Demand Planning and Forecasting Session 5

1
Demand Planning and ForecastingSession 5

• Selecting and Monitoring a Forecast System
• By
• K. Sashi Rao
• Management Teacher and Trainer

2
Choice of Forecasting Method
3
Key Methodology Issues

• Choice of method/model determined by forecast
  need and purpose
• Accuracy level required for predictions and
  trade-offs involved
• Extent to which chosen model assumes past
  relationships and patterns which are expected
  to remain so in future
• Appropriate forecasting horizon that reflects
  current lead time for capacity creation
• Need to match existing business patterns(
  constant, trend, seasonal et al) with selected
  forecasting model

4
Developing the Forecasting Logic
Identify PURPOSE -use -horizon -data needed
Idnetify
/////identify
STARTSSTT
Identify TECHNIQUE -collect/analyze past
data -select suitable model
START
Develop
Develop FORECASTING LOGIC -establish model
parameters -build the model
NO
STOP
Test
Test MODEL ADEQUACY -test using historical data
STOP
MODEL SATISFACTORY ??
YES
5
Basic Approach toForecasting

• Understand the objectives of forecasting
• Integrate demand planning and forecasting
• Identify major factors that influence the demand
  forecast
• Understand and identify customer segments
• Determine the appropriate forecasting technique
• Establish performance and error measures for the
  forecast

6
Developing a Forecasting System

• Three step process involves
• Developing a forecasting logic by identifying
  the purpose, data, models and methods to be used
• Establishing a control mechanism to obtain
  reliable forecasts
• Incorporating managerial considerations in using
  the forecasting system

7
Establishing a Forecast System

•          Importing historical sales data
•          Creating statistical forecasts
•          Importing customer forecasts
•          Collaborating with customers
•          Managing forecasts
•          Building consensus forecasts
•          Supply and demand collaboration
•          Securing constrained forecasts
•          Confirmation with customers
•          Re-examining data and adjusting
  planning accordingly.
•  

8
Features of a Good Forecast

• Accuracy- with small errors
• Unbiased- so as not to overestimate or
  underestimate demand
• Responsiveness to frequent changes in demand(
  high alpha choice)
• Stability- strike balance with responsiveness(
  low alpha choice)
• Accommodative of odd unusual figures
• Timeliness to meet its purpose
• Cost-effective- forecasts benefits more than
  costs
• Easiness - to understand and operate
• Sustaining- useful for reasonable time before
  needed changes

9
Forecasting Principles

• Unless method is 100 accurate( very rarely so!),
  the chosen one should be simple to understand and
  use by all concerned at all levels
• Every forecast should state the estimate of
  likely error as /- error ( as measure of
  accuracy)
• Individual items best forecast for shorter
  periods with only as groups/families of items for
  longer periods
• Managerial judgments should be used, if
  necessary, over forecasts generated by any method
• Forecasts are not an end in themselves, but
  done for a purpose, and should finally be useful
  for decision-making

10
Forecasting in Practice

• Recognize that forecasting is an essential part
  of business planning
• Involve all those who can contribute to building
  good forecasts
• Actively collaborate with all above in building
  forecasts
• The value of data depends on where you are in the
  supply chain
• Be sure to distinguish between demand and sales
  forecasting
• Ensure that forecasting models are periodically
  reviewed for their utility of purpose

11
Forecasting Issues in Small/New Ventures

• They also need forecasts to plan and run their
  businesses
• But they dont have data-rich environments
• Forecasting for new/constantly evolving products
  more challenging
• There may not be much historical/time series data
• They lack the resources and/or capability to
  develop and use advanced forecasting techniques
• Should make do with external data/forecasting
  resources
• Will still have to develop own practical
  forecasting solutions conscious of forecast
  inaccuracies

12
Managerial Issues in Forecasting
COST
D
TIMTI
COST- BENEFIT
KEY INFERENCES
DATA AVAILABILTY
TIME FRAME
G

• GETTING STARTED
• Choice of model
• Estimation of parameters

NEW COMPETITION
USING THE SYSTEM -how to incorporate external
information -stability versus responsiveness
SALES PROMOTION
CHANGING THE SYSTEM -parameter re-estimation
versus model change
CH
FORECAST RELIABILITY
13
Some Last Words

• The future is not knowable in the sense of
  exact predictions. Life is filled with
  surrealistic surprise. Even the seemingly
  hardest models and data are frequently based
  on soft assumptions Alvin Toffler, Future
  Shock
• It is far better to foresee into the future
  without certainty than not to foresee at all-
  Henry Poincare

14
Thank you

• All the very best