Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

• Cycle

• Irregular Patterns
• Drivers of Sales
• Marketing
• Distribution
• Availability
• Pricing
• Consumer’s Capability
• Consumer’s Behavior
Roles of Forecasting Models
• Customer Relations Management
• Activation
Forecasting • Promotion
• Is an activity that calculates/predicts some future events • Loyalty
or conditions, usually as a result of rational study or • Pricing Strategy
analysis of pertinent data • Business Insights (minimize the pains of trial and errors)
Qualitative and Quantitative Forecasting • Maintain C, Expand DE, Attract AB
• Qualitative is an intuitive and educated guess • Consumer Insights (deliver the right goods and services)
• Quantitative is based on some deterministic or statistical • Preference
model and historical data • Profile
Requirements in Forecasting • Needs
• Sophisticated modeling • Implications
• Efficient data architecture, warehouse • Minimize adhoc, routine studies
• Computing technology • More in-house research capabilities required
• Computational statistics Statistical Forecasting Techniques
• The data/time series • A Statistical Forecasting Technique is one that
Sources of Data generates forecasts by extrapolating patters in historical
• POS database data
• Credit history • Forecasting is linked to the building of statistical models
• Usage history in the sense that before one can forecast a variable of
• Economic indicators interest, one must build a model for it and estimate the
• Consumer panel survey model’s parameters using historical data
• Sales monitoring • The estimated model summarizes the dynamic patters in
• Retail audit the data. It provides a statistical characterization of the
• Data retailers links between the past and future values
• Meta-analysis
Demand Forecasting
• New product is launched/Existing one is relaunched
• Increase demand?
• Coping with the demand
• Strategic planning
• Demand is not limited
• There is competition
• Regulations
• Can we keep up with the demand?
• How do we create new demand? Features Common to All Forecast
• Activation • Forecasting techniques generally assume that the same
• Promotion underlying causal system that existed in the past will
• Loyalty incentives continue to exist in the future
Sales Performance • Forecasts are rarely perfect; actual results usually differ
• Historical vs Drivers of Sales from predicted values
• Historical • Forecasts for groups of items tend to be more accurate
• Patterns than forecasts for individual items because forecasting
• Seasonality errors among items in a group usually have a canceling
• Trend effect
• Forecast accuracy decreases as the time period covered Irregular and Random Variation
by the forecast—the time horizon—increases • Irregular variations are due to unusual circumstances
Elements of a Good Forecast such as severe weather conditions, strikes, or a major
• Features Common to All Forecast change in a product or service.
• The forecast should be accurate and the degree of • Random variations are residual variations that remain
accuracy should be stated after all other behaviors have been accounted for
• The forecast should be reliable Naïve Methods
• The forecast should be expressed in meaningful units • A naive forecast uses a single previous value of a time
• The forecast should be in writing series as the basis of a forecast. The naive approach can
• The forecasting technique should be simple to be used with a stable series (variations around an average),
understand and use with seasonal variations, or with trend.
Forecast Based on Time Series • Stable series – last data point becomes the forecast for
• A time series is a time-ordered sequence of observations the next period
taken at regular intervals over a period of time (e.g., • Seasonal Variation – forecast for this “season” is equal
hourly, daily, weekly, monthly, quarterly, annually). to the value of the series last “season”
• Analysis of time series data requires the analyst to • Trend – the forecast is equal to the last value of the
identify the underlying behavior of the series. This can series plus or minus the difference between the last two
often be accomplished by merely plotting the data and values of the series
visually examining the plot.
Trend
• Trend refers to a long-term upward or downward
movement in the data. Population shifts, changing
incomes, and cultural changes often account for such
movements

Forecasting System
• Cycles are wavelike variations of more than one year’s
duration. These are often related to a variety of economic,
political, and even agricultural conditions.

Seasonality
• Seasonality refers to short-term, fairly regular variations
generally related to factors such as the calendar or time of
day. Restaurants, supermarkets, and theaters experience
weekly and even daily “seasonal” variations.

Weighted Moving Average


• A weighted average is similar to a moving average,
except that it assigns more weight to the most recent
values in a time series
Exponential Smoothing
• Exponential smoothing is a sophisticated weighted
averaging method that is still relatively easy to use and
understand. Each new forecast is based on the previous
forecast plus a percentage of the difference between that
forecast and the actual value of the series at that point:

Next Forecast = Previous forecast + α (Actual –


Previous forecast)

where (Actual = Previous forecast) represents the


forecast error and = is a percentage of the error

• Ft = Forecast for period t


• Ft-1 = Forecast for the previous period
• α = Smoothing constant
• At-1 = Actual demand or sales for the previous period

Trend Equation

Seasonal Relatives
• commonly used method for representing the trend
portion of a time series
• Computations and the resulting values are the same as
those for a moving average forecast. However, the
values are not projected as in a forecast; instead, they
are positioned in the middle of the periods used to
compute the

moving average.

Simple Linear Regression


yc = a + bx
yc = predicted (dependent)
variable X = predictor (independent)
variable b = slope of the line
a = value of yc when x = 0
Weaknesses of Linear Regression
• Simple linear regression applies only to linear
relationships with one independent variable
• One needs a considerable amount of data to establish
the relationship—in practice, 20 or more observations
• All observations are weighted equally
Measuring Forecast Accuracy
• Forecast accuracy is a significant factor when deciding
among forecasting alternatives. Accuracy is based on the
historical error performance of a forecast
• Mean Absolute Deviation (MAD)

• Mean Squared Error (MSE)

CONTROLLING THE FORECAST


• It is necessary to monitor forecast errors to ensure that
the forecast is performing adequately.
• The model may be inadequate due to (a) the
omission of an important variable, (b) a change or
shift in the variable that the model cannot deal with
(e.g., sudden appearance of a trend or cycle), or (c)
the appearance of a new variable (e.g., new
competitor).
• Irregular variations may occur due to severe weather or
Correlation other natural phenomena, temporary shortages or
• A measure of the strength and direction of relationship breakdowns, catastrophes, or similar events
between two variables • The forecasting technique may be used incorrectly or
• can range from -1.00 to +1.00 the results misinterpreted
• +1.00 indicates that changes in one variable are always • There are always random variations in the data.
matched by changes in the other Randomness is the inherent variation that remains in the
• -1.00 indicates that increases in one variable are data after all causes of variation have been accounted for
matched by decreases in the other TRACKING SIGNAL
• 0 = indicates little linear relationship between two
variables

Assumptions on Linear Regression Analysis


• Variations around the line are random
• Deviations around the line should be normally
distributed
• Predictions are being made only within the range of Control Limits
observed values
Guidelines in Getting the Best Result
• Always plot the data to verify that a linear relationship
is appropriate Where:
• The data may be time-dependent Square Root of MSE = standard deviation
• A small correlation may imply that other variables are z = Number of standard deviations; 2 and 3 are the
important typical values
Summary

You might also like