Post on 05-Apr-2018
7/31/2019 Wk4 Forecasting
1/22
7/31/2019 Wk4 Forecasting
2/22
Marketing.. Demand Forecasting, Marketshare, Trend in pricesOperations.. Material requirements,Material and Labour costs, Idle time,Inventory, Defective partsFinance.. Cash flows, Expenses,Revenues, Costs
Personnel.. Labour Turnover,Absenteeism..
7/31/2019 Wk4 Forecasting
3/22
Sources of data Sales force estimates.Point of sales (POS) data systems.Trade /Association Journals.Economic Surveys and indicators.
7/31/2019 Wk4 Forecasting
4/22
Coordinate and Control all the sources of demand to- Use Production system efficiently- Delivery on Time
Types of Demand- Dependent Demand : part of a product
- Independent Demand : either influenceDemand or respond to Demand
7/31/2019 Wk4 Forecasting
5/22
Demand Management
A
B(4) C(2)
D(2) E(1) D(3) F(2)
Independent Demand
Dependent Demand
Independent demand is uncertain. Dependent demand is certain.
7/31/2019 Wk4 Forecasting
6/22
PATTERNS OF DEMAND
Average Demand - steady requirement TREND - LONG-RUN GENERAL MOVEMENTS INCREASING
OR DECREASING
SEASONAL - RECURRENT AND PERIODIC EVERY 12
MONTHS CYCLICAL - CAUSED BY ECONOMIC EXPANSIONS AND
CONTRACTIONS, TECHNOLOGICAL, DEMOGRAPHICAL, ETC.VERY HARD TO FORECAST
RANDOM - NO DISCERNIBLE PATTERN
7/31/2019 Wk4 Forecasting
7/22
Demand Forecasting
Qualitative analysis Quantitative analysis
Customer survey
Sales forcecomposite
Executiveopinion
Delphimethod
Past analogy
Time seriesanalysis
Causalanalysis
Forecast by linear regression
Simplemovingaverage
Simpleexponentialsmoothing
Trend analysis
Holts doubleExponentialsmoothing
Winters tripleExponentialsmoothing
7/31/2019 Wk4 Forecasting
8/22
Forecasting methods
Qualitative. . Where no data available, useful for new products. Quantitative
Time series .. Based on past data, use of computers for fasterprocessing Causal.. Based on factors influencing demand e.g.
Advertisement, quality, competition, economic factors, Govtpolicies..
7/31/2019 Wk4 Forecasting
9/228
Forecasting...Qualitative Methods
Grass Roots - Talk to Sales Force, Talk toCustomersMarket Research - Surveys of Customers,Experimental Test Markets
Panel Consensus - Bring in ExpertsExecutive Judgment - Surveys or Formal Input fromExecutivesHistorical Analogy - For New Products/ Technology,Find a Similar ProductDelphi Method - Formal, Sequential Method of Polling and Pooling Expert Opinions
7/31/2019 Wk4 Forecasting
10/22
Forecasting methods... Time series analysis
Simple Moving average : Neither growing nor declining, w/oseasonal characteristics, e.g.items in inventory
Weighted Moving Average : More weightage to recentpast.. .5/.3/.2, sales in a dept stores
Simple exponential smoothing :Accurate, easilyunderstandable, less computation, retail/wholesaletrade,services
Regression Analysis: useful for long-term forecasting of family of products, past and future fall in a straight line
7/31/2019 Wk4 Forecasting
11/22
Casual :
The demand for product or service is dependent ondifferent factors or variables like price, quality,availability of substitute and/or complementary
products/ services, income levels of customers,
number of competitors, etc.A causal method evaluates the relationship betweendifferent variables and their influence on each other.Causal methods include linear regression and
multiple regression analysis.
7/31/2019 Wk4 Forecasting
12/22
7/31/2019 Wk4 Forecasting
13/22
18
In-Class Exercise
Week Demand
1 11
1 11
1 11
1 11
1 11
1 11
1 11
Develop 3-week and 5-week moving average
forecasts for demand. Assume you only have 3
weeks and 5 weeks of
actual demand data for the respective forecasts
7/31/2019 Wk4 Forecasting
14/22
19
In-Class Exercise (Solution)
Week Demand -Week1 -Week1
1 111
1 1111 111
1 111 .11111
1 111 .11111
1 111 .11111 .1111
1 111 .11111 .1111
7/31/2019 Wk4 Forecasting
15/22
20
Weighted Moving Average
F = w A + w A + w A +. ..+w At 1 t- 1 1 t- 1 1 t- 1 n t-n
w =ii=1
n
Determine the 3-periodweighted moving averageforecast for period 4.
Weights:t-1 .5t-2 .3
t-3 .2
Week Demand
1 11
1 666
1 111
7/31/2019 Wk4 Forecasting
16/22
21
Solution
Week Demand Forecast
1 111
1 666
1 111
1 .666 6
F = . ( )+.1111 ( )+. (1111 1 )1111
7/31/2019 Wk4 Forecasting
17/22
22
In-Class Exercise
Determine the 3-periodweighted moving averageforecast for period 5.
Weights:t-1 .7
t-2 .2t-3 .1
Week Demand1 11
1 11
1 11
1 11
7/31/2019 Wk4 Forecasting
18/22
23
Solution
Week Demand Forecast
1 111
1 111
1 111
1 111
1 11
7/31/2019 Wk4 Forecasting
19/22
7/31/2019 Wk4 Forecasting
20/22
Causal Models for
forecastingA set of independent variables are identified and associatedwith the dependant variable through a functionalrelationshipIn general the forecast for a dependant variable Y using nindependent variables X1, X2, X3, , Xn involvesdeveloping a functional relationship as follows:
Y = f(X1, X2, X3, , Xn)There are several computer packages available today suchas SPSS to help the forecast designer in this process
M f F i
7/31/2019 Wk4 Forecasting
21/22
21
Measures of ForecastingAccuracy
A forecasting error is the differencebetween the forecasted demand for aparticular period and the actual demandin that period.
To determine how well the forecastsfrom a forecasting model fit with theactual demand pattern, the averageerror of the model is calculated.
7/31/2019 Wk4 Forecasting
22/22
Rest in the next