targets, or limited information.
Delphi Method
The Delphi method involves obtaining forecasts from a group of experts through structured and often anonymous rounds of assessment.
Experts provide their estimates independently.
The responses are analyzed and shared in summarized form, after which experts may revise their estimates.
This process can continue until a reasonable level of agreement is reached.
The method is useful when forecasting complex situations where reliable historical data is unavailable.
Quantitative Forecasting Methods
Quantitative forecasting uses numerical data to estimate future demand.
These methods are particularly useful when historical demand data is available and the assumption that historical patterns contain useful information about future demand is reasonable.
Common quantitative approaches include:
- Moving averages.
- Weighted moving averages.
- Exponential smoothing.
- Trend analysis.
- Regression analysis.
- Time-series forecasting.
Moving Average
A moving average calculates an average of demand over a selected number of previous periods.
Suppose a company recorded the following monthly demand:
January = 100 units
February = 120 units
March = 140 units
A three-month moving average forecast for April would be:
(100 + 120 + 140) ÷ 3 = 120 units
Therefore, the forecast for April would be 120 units.
As new information becomes available, the oldest observation is removed and the newest observation is included.
For example, when forecasting May, the organization may use February, March, and April demand.
Moving averages are useful when demand is relatively stable.
Weighted Moving Average
A weighted moving average gives different levels of importance to previous periods.
More recent demand may receive greater weight because it may provide a better indication of current customer behavior.
Suppose demand for the last three months was:
January = 100
February = 120
March = 140
Management assigns weights of:
January = 20%
February = 30%
March = 50%
The forecast would be:
(100 × 0.20) + (120 × 0.30) + (140 × 0.50)
= 20 + 36 + 70
= 126 units
The forecast for April would therefore be 126 units.
Exponential Smoothing
Exponential smoothing is a forecasting technique that gives greater importance to recent observations while still incorporating previous forecasts.
A simplified formula is:
New Forecast = Previous Forecast + α(Actual Demand − Previous Forecast)
Where α is the smoothing constant, usually between 0 and 1.
A higher value of α makes the forecast respond more quickly to recent changes.
A lower value makes the forecast smoother and less responsive to short-term fluctuations.
Exponential smoothing is useful when organizations want a forecasting method that can respond to changing demand without completely ignoring historical information.
Trend Analysis
Trend analysis examines the general direction of demand over time.
Demand may show an upward trend, downward trend, or relatively stable pattern.
Suppose annual demand is:
2023 = 5,000 units
2024 = 5,500 units
2025 = 6,000 units
2026 = 6,500 units
The data suggests an upward trend.
Management may investigate the reason for this increase and determine whether it is likely to continue.
Trend analysis is particularly useful for identifying long-term changes in customer demand.
Seasonality
Seasonality refers to predictable demand changes that occur at particular times of the year.
Examples include:
- Increased umbrella demand during rainy seasons.
- Increased school-supply demand before school terms.
- Increased gift demand during holiday periods.
- Increased tourism-related demand during peak travel seasons.
Seasonality is important because using a simple annual average may hide significant fluctuations.
For example, a retailer selling Christmas decorations may have very low demand for most of the year and extremely high demand during November and December.
A forecast that ignores seasonality could result in serious inventory problems.
Cyclical Demand
Cyclical demand refers to longer-term fluctuations associated with economic or business cycles.
Demand may increase during periods of economic growth and decline during recessions.
For example, demand for expensive consumer goods may decrease when customers have lower disposable income.
Organizations operating in cyclical markets need to consider broader economic conditions when developing forecasts.
Random Demand Variation
Not all changes in demand can be explained by trends or seasonal patterns.
Unexpected events can cause random demand fluctuations.
Examples include sudden changes in customer preferences, unexpected competitor actions, supply disruptions, extreme weather, or unusual events.
Random variation makes forecasting more difficult because it cannot always be predicted from historical patterns.
Organizations therefore need contingency plans in addition to forecasts.
Market Analysis
Market analysis involves studying external factors that can influence future demand.
An organization should not rely exclusively on its internal sales history.
External factors may include:
- Competitor behavior.
- Economic conditions.
- Customer income.
- Population changes.
- Technology.
- Government regulations.
- Consumer preferences.
- Industry trends.
For example, a company selling petrol-powered vehicles must consider the increasing adoption of electric vehicles when forecasting long-term demand.
Competitive Analysis
Competitors can significantly affect demand.
If a competitor introduces a cheaper or higher-quality product, demand for an organization’s existing products may decrease.
Conversely, if a major competitor exits the market, the organization’s demand may increase.
Forecasting should therefore consider expected competitor behavior whenever reliable information is available.
Economic Conditions
Economic conditions can strongly influence demand.
During periods of economic expansion, consumers and businesses may increase spending.
During economic downturns, customers may reduce purchases, delay investments, or shift toward cheaper alternatives.
For example, demand for luxury products may decline significantly during economic difficulties while demand for essential products remains relatively stable.
Economic indicators can therefore provide useful information for forecasting.
Sales Forecasting
Sales forecasting is the process of estimating future sales revenue or sales volume.
It is closely related to demand forecasting but is not exactly the same.
Demand forecasting focuses on expected customer requirements.
Sales forecasting often focuses on the organization’s expected sales performance.
For example, customer demand for a product may be 10,000 units, but the company may forecast sales of only 8,000 units because of limited production capacity or expected supply constraints.
Understanding this distinction is important when developing inventory plans.
Demand Forecasting Versus Sales Forecasting
| Aspect | Demand Forecasting | Sales Forecasting |
|---|---|---|
| Main Focus | Customer requirements | Expected company sales |
| Considers Market Demand | Strongly | Yes |
| Considers Company Capacity | Not always | Usually |
| Purpose | Plan supply and resources | Plan revenue and sales |
| Example | Customers may need 10,000 units | Company expects to sell 8,000 units |
Demand forecasting is therefore broader because it attempts to understand the market requirement, while sales forecasting often focuses on what the organization expects to sell.
Demand Trends
Demand trends represent patterns or directions in customer demand over time.
A trend may be:
Increasing — demand is growing.
Decreasing — demand is declining.
Stable — demand remains relatively consistent.
Seasonal — demand changes according to predictable periods.
Cyclical — demand changes in response to longer economic or industry cycles.
Identifying the correct pattern helps organizations select appropriate forecasting methods.
Forecast Accuracy
Forecast accuracy measures how close a forecast is to actual demand.
No forecast is perfectly accurate, but organizations can monitor forecast errors to determine whether forecasting methods are performing effectively.
Suppose the forecast for a month is 1,000 units but actual demand is 1,100 units.
Forecast error is:
Actual Demand − Forecast Demand
1,100 − 1,000 = 100 units
The organization underestimated demand by 100 units.
Forecast Error
Forecast error can be positive or negative depending on the relationship between actual demand and forecast demand.
If actual demand is greater than forecast demand, the organization underestimated demand.
If actual demand is lower than forecast demand, the organization overestimated demand.
For example:
Forecast = 1,000
Actual = 900
Error:
900 − 1,000 = −100
The negative value indicates that demand was overestimated by 100 units.
Mean Absolute Deviation (MAD)
Mean Absolute Deviation, or MAD, measures the average absolute size of forecast errors.
The basic calculation is:
MAD = Sum of Absolute Forecast Errors ÷ Number of Forecast Periods
Suppose forecast errors for three months are:
January = 50
February = −30
March = 40
Absolute errors are:
50, 30, and 40.
Therefore:
MAD = (50 + 30 + 40) ÷ 3
MAD = 40 units
This means the forecast was off by an average of 40 units per period, ignoring the direction of the errors.
Mean Absolute Percentage Error (MAPE)
MAPE expresses forecast error as a percentage of actual demand.
It is useful for comparing forecasting performance across products with different demand volumes.
The basic concept is:
MAPE = Average of Absolute Percentage Errors
For example, if the organization consistently forecasts demand within 5% of actual demand, forecasting performance is generally better than if errors are consistently around 20%.
However, MAPE can present problems when actual demand is zero or very close to zero.
Forecast Accuracy and Inventory
Forecast accuracy has a major impact on inventory.
If forecasts are consistently too high, organizations may purchase excessive inventory.
This increases carrying costs and may result in obsolete stock.
If forecasts are consistently too low, organizations may experience stockouts.
Therefore, improving forecast accuracy can help reduce both excess inventory and shortages.
However, an organization should not focus only on producing the most accurate mathematical forecast. It must also consider the cost of forecast errors.
A small forecasting error for an inexpensive item may have little financial impact, while the same percentage error for an expensive product may be significant.
Demand Planning
Demand planning is the broader process of using demand information to coordinate future business activities.
It combines forecasting with decisions regarding inventory, purchasing, production, capacity, distribution, and customer requirements.
Demand planning therefore transforms forecasts into operational decisions.
For example, if the forecast indicates that demand for a product will increase by 30%, the organization may need to increase purchases, allocate additional warehouse space, schedule additional production, and arrange additional transportation capacity.
Collaborative Demand Planning
Demand forecasting can be improved when different departments share information.
Sales teams may provide information about customer orders.
Marketing teams may provide information about upcoming promotions.
Procurement teams may provide information about supplier lead times.
Warehouse teams may provide information about current inventory.
Finance teams may provide information about budgets.
When these groups collaborate, the organization can develop a more realistic demand plan.
Example: Demand Forecasting at TechNova
TechNova Electronics sells laptops.
The company records monthly laptop sales:
| Month | Actual Demand |
|---|---|
| January | 500 |
| February | 550 |
| March | 600 |
| April | 650 |
| May | 700 |
The data shows an increasing trend.
A simple three-month moving average for June would be:
(600 + 650 + 700) ÷ 3
= 650 units
However, the increasing trend suggests that a simple moving average may underestimate future demand because the most recent months have been consistently higher.
Management may therefore consider a weighted moving average that gives greater weight to recent months.
If June demand is expected to be higher due to a marketing campaign, management should also incorporate that information rather than relying solely on historical data.
This demonstrates an important principle: forecasting should combine historical data with relevant knowledge about future conditions.
Forecasting and Promotions
Marketing promotions can significantly change demand.
Suppose normal weekly demand for a product is 1,000 units.
The marketing department plans a 20% discount campaign.
Historical data from similar promotions indicates that demand usually increases by approximately 50%.
The inventory team should not simply forecast demand at 1,000 units.
A reasonable starting estimate based on the historical promotion effect could be:
1,000 × 1.50 = 1,500 units
The company may therefore need to increase inventory before the promotion begins.
Failure to incorporate promotions into forecasting can cause stockouts.
Forecasting and New Products
New products present a special forecasting challenge because historical sales data may not exist.
Organizations may use information from:
- Similar products.
- Market research.
- Customer surveys.
- Competitor products.
- Test markets.
- Expert judgment.
- Pilot sales.
For example, if TechNova launches a new laptop model, it can study historical demand for similar models and combine that information with market research.
Forecasts for new products should generally be treated as more uncertain than forecasts for established products with long and reliable demand histories.
Forecasting Challenges
Several factors can reduce forecasting accuracy.
Poor-quality historical data can produce unreliable forecasts.
Sudden market changes can make historical patterns less relevant.
New competitors can change customer behavior.
Promotions can create temporary demand spikes.
Seasonality can complicate forecasting.
Product life cycles can cause demand to increase and eventually decline.
External events such as economic crises, regulatory changes, or supply disruptions can also affect demand.
Organizations should therefore review forecasts regularly rather than treating them as permanent predictions.
Forecast Review and Adjustment
Forecasts should be reviewed as new information becomes available.
Suppose an organization originally forecasts monthly demand of 5,000 units.
After receiving new customer orders and market information, expected demand increases to 6,000 units.
The organization should update its forecast.
Forecast review is particularly important when there are significant changes in:
- Customer orders.
- Market conditions.
- Pricing.
- Promotions.
- Supplier capacity.
- Competitor behavior.
- Economic conditions.
A forecast is therefore a living planning tool rather than a fixed number.
Technology in Demand Forecasting
Modern organizations use ERP systems, warehouse management systems, business intelligence tools, and specialized forecasting software to support demand planning.
These systems can collect large amounts of historical sales and inventory data.
They can identify trends, calculate forecast errors, generate replenishment recommendations, and support scenario analysis.
However, technology does not eliminate the need for human judgment.
A system may identify a historical trend, but management must determine whether the conditions responsible for that trend are likely to continue.
Demand Forecasting and Business Central
In an ERP environment such as Microsoft Dynamics 365 Business Central, demand-related information can be connected to inventory, sales, purchasing, planning, and replenishment processes.
Historical transactions can provide valuable information about item demand.
Planning functionality can use inventory availability, expected demand, supply orders, and other planning parameters to help organizations determine future replenishment requirements.
This demonstrates the relationship between demand forecasting and broader inventory planning: forecast information becomes useful when it is converted into purchasing, production, and replenishment decisions.
Key Takeaways
Demand forecasting is the process of estimating future customer demand.
It is essential for inventory management because organizations must make purchasing, production, warehouse, and distribution decisions before actual demand occurs.
Forecasting methods can be broadly divided into qualitative and quantitative approaches.
Qualitative methods rely on expert judgment, market research, sales-force estimates, and structured expert consultation.
Quantitative methods use historical numerical data and include moving averages, weighted moving averages, exponential smoothing, trend analysis, and other statistical approaches.
Moving averages use historical demand to estimate future demand, while weighted moving averages give greater importance to selected periods.
Exponential smoothing gives greater importance to recent observations while retaining information from previous forecasts.
Demand can exhibit trends, seasonality, cycles, and random variation.
Market analysis is important because customer demand is influenced by competitors, economic conditions, technology, regulations, and changing customer preferences.
Demand forecasting and sales forecasting are related but not identical. Demand forecasting focuses on expected customer requirements, while sales forecasting often focuses on the organization’s expected sales.
Forecast accuracy measures how closely forecasts correspond to actual demand.
Forecast errors can result in either excess inventory or stockouts.
Measures such as MAD and MAPE can be used to evaluate forecasting performance.
Demand planning goes beyond forecasting by converting expected demand into coordinated decisions concerning purchasing, production, inventory, warehousing, and distribution.
Forecasts should be regularly reviewed and updated as new information becomes available.
Ultimately, effective demand forecasting does not attempt to predict the future perfectly. Instead, it provides a structured and evidence-based estimate of future demand that allows an organization to prepare its inventory, resources, suppliers, warehouse capacity, and operations before customer requirements actually occur.