Day 1: Importance of Demand Forecasting
1.1 What is Demand Forecasting?
Demand forecasting is a critical process in supply chain management where businesses predict
future customer demand for a product or service. Accurate forecasting helps optimize
production schedules, inventory levels, and supply chain operations. The ability to predict
demand enables companies to avoid stockouts, reduce holding costs, and plan their resources
efficiently.
1.2 Why is Demand Forecasting Important?
● Optimized Inventory Management: Forecasting demand allows businesses to maintain
the right inventory levels. Overstocking leads to higher storage costs, while
understocking can result in missed sales.
● Efficient Production Planning: Forecasting demand helps manufacturers plan their
production schedules, allocate resources, and optimize labor costs.
● Reduced Costs: With accurate forecasts, businesses can reduce the costs of expedited
shipping, production delays, and markdowns from unsold inventory.
● Improved Customer Satisfaction: Meeting customer demand by having products
available at the right time leads to improved customer experience and loyalty.
● Strategic Decision-Making: Good forecasts help businesses make long-term strategic
decisions, such as expansion plans or investing in new technology.
1.3 Types of Forecasting Needs
● Short-Term Forecasting: Focuses on immediate future (days, weeks). Used for
operational decisions such as production planning and inventory management.
● Medium-Term Forecasting: Focuses on a few months to a year. Used for sales
promotions, product lifecycle planning, and workforce planning.
● Long-Term Forecasting: Focuses on the horizon of several years. Used for strategic
planning, new product development, and capital investments.
Day 2: Types of Forecasting Methods
2.1 Qualitative Forecasting Methods
Qualitative methods are based on subjective judgment, often used when historical data is
unavailable, and future demand is uncertain. These methods rely on the expertise and intuition
of individuals or groups to generate predictions.
2.2 Common Qualitative Methods
● Expert Opinion: Involves consulting industry experts or internal personnel with in-depth
knowledge about the market or product. For example, a tech company might consult with
R&D engineers to forecast demand for a new gadget.
○ Exercise: Suppose a company is launching a new product and wants to forecast
initial demand. Organize a panel of experts (e.g., engineers, marketing
professionals) to estimate the potential sales for the next three months.
● Market Research: This includes surveys, focus groups, or customer interviews to gauge
demand. For example, a company might send out surveys to determine interest in a new
service.
○ Exercise: Conduct a small market survey (either real or simulated) to gauge
customer interest in a new product or service. Based on the survey results,
forecast the potential demand for the next quarter.
● Delphi Method: A structured process where experts provide individual forecasts, then
revise them after receiving anonymous feedback from the group until a consensus is
reached.
○ Exercise: Organize a Delphi-style exercise by asking several colleagues to
independently forecast demand for a product, followed by a round of feedback
and revision.
● Executive Opinion: Senior management's collective judgment based on their
knowledge of the market and product history.
○ Exercise: Ask senior executives in a company to provide a demand forecast for
an upcoming quarter based on their experience and market knowledge. Compare
these opinions to see the range of forecasts.
2.3 When to Use Qualitative Methods
● When launching new products with little historical data.
● During periods of significant market disruption (e.g., a pandemic or a technological
breakthrough).
● In emerging markets where consumer preferences are unknown or constantly changing.
Day 3: Quantitative Forecasting Methods
3.1 What Are Quantitative Forecasting Methods?
Quantitative forecasting methods rely on historical data to predict future demand. These
methods use statistical techniques to identify patterns and trends in past data, which can then
be used to make informed predictions.
3.2 Moving Average (MA)
● Definition: Moving average smooths out fluctuations in historical data to predict future
demand. It averages data over a specific number of periods, typically 3, 6, or 12 months,
and uses this average as a forecast for the next period.
Example: A retailer forecasts monthly sales of a product using a 12-month moving average. If
the demand for the last 12 months is:
150, 160, 155, 180, 175, 200, 210, 215, 230, 220, 240, 250
● The 12-month moving average would be:
MA=150+160+155+180+175+200+210+215+230+220+240+25012=201.25MA =
\frac{150 + 160 + 155 + 180 + 175 + 200 + 210 + 215 + 230 + 220 + 240 + 250}{12} =
201.25
Exercise: Calculate the 6-month moving average for a product given the following demand
data:
Month 1: 100, Month 2: 120, Month 3: 110, Month 4: 150, Month 5: 180, Month 6: 160
●
3.3 Exponential Smoothing
● Definition: Exponential smoothing gives more weight to recent data while still
considering past observations. It uses a smoothing constant (α\alpha) to determine the
weight given to the most recent observation.
Formula:
Ft+1=αDt+(1−α)FtF_{t+1} = \alpha D_t + (1 - \alpha) F_t
Where:
○ Ft+1F_{t+1} is the forecast for the next period,
○ DtD_t is the actual demand at time tt,
○ FtF_t is the forecast for time tt,
○ α\alpha is the smoothing constant (0 < α\alpha < 1).
● Example: A company has an actual demand of 500 units in the last period. The forecast
for that period was 450 units, and α=0.2\alpha = 0.2. The new forecast will be:
Ft+1=0.2(500)+0.8(450)=460F_{t+1} = 0.2(500) + 0.8(450) = 460
Exercise: Apply exponential smoothing with α=0.3\alpha = 0.3 and the following data:
Last period's demand (D): 300 units, Last period's forecast (F): 280 units
●
Day 4: Forecasting Metrics
4.1 Importance of Forecasting Metrics
Forecasting metrics help assess the accuracy of demand forecasts. These metrics can help
identify areas for improvement and give a more objective view of forecast performance.
4.2 Mean Absolute Deviation (MAD)
● Definition: MAD measures the average absolute difference between forecasted and
actual demand.
Formula:
MAD=∑t=1n∣Dt−Ft∣nMAD = \frac{\sum_{t=1}^{n} |D_t - F_t|}{n}
Where:
○ DtD_t is the actual demand at time tt,
○ FtF_t is the forecasted demand at time tt,
○ nn is the number of periods.
Exercise: Calculate the MAD for the following data:
Actual demand: 150, 160, 180, 170
Forecasted demand: 140, 155, 175, 165
●
4.3 Mean Squared Error (MSE)
● Definition: MSE gives more weight to larger errors by squaring the forecast errors. This
makes it more sensitive to outliers.
Formula:
MSE=∑t=1n(Dt−Ft)2nMSE = \frac{\sum_{t=1}^{n} (D_t - F_t)^2}{n}
● Exercise: Calculate the MSE for the same data above.
4.4 Root Mean Squared Error (RMSE)
● Definition: RMSE is the square root of the MSE and provides an error measure in the
same units as the demand.
● Exercise: Calculate the RMSE using the data from the previous exercise.
Day 5: Practical Applications
5.1 Case Study: Retailer Demand Forecasting
A retail company is forecasting sales for the upcoming holiday season. The company uses a
combination of quantitative forecasting (moving averages) and qualitative forecasting (expert
opinion) to predict demand.
● Exercise: Practice by creating a forecast for a retail company using a 6-month moving
average for their past sales data. Incorporate expert opinions to adjust the forecast for
expected seasonal spikes.
Day 6: Forecasting in the Real World
6.1 Case Study: Electronics Manufacturer
A smartphone manufacturer uses exponential smoothing to forecast demand based on historical
sales and anticipated consumer trends.
● Exercise: Forecast demand for the next quarter based on historical demand and a given
smoothing constant.
Day 7: Challenges in Demand Forecasting
7.1 Common Challenges
● Data Inaccuracy: Forecasts are only as good as the data used. Poor quality or
incomplete data will result in inaccurate predictions.
● External Disruptions: Unexpected events (e.g., natural disasters, pandemics) can
drastically alter demand patterns.
● Exercise: List some methods to overcome data quality issues and disruptions in
forecasting, considering your case studies.
