Trends, Forecasts and Patterns Flashcards
Business Strategy Issues
Long term trend - T
Seasonal variations - S
Cyclical variations - C
Random variations - R
Time Series Analysis
Involves the analysis of past observations in order to forecast a variable into the future
Steps of Time Series Analysis
Find the trend
Calculate the seasonal variation
Forecast the next year of numbers
Seasonal variation can be found through additive model or multiplicative model
Additive Model
Used when seasonal variations are fixed amounts. On a graph, the gap above and below the trendline remains consistent over time
Multiplicative Model
Used when seasonal variations are a percentage amount. On a graph, the gap above and below the trendline is always getting bigger over time
Forecasting on a Graph
To forecast using a graph, simply extend the graph, using the same patterns found in previous periods
Adjusting the Trend Line: Additive Model
Additive model assumes variations to be a consistent amount:
Y = T + S + C + R
Y is income T is long term trend S is seasonal variation C is cyclical variation R is random variation
Adjusting the Trend Line: Multiplicative Model
Additive model assumes the seasonal variation to be a constant proportion:
Y = T x S x C x R
Y is income T is long term trend S is seasonal variation C is cyclical variation R is random variation
Calculating the Adjusted Value Using the Multiplicative Model
Actual value
Adjusted value = ——————————-
Seasonal component
Can be rearranged to find seasonal comp or actual value
Calculating the Adjusted Value Using the Additive Model
Adjusted value = trend + seasonal component
Forecasting Future Sales Figures Using a Sales Equation
Step 1: calculate the trend
Step 2: calc seasonal variation using additive variation (units sold - trend)
Step 3: forecast sales using additive variation (the equation)
Step 4: Calculate the seasonal variations using multiplicative variation (% difference between actual figures and the trend)
Step 5: forecast sales using multiplicative variation (trend figures x by %)
Calculating the Trend Given the Actual Sales and the Seasonal Variation
Step 1: express multiplicative variation as a decimal (e.g. 8% = 1.08)
Step 2: use the addictive model formula to calculate the trend:
Y = T + S, rearranged as
Y
T = —–
S
Finding the Linear Trendline Using the Least Squares Method
T = a + bt
T = trend a = where the trendline cuts the vertical axis at 0 b = increase or decrease in one time period t = time period
The Increase or Decrease in One Time Period in a Linear Trend
NΣxy - ΣxΣy
b = ——————-
NΣx² - (Σx)²
b = increase or decrease in one time period x = the quarter number y = the actual data n = the number of items of data
Where the Trendline Cuts the Vertical Axis at Time 0
a = ȳ - bx̄
b = increase or decrease in one time period x̄ = average of x ȳ = average of y