Outside Sales Projections by uki18959


Outside Sales Projections document sample

More Info
									Forecasting - Chapter 4
Forecasting at Tupperware
    Monthly, quarterly, and 12-month sales
    85% of sales outside of US
    Use all techniques discussed in text
    Three factors - # of dealers, % of active dealers,
    sales/active dealer
    p77 just like Outdoor Technologies

¨What is Forecasting?
    ¨*Process     of predicting a future event
        ¨Order   inventory without knowledge of future

    ¨Need   to make good estimates
    ¨Take  historical data and project them into
    the future
    ¨Many  different methods, seldom one
    clearly superior method, seldom perfect
    ¨Underlying      basis of all business decisions
        ¨Production, Inventory, Personnel,   Facilities
¨ *Types of Forecasts by Time Horizon

    ¨   Short-range - Up to 1 year; usually < 3 months
         ¨   Job scheduling, worker assignments

    ¨   Medium-range - 3 months to 3 years
         ¨   Sales & production planning, budgeting

    ¨   Long-range forecast
         ¨ 3+ years
         ¨ New product planning, facility location

¨ *Medium/long range - more comprehensive issues,
  more qualitative methods.
¨ *Short-term - tend to be more accurate, more
¨ Forecasting ability has improved, but it has been
  outpaced by an increasingly complex world
¨ Forecasting is influence by Product Life Cycle as
  product passes through stages
¨   *3 Types of Forecasts
     ¨ 1) Economic forecast, business cycle, inflation rate,
        money supply
     ¨ 2) Technological forecasts
          ¨   Predict technological change, and new product sales
     ¨   3) Demand forecasts
          ¨   Predict existing product sales

¨   *Product forecast impact
     ¨ Human Resources, hiring, training, laying off workers
     ¨ Capacity, Undependable delivery means you lose
       customers, overproduction means you incur storage

¨ Seven Steps in Forecasting, p80
     ¨   Determine the use of the forecast
     ¨   Select the items to be forecast
     ¨   Determine the time horizon of the forecast
     ¨   Select the forecasting model(s)
     ¨   Gather the data (can be difficult)*
     ¨   Make the forecast
     ¨   Validate and implement results

     ¨   Feed back loop
¨ Forecasts are seldom perfect, many outside factors
  that cannot be predicted or controlled
¨ *2 Forecasting Approaches
    ¨   1) Qualitative Methods - Used when situation is
        vague & little data exist
          ¨ New products, New technology
          ¨ Involves intuition, experience
          ¨ forecasting sales on Internet

    ¨   2) Quantitative Methods - Used when situation is
        „stable‟ & historical data exist
          ¨ Existing products, Current technology
          ¨ Involves mathematical techniques
          ¨ forecasting sales of color televisions

¨ Qualitative Methods
    ¨   Jury of executive opinion- Pool opinions of high-level
        executives “Group-think”
    ¨   Sales force composite - estimates from individual
        salespersons overly optimistic
    ¨   Delphi method - Panel of experts, queried
        interactively, 3 types of people:Decision makers,
        Staff, Respondents, Reduces „group-think‟
    ¨   Consumer Market Survey - Ask the customer, “Halo
¨ Quantitative Approaches (time series models)
   ¨ Assume that the future is a function of the past.
   ¨ Use historical knowledge (evenly spaced data
     points) of the event you tying to predict

    ¨   **4 components of a time series f4.1 p83
         ¨   a) Trend Component
               ¨   Persistent, overall upward or downward pattern
         ¨   b) Seasonal Component
               ¨   Regular pattern of up & down fluctuations
               ¨   Due to weather, customs
               ¨   Occurs within 1 year
               ¨   Could be day, week, hour, months, year, quarter
         ¨   c) Cyclical Component
               ¨   Repeating up & down movements
               ¨   Due to interactions of factors influencing economy
               ¨   Usually 2-10 years duration
         ¨   d) Random Component
               ¨   Erratic, unsystematic, „residual‟ fluctuations
               ¨   Random variation or unforeseen events
               ¨   Short duration & nonrepeating

    ¨   LL Bean - time series models forecast incoming
        calls, if wrong, unanswered calls or over-
        scheduling of operators
¨ 4 time series models

    ¨   1) Naïve approach
         ¨   demand next period is = to demand this period
         ¨   provides a starting point

    ¨   2) Moving averages, E1 p84
         ¨   use the avg. of the n most recent periods
         ¨   Makes model less sensitive, does not pick up
             trends as well
         ¨   requires extensive records from the past
         ¨   Weighted Moving Average Method
               ¨   Used when trend is present
               ¨   Older data usually less important

    ¨   3) Exponential smoothing
         ¨   Requires smoothing constant ()
         ¨   Use a fraction (alpha) of the amount that you
             were off last period
               ¨   Ranges from 0 to 1, but is generally .05 to .5
               ¨   Subjectively chosen
         ¨   Very little record keeping is needed
¨   4) Trend projection, f4.4 p93
     ¨   Fits a trend line to a series of historical data points
         and then projects the line into the future
     ¨   Assumes relationship between response variable,
         Y, and time, X, is a linear function
           ¨   other options, log, squared, cubed
     ¨   Estimated by least squares method - Minimizes
         sum of squared errors
     ¨   Slope (b) and Y-Intercept (a) :
     ¨   Can only predict into the near future
     ¨   Assume that deviations around the line are
         random and normally distributed

     ¨   This is like correlation and regression on the
         next overhead but with trend projection you
         are using the variable to predict itself
¨ Correcting for Seasonal Data, p96
     ¨   1) Find average historical demand for each “season”.
     ¨   2) Compute the average demand over all seasons.
     ¨   3) Compute a seasonal index by dividing that
         season‟s historical demand (step 1) by the average
         demand over all seasons (step 2).
     ¨   4) Estimate next year‟s total demand.
     ¨   5) Adjust it by the seasonal index.
     ¨   Examples: suntan oil, snow blower, lawn mowers
         A/C units, x-mass trees, chicken wings, hotdogs, fuel
         oil, golf clubs

¨   Quantitative Approaches (Associative model)
     ¨ *Use variables that might influence or predict the
       variable you are forecasting
     ¨ *Usually consider several variables that are related to
       the quantity being predicted
     ¨ Linear Regression Model - shows linear relationship
       between dependent & explanatory (independent)
     ¨ IBM p163: Sales & advertising (not time)
     ¨ Interpretation of Coefficients, p101
          ¨   Dependent Variable (y hat), Slope (b), Y-intercept (a),
              independent variable (x)
          ¨   y-hat is what we are forecasting (predicting)
     ¨   Might have to forecast IVs to predict DV
     ¨   Glidden paint p104
¨ *Correlation, f4.10 p105
   ¨ Answers: „how strong is the linear relationship
     between the variables?‟ degree of association
   ¨ Coefficient of correlation r
   ¨ Values range from -1 to +1
   ¨ Not “cause-and-effect”
   ¨ Coefficient of determination r2
          ¨ r2 = +/- .8 or so, highly correlated
          ¨ r2 = +/- .5 or so, correlated
          ¨ r2 = +/- .2 or so, not correlated
          ¨ % of variation in the DV explained by the IV(s)

¨ Multiple Regression Analysis
   ¨ *More than one IV, betas e4.17 p106
          ¨   y hat = a + b1x1 + b2x2 + …
     ¨   the mathematics of multiple regression becomes
         quit complex - computer

¨   DV - KIA yourself or others while driving
     ¨ MADD - 50% of fatal auto accidents involve alcohol.
     ¨ 66% involve large differences in speed.
     ¨ 60-70% in rage.
     ¨ ??% involve cell phones, like a BAL of .08
     ¨ ??% Sleepy
¨   Monitoring Forecasts
     ¨ Measures how well forecast is predicting actual
     ¨ *Should be within upper and lower control limits
       f4.11 p108
         ¨ +/- 2 stdev
         ¨ Mean square error (MSE)
         ¨ Mean absolute deviation (MAD)

¨   Focus Forecasting
     ¨ the computer selects the forecast method that yielded
       the least error in the previous period

¨   Software for forecasting, SAS, SPSS, BIOMED,
    SYSTAB, Minitab, Excel, POM for Windows, Excel

¨   *Forecasting in the Service Sector
     ¨ special need for short term records for short term
     ¨ issues of holidays and calendar
     ¨ holidays, day of the week
     ¨ unusual events, weather
     ¨ Hourly demand, f4.12 p110
     ¨ Point-of-sale can track sales

To top