Static forecasting example

(2018) 7th Edition. Boston, MA: Pearson Education. 4 6/12/2019 Types of Forecasting Methods • Qualitative forecasting methods – Subjective – Rely on human judgment – Most appropriate when: • Little historical data is available • Experts have information about the market that influence forecast – May be needed when forecasting demand for the next several years in a new industry • Time series forecasting methods – Use historical demand data to make a forecast – Most appropriate when: • Basic demand pattern does not vary significantly from year to year – Simplest forecasting methods (Source) Supply Chain Management: Strategy, Planning, and Operation. Sunil Chopra. (2018) 7th Edition. Boston, MA: Pearson Education. Types of Forecasting Methods • Simulation forecasting methods – Imitate customer choice to create forecast – Use simulation to combine: • Time series • Causal methods to answer questions such as: • What will be the impact of a price promotion? • What will be the impact of a competitor opening a store near by? – Example • Airlines simulate customer buying behavior to forecast demand for higher-fare seats when no lower-fare seats are available (Source) Supply Chain Management: Strategy, Planning, and Operation. Sunil Chopra. (2018) 7th Edition. Boston, MA: Pearson Education. Types of Random Element in Forecasting • Systematic component – Measures the expected value of demand – Consists of level, trend, and seasonality • Level: current deseasonalized demand (i.e., demand at reference time) • Trend: rate of growth or decline in demand for the next period • Seasonality: predictable seasonal fluctuations in demand • Random component – Part of forecast that deviate from systematic component – Only size and variability of random component can be predicted (not direction) – Size and variability of random component provide measure of forecast error (Source) Supply Chain Management: Strategy, Planning, and Operation. Sunil Chopra. (2018) 7th Edition. Boston, MA: Pearson Education. Types of Forecasting Methods • Causal forecasting methods – Assume that demand forecast is highly correlated with certain factors in the environment • State of economy • Interest rates • Etc. – Thus, • Find correlations (relationships) between demand and environmental factors, • Estimate future states of environmental factors, and • Use estimates of these factors to forecast future demand – For example, • Product demand is strongly correlated with product price • Thus, company can estimate impact of price promotion on product demand (Source) Supply Chain Management: Strategy, Planning, and Operation. Sunil Chopra. (2018) 7th Edition. Boston, MA: Pearson Education. Components of Demand • Among four types of forecasting methods, we focus on timeseries methods • Time-series methods are most appropriate when future demand is related to: – Historical demand – Growth patterns – Seasonal patterns • No matter which forecasting method is used, there is always: – Random element that cannot be explain only from historical demand patterns • Thus, observed (actual) demand can be broken down into two components 𝑂𝑏𝑠𝑒𝑟𝑣𝑒𝑑 𝑎𝑐𝑡𝑢𝑎𝑙 𝑑𝑒𝑚𝑎𝑛𝑑 𝐷 = 𝑠𝑦𝑠𝑡𝑒𝑚𝑎𝑡𝑖𝑐 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡 𝑆 + 𝑟𝑎𝑛𝑑𝑜𝑚 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡 (𝑅) (Source) Supply Chain Management: Strategy, Planning, and Operation. Sunil Chopra. (2018) 7th Edition. Boston, MA: Pearson Education. Types of Random Element in Forecasting • Objective of forecasting – Filter out the random component (noise) and – Estimate the systematic component • Forecast error measures difference between forecast and actual demand 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 𝑒𝑟𝑟𝑜𝑟 = 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 − 𝑂𝑏𝑠𝑒𝑟𝑣𝑒𝑑 𝑎𝑐𝑡𝑢𝑎𝑙 𝑑𝑒𝑚𝑎𝑛𝑑 • Good forecasting method should have small forecast errors, that is: "𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡"