Learning Objectives
By the end of this lesson, learners should be able to:
- Apply common forecasting methods.
- Calculate moving averages and exponential smoothing.
- Select suitable forecasting techniques for business problems.
Naïve Forecast
The next period is assumed to equal the most recent actual value.
Example
If June sales were USD 120,000, the July forecast is USD 120,000.
Useful as a benchmark.
Moving Average
A moving average uses the average of recent periods.
Example
Sales:
- April = 100
- May = 110
- June = 120
3-month moving average for July:
(100 + 110 + 120) ÷ 3 = 110
Moving averages smooth short-term fluctuations.
Weighted Moving Average
More recent observations receive greater weight.
Example
Weights:
- June = 0.5
- May = 0.3
- April = 0.2
Forecast = (120×0.5) + (110×0.3) + (100×0.2)
= 60 + 33 + 20 = 113
This responds more quickly to recent changes.
Exponential Smoothing
Forecast formula:
New Forecast = α(Actual) + (1−α)(Previous Forecast)
Where α is the smoothing constant between 0 and 1.
Higher α gives more weight to recent observations.
Trend Projection
A trend line is fitted to historical data and extended into the future.
Commonly used for medium- and long-term planning.
Choosing A Forecasting Method
|
Situation |
Recommended Method |
|
Stable demand |
Moving average |
|
Recent changes important |
Exponential smoothing |
|
Clear long-term trend |
Trend projection |
|
Benchmark comparison |
Naïve forecast |
International Example
A consumer electronics company in Japan uses exponential smoothing for monthly smartphone sales because demand changes rapidly after new product launches.
Common Mistakes
- Using long moving averages when demand changes quickly,
- Ignoring seasonality,
- Choosing smoothing constants arbitrarily,
- Using one method for all products.
Practical Exercise
Using monthly sales data for the last six months:
- Calculate a 3-month moving average,
- Calculate a weighted moving average,
- Compare the results and explain which responds faster to recent changes.
Learning Materials / Reference Materials
- Makridakis, Wheelwright & Hyndman. Forecasting Methods and Applications.
- Excel Forecasting Functions Documentation.
Lesson Summary
Forecasting methods differ in complexity and responsiveness. Selecting the appropriate method depends on demand patterns, trend strength, seasonality, and business objectives.