Ch2: Forecasting and Demand Planning

forecasting

developed through data analysis and judgement, estimates future demand

demand planning

combines statistical forecasting techniques and judgement to construct demand estimates for products or services

demand

need for a particular product or component

independent demand

demand for an item unrelated to the demand for other items

ex) finished product, spare part, service part

dependent demand

demand for an item directly related to other items or finished products

ex) component, material

short-term forecasting

less than 3 months, used mainly for tactical decisions

medium-term forecasting

3months-2yrs , used to develop a strategy over the next 6-18months

long-term forecasting

used to detect general trends and identify major turning points

forecasting error

goal of forecasting and D.P. is to minimize forecast error, hope for consistently accurate as possible

qualitative forecasting

opinion and intuition

quantitative forecasting

math models and historical data

jury of exec opinion Qual. FC. model

management + execs panel create estimates

delphi method

same as jury of exec opinion but each person’s input collected separately so no outside influence

historical analogy

identify a sales history comparable to a present situation

customer survey

directly approached and asked for opinions about particular product

time-series

based on assumption that future is an extension of the past

• most frequently used

cause and effect

assumes that 1 or more factors (independent variables) predict future demand

• 2 models: simple linear regression + multiple linear regression

naive forecast

last period’s actuals are used as this period’s forecast without adjustment

moving average

as time progresses and each month becomes actual, the oldest month is dropped and the newest month is added

weighted moving average

puts more weight on recent data and less on past data through a weighing factor

exponential smoothing

requires 3 parts: last period’s actual demand, previous period’s forecast, and a smoothing factor

linear trend forecast

best-fit line across demand data of an entire time series

simple linear regression

relationship between a single independent variable and a dependent variable (demand) by fitting a linear equation to observed data

multiple linear regression

between 2 or more independent variables

bullwhip effect

small shifts in end-consumer demand cause progressively larger fluctuations in demand forecasts as you move further up the supply chain