Docstoc

REL-DU-EMBA

Document Sample
REL-DU-EMBA Powered By Docstoc
					THE FERRIS COMPANY



                 Ferris Forecasting
                              May 9, 2007




                 As promising as the view from a Ferris Wheel
                                   Goal



           Build a tool that will allow our marketing
           department to accurately forecast sales
              based on our products’ attributes




THE FERRIS COMPANY
 As promising as the view from a Ferris Wheel           2
        Linear Regression – Take 1

• 2 Fundamental Mistakes
  – Sales vs. Potential Market Share
      » Using sales history to predict demand provides a skewed view,
        as it includes the affects of stockouts by the competition
      » Using Potential Market Share removes this issue

  – ‘Absolute’ vs. ‘Relative’ Attribute Measurements
      » We originally used our absolute price (e.g. $35.00), MTBF
        (17500), etc.
      » It is not our absolute price that is important, it is our price
        relative to our competitors in the market segment




     THE FERRIS COMPANY
       As promising as the view from a Ferris Wheel                       3
                       Linear Regression

• Minitab helps us understand most important
  attributes affecting potential market share:

   –   Expected Market Share (Segment Demand / # of Products in Segment)
   –   Relative Price (Our Price minus Expected Average Price)
   –   Relative MTBF (Our MTBF minus Expected Average MTBF)
   –   Relative Awareness (Our Awareness minus Expected Average
       Awareness)


• Though other factors are at play, statistics show that
  these are the attributes that have the bulk of the
  effect on demand


       THE FERRIS COMPANY
         As promising as the view from a Ferris Wheel              4
               Linear Regression




THE FERRIS COMPANY
 As promising as the view from a Ferris Wheel   5
                 Forecasting Tool

                                     Known – From Simulation

                                    Assumptions – Must Predict Competitor Behavior


                                     Known – From our Product Attributes

                                     Results – From Regression & Assumptions


                                     Error – How good were our assumptions ?



                 Underestimated competitor’s price cuts




THE FERRIS COMPANY
 As promising as the view from a Ferris Wheel                                  6
                                     Process

• Update Regression after each round
   – Update coefficients based on total rounds to date
• Look for new factors that are beginning to affect
  demand
   – If the current factors converge for all products in the segment, other
     factors will more heavily impact demand
• Learn from error
   – Did we make a bad assumption on competitor’s price / MTBF / etc. ?




     THE FERRIS COMPANY
        As promising as the view from a Ferris Wheel                 7
                          Key Learnings

• Forecasting is ‘Art’ as well as ‘Science’ … can never
  perfectly predict competitive behavior
• Must evaluate your product vs. competition (market
  conditions)
• Process was simplified for us due to our primary
  involvement in a single market segment




     THE FERRIS COMPANY
      As promising as the view from a Ferris Wheel   8
                               Questions?


THE FERRIS COMPANY



        For more information, contact any member of our Ferris Family:

                  David Domnisch            Shannon Koerber
                  (302) 999-3240            (302) 695-1598

                  Kristen Falcone           Bill Potts
                  (302) 992-2195            (302) 992-2164




                   As promising as the view from a Ferris Wheel

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:0
posted:5/8/2013
language:English
pages:9