Forecasting

1
Forecasting

• By Steve Quiat CPIM, CSCP
• Hunter Douglas, Inc.

2
Who Am I Steve Quiat

• Work as a Business Process Analyst for Hunter
  Douglas, Inc.
• Worked in manufacturing for over 35 years.
• Worked in production and materials planning for
  20 years.
• Taught classes on Production Planning,
  Purchasing, MRP, SOP, and other functions.
• APICS certified CPIM, CSCP
• Email steve.quiat_at_hunterdouglas.com

3
Todays Discussion Agenda

• Purpose of Forecasting
• Prerequisite to Forecasting
• Forecasting Definition and Assumption
• Forecast Methods
• Statistical Forecasting Concepts
• SOP Process

4
Purpose of Forecasting

• Provide basis for an operations and procurement
  business plan.

5
Forecast Prerequisites

• Develop Inventory Management Strategies
• Requires extensive attention.
• Different strategies for different inventory
  categories.
• Examples
• Good strategies should simultaneously reduce
  inventory and improve service levels.
• Forecasting is not appropriate in some cases.

6
Forecast Prerequisites

• SOP Process
• Formalize your process for Sales and Operations
  Planning
• Key players upper management must participate
• Discuss forecast deviations and resulting
  adjustments.

7
Forecast Definition and Assumption

• A forecast is a prediction of the future.
• By definition, the forecast is always wrong.
• If the forecast is always wrong, why do it?

8
Forecast Methods

• Aggregate customer forecasts.
• Consensus qualitative and quantative inputs.
• Statistical forecast
• Last Period.
• Average.
• Simple Moving Average (smoothed average).
• Weighted Moving Average.
• Exponential Smoothing (and modified)
• Delphi Method Multiple rounds of consensus of
  experts.
• Market Research

9
Exponential Smoothing

• Basic Concept
• Weighted average of the previous forecast and
  and previous actual consumption. By weighting
  one more than the other, you rely more on the
  most recent period, or all previous periods.
• The raw data sequence is often represented by
  xt, and the output of the exponential smoothing
  algorithm is commonly written as st, which may
  be regarded as a best estimate of what the next
  value of xwill be. When the sequence of
  observations begins at time t  0, the simplest
  form of exponential smoothing is given by the
  formulas1
• where a is the smoothing factor, and 0 lt a lt 1.
• Source Wikipedia

10
Modifications To Exponential Smoothing

• Seasonality
• Trend
• Others

11
Summary

• Purpose of Forecasting
• Business Plan
• Prerequisites to Forecasting
• Good Inventory Strategies
• Formal SOP Process
• Forecasting Definition and Assumption
• Predition of Future
• Forecast is always wrong (measure the error)
• Forecast Methods
• Statistical Forecasting Concepts

12
Thank You for Coming.