© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Operations Management Forecasting Chapter 4

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J What is Forecasting?  Process of predicting a future event  Underlying basis of all business decisions  Production  Inventory  Personnel  Facilities Sales will be $200 Million!

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Types of Forecasts  Economic forecasts  Address business cycle, e.g., inflation rate, money supply etc.  Technological forecasts  Predict rate of technological progress  Predict acceptance of new product  Demand forecasts  Predict sales of existing product

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J  Short-range forecast  Up to 1 year; usually less than 3 months  Job scheduling, worker assignments  Medium-range forecast  3 months to 3 years  Sales & production planning, purchasing, budgeting  Long-range forecast  3 + years  New product planning, facility location Types of Forecasts by Time Horizon

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J What Do We Forecast - Aggregation Clustering goods or services that have similar demand requirements and common processing, labor, and materials requirements: Red shirts White shirts Blue shirts Big Mac Quarter Pounder Regular Hamburger Shirts Pounds of Beef $ $$ $ Why do we aggregate? What about units of measurement ?

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Realities of Forecasting  Forecasts are seldom perfect  Most forecasting methods assume that there is some underlying stability in the system  Both product family and aggregated product forecasts are more accurate than individual product forecasts

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J  Persistent, overall upward or downward pattern  Due to population, technology etc.  Several years duration Trend Component Time Response

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J  Regular pattern of up & down fluctuations  Due to weather, customs etc. Time Response Seasonal Component Summer

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J  Repeating up & down movements  Due to interactions of factors influencing economy  Usually 2-10 years duration Time Response Cyclical Component Cycle

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Product Demand Year 1 Year 2 Year 3 Year 4 Actual demand line Demand for product or service Seasonal peaksTrend component Average demand over four years Random variation

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecasting Approaches  Used when situation is ‘stable’ & historical data exist  Existing products  Current technology  Involves mathematical techniques  e.g., forecasting sales of color televisions Quantitative Methods  Used when situation is vague & little data exist  New products  New technology  Involves intuition, experience  e.g., forecasting sales on Internet Qualitative Methods

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Overview of Qualitative Methods  Jury of executive opinion  Pool opinions of high-level executives, sometimes augmented by statistical models  Delphi method  Panel of experts, queried iteratively  Sales force composite  Estimates from individual salespersons are reviewed for reasonableness, then aggregated  Consumer Market Survey  Ask the customer

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Overview of Quantitative Approaches  Naïve approach  Moving averages  Exponential smoothing  Trend projection  Seasonal variation  Linear regression Time-series Models Associative Models

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J  Set of evenly spaced numerical data Observing the response variable at regular time intervals  Forecast based only on past values Assumes that factors influencing the past and present will continue to influence the future  Example Year: Sales: What is a Time Series?

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Naive Approach  Assumes demand in next period is the same as demand in most recent period  e.g., If May sales were 48, then June sales will be 48  Sometimes cost effective & efficient © 1995 Corel Corp.

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecastn n Demand in Previous Periods Periods Simple Moving Average F = A + A + A + A F = A + A + A t t–1 t–2 t–3 t–4 t t–1 t–2 t–3 4 3

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecast = Σ (Weight for period n) (Demand in period n) Σ Weights Weighted Moving Average F =.4A +.3A +.2A +.1A F =.7A +.2A +.1A t t–1 t–2 t–3 t–4 t t–1 t–2 t–3

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J  Form of weighted moving average  Weights decline exponentially  Most recent data weighted most  Requires smoothing constant (  )  Ranges from 0 to 1  Subjectively chosen  Involves little record keeping of past data Exponential Smoothing Method

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Exponential Smoothing F =  A + (1 –  ) (F ) =  A + F –  F = F +  (A – F ) t t–1 t–1 t–1 t–1 t–1 Forecast =  (Demand last period) + (1 –  ) ( Last forecast) Forecast = Last forecast +  (Last demand – Last forecast)

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J T t =  (Forecast this period – Forecast last period) + (1-  ) (Trend estimate last period) =  (F t - F t-1 ) + (1-  )T t-1 Exponential Smoothing with Trend Adjustment Forecast = Exponentially smoothed forecast (F ) + Exponentially smoothed trend (T ) t t

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Seasonal Variation QuarterYear 1Year 2Year 3Year Total Average Seasonal Index = Actual Demand Average Demand = = Forecast for Year 5 = 2600

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Quarter Year 1 Year 2 Year 3 Year 4 145/250 = /300 = /450 = /550 = /250 = /300 = /450 = /550 = /250 = /300 = /450 = /550 = /250 = /300 = /450 = /550 = 0.39 QuarterAverage Seasonal Index 1( )/4 = ( )/4 = ( )/4 = ( )/4 = 0.50 Seasonal Variation Forecast 650(0.20) = (1.30) = (2.00) = (0.50) = 325

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Overview of Quantitative Approaches  Naïve approach  Moving averages  Exponential smoothing  Trend projection  Seasonal variation  Linear regression Time-series Models Associative Models

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Linear Regression Independent Dependent Variables Variable Factors Associated with Our Sales Advertising Pricing Competitors Economy Weather Sales

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Scatter Diagram Sales vs. Payroll Area Payroll (in $ hundreds of millions) Sales (in $ hundreds of thousands) Regression Line Now What?

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J  Short-range forecast  Up to 1 year; usually less than 3 months  Job scheduling, worker assignments  Medium-range forecast  3 months to 3 years  Sales & production planning, budgeting  Long-range forecast  3 + years  New product planning, facility location Types of Forecasts by Time Horizon Time Series Associative Qualitative

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecast Error

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecast Error + 5 – 3 E = A – F t t t

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecast Error - CFE CFE =  E t CFE – Cumulative sum of Forecast Errors Positive errors offset negative errors Useful in assessing bias in a forecast

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecast Error - MSE MSE – Mean Squared Error Accentuates large deviations MSE =  E t n 2

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecast Error - MAD   |E t | n MAD = MAD – Mean Absolute Deviation Widely used, well understood measurement of forecast error

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecast Error - MAPE MAPE = 100   |E t | / A t n MAPE – Mean Absolute Percent Error Relates forecast error to the level of demand

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecast Error E t = A t – F t CFE =  E t 100  |E t | / A t n MAPE =   |E t | n MAD = MSE =  E t n 2

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Monitoring & Controlling Forecasts We need a TRACKING SIGNAL to measure how well the forecast is predicting actual values TS = Running sum of forecast errors ( CFE ) Mean Absolute Deviation ( MAD ) =  E  |  E | / n t t

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Plot of a Tracking Signal Time Lower control limit Upper control limit Signal exceeded limit Tracking signal CFE / MAD Acceptable range + 0 -

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecasting in the Service Sector Presents unusual challenges  special need for short term records  needs differ greatly as function of industry and product  issues of holidays and calendar  unusual events

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Forecast of Sales by Hour for Fast Food Restaurant

© 2004 by Prentice Hall, Inc., Upper Saddle River, N.J Summary  Demand forecasts drive a firm’s plans - Production - Capacity - Scheduling  Need to find the forecasting method(s) that best fit our pattern of demand – no one right tool - Qualitative methods e.g. customer surveys - Time series methods (quantitative) rely on historical demand to predict future demand - Associative models (quantitative) use historical data on independent variables to predict demand e.g. promotional campaign  Track forecast error to determine if forecasting model requires change