Sales Forecasting

Prepared by: Brian Rutherford

L14

Objectives
How do we forecast sales?

Sales force composite Jury of executive opinion Survey of buyer intentions Trend projections Moving averages Exponential smoothing Regression Econometric models

Why do we need to know how to forecast/predict sales?

Starting point of short term (3 to 6 months), medium term (6 to 24 months), and long term planning.

Short/medium term forecasting is the basis for determining a companys production schedule.
Long term forecasting is useful in determining what new facilities, labor, and funding will be needed.

(Source: Hite and Johnston)

Forecasting helps:

In setting up territories Assigning quotas Comparing sales performance

(Source: Dalrymple et al)

Forecasting is important for at least five reasons

Basis for setting and maintaining a production schedule Determine the quantity and time of needs for labor, equipment, tools, parts, are raw materials Influence the amount of borrowing needed Basis for assigning a sales quota Help determine the companys business and marketing plans
(Source: Futrell)

Other reasons forecasting is important

Predict employee turnover Predict employee satisfaction (dissatisfaction) Predict organizational commitment Predict sales force burnout

Define

Sales forecast- the estimated dollar or unit sales for a specific future time period
(Source: Futrell)

Market potential- estimate of maximum demand in a time period based on the number of potential users and the purchase rate

Sales potential- a companies portion of total industry demand

(Source: Dalrymple et al)

Sales quotas- the sales goal, in either dollars of physical units, of a given product, product line, or service assigned to a particular individual or part of the sales force in a given territory during a given period of time
(Source: Hite and Johnston)

Qualitative Sales Forecasting

Guess Sales force composite Jury of executive opinion Asking buyer intentions

Guess
What we did last time is what we will do this time

Qualitative Sales Forecasting

Sales force composite Asks the options of sales personnel concerning future sales Salesperson gives an estimate of future sales After the salesperson gives the estimate the sales manager may adjust it Best used when a highly trained and specialized sales force is used

Sales force composite

Advantages

Attention is called to the actual sales potential for each territory It uses the knowledge of those closest to the market Salesperson is more likely to believe their forecast is correct It places more responsibility on the salesperson

(Source: Hite and Johnston)

Qualitative Sales Forecasting

Sales force composite

Disadvantages

Salespeople are not trained in forecasting May try to underestimate due to quotes Salespeople may not spend the time needed

(Source: Hite and Johnston)

Qualitative Sales Forecasting

Jury of executive opinion

Uses two to ten company executives and their responses are pooled into one forecast (Source: Hite and Johnston) Use one seasoned individual (most often used when the company is small) Best used when executives have a strong working knowledge of the area and other sources of data is hard to find

Jury of executive opinion Advantages Can be done easily and quickly without a lot of elaborate statistical manipulations Incorporated a variety of opinions from executives

(Source: Hite and Johnston)

Qualitative Sales Forecasting

Jury of executive opinion

Disadvantages

Over reliance on personal opinion that is not backed up with facts People with little marketing and sales knowledge may be involved in making the decisions Lack of statistics May be undue influence in certain cases
(Source: Hite and Johnston)

Qualitative Sales Forecasting

Jury of executive opinion

Variations

Factor Listing- when each member of the jury is required to list factors that will have a positive or negative impact. A consensus is then sought on the magnitude of each factor so a prediction can be made. Delphi Technique- members of the jury never meet and make anonymous forecasts. The leader averages and returns a median or mean forecast to each member of the jury. Each member evaluates and revises the forecast until a consensus is met.
(Source: Hite and Johnston)

Qualitative Sales Forecasting

Asking buyer intentions

When buyers are asked what their future intentions are

(Source: Hite and Johnston)

Do.. Customers have the ability to project their purchasing needs Customers have a good record of following through Customers are financial capable of following through Willing to gave an honest disclosure of future needs
(Source: Dodge) Asking buyer intentions

Advantages

People that are going to purchase the product are involved May be the only way to predict future sales (new firm)

Qualitative Sales Forecasting

Asking buyer intentions

Disadvantages

People may not know what they are going to purchase They may report what they want to buy, but not what they are capable of buying Customers may not want to disclose that information Effects of derived demand may make forecasting difficult
(Source: Hite and Johnston)

Quantitative Sales Forecasting

Trend Projections Moving Averages Exponential Smoothing Regression Analysis Econometrics

Trend projections
Year 1-14 1 2 3 4 5 6 7 Year 15-28 15 16 17 18 19 20 21 Sales 1-14 197 211 203 247 239 269 308 Year 15-28 305 308 356 393 363 386 443

8
9 10

22
23 24

262
258 256

308
358 384

11

25

261

12

26

288

13

27

296

14

28

276

Trend projections
Trend
500 400

Sales

300 Series2 200 100 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 Year

Trend
Trend Projection
500 400

Sales

300 200 100 0 1 3 5 7 9 11 13 15 17 19 21 23 25 27 Year

Sales Trend Line

Moving Averages
Month 1 2 3 4 5 6 7 8 9 10 11 Actual 60 65 68 73 69 65 62 59 63 66 58

12

60

Moving Averages
Month Actual 3 month moving average

60

2 3 4 5

65 68 73 69 64 69 70

6
7 8 9 10 11

65
62 59 63 66 58

69
65 62 61 63 62

12

60

61

Moving Averages
Month Actual 3 month moving average Forecast 1 60

65

3 4 5 6 7 8 9 10 11 12

68 73 69 65 62 59 63 66 58 60

64 69 70 69 65 62 61 63 62 61 64 69 70 69 65 62 61 63 62

61

Moving Averages
Forecast with Moving Average
75 70

Sales

65 60 55 50 1 2 3 4 5 6 7 Time 8 9 10 11 12 13

Actual Moving average Forecast

Quantitative Sales Forecasting

Exponential Smoothing Heavily weighted toward recent periods May still lag behind recent trends Smoothing constant is chosen subjectively (Source: Hite and Johnston)

Exponential Smoothing
Month
1 2 3 4 5 6 7 8 9 10 11 12

Actual
200 135 195 197.5 310 175 155 130 220 277.5 235

0.1

0.5

0.9

200 193.5 193.65 194.035 205.6315 202.5684 197.8115 191.0304 193.9273 202.2846

200 167.5 181.25 189.375 249.6875 212.3438 183.6719 156.8359 188.418 232.959

200 141.5 189.65 196.715 298.6715 187.3672 158.2367 132.8237 211.2824 270.8782

Exponential Smoothing

Forecast for period Twelve

235(.1) + 202.3(.9) = 205.6 235(.5) + 233(.5) = 234.0 235(.9) + 270.9(.1) = 238.6

a=.1 a=.5 a=.9

Exponential Smoothing
Month
1 2 3 4 5 6 7 8 9 10 11 12

Actual
200 135 195 197.5 310 175 155 130 220 277.5 235

0.1

0.5

0.9

200 193.5 193.65 194.035 205.6315 202.5684 197.8115 191.0304 193.9273 202.2846

200 167.5 181.25 189.375 249.6875 212.3438 183.6719 156.8359 188.418 232.959

200 141.5 189.65 196.715 298.6715 187.3672 158.2367 132.8237 211.2824 270.8782

205.5561

233.9795

238.5878

Regression

Adjusted R^2 Significance level of .05

(Constant does not matter if non-significant)

Unstandardized coefficient B

(Prediction equation)

Simple Regression
Years 2.00 3.00 4.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 Sales 3.00 6.00 8.00 4.00 10.00 14.00 8.00 12.00 14.00 12.00 16.00

Simple Regression
M odel Sum m ary A djusted R Square .675 Std. Error of the Estimate 2.44799

Model 1

R .841 a

R Square .707

a. Predictors: (Constant), Y ears w ith the f irm

a Coe fficients

Model 1

(Constant) Years w ith the firm

Unstandardized Coefficients B Std. Error 2.170 1.781 .978 .210

Standardized Coefficients Beta .841

t 1.218 4.662

Sig. .254 .001

a. Dependent Variable: Sales for Period 1

Simple Regression

Significance level of .05

(Constant does not matter if non-significant)

Unstandardized coefficient B

(Prediction equation)

a Coe fficients

Model 1

(Cons tant) Years w ith the f irm

Unstandardiz ed Coef f icients B Std. Error 2.170 1.781 .978 .210

Standardized Coef f icients Beta .841

t 1.218 4.662

Sig. .254 .001

a. Dependent Variable: Sales f or Period 1

Sales for Period 1= 2.170 + .978(X1)

.001<.05 significant and can be used

Simple Regression

Adjusted R^2
M odel Sum m ary A djusted R Square .675 Std. Error of the Estimate 2.44799

Model 1

R .841 a

R Square .707

a. Predictors: (Constant), Y ears w ith the f irm

Adjusted R Square equals .675

Simple Regression
Sales for Period 1= 2.170 + .978(X1) Predict sales for the following values 1 5 10

Sales for Period 1= 2.170 + .978(X1)

Predict sales for the following values

1 = 3.148 5 = 7.06 10= 11.95

Instead of one X and one Y, we have multiply Xs predicting Y Y= B1(X1) + B2(X2) + B3(X3) + B4(X4)

Multiply Regression
Clients Hours 4.00 40.00 9.00 40.00 10.00 35.00 4.00 45.00 12.00 40.00 15.00 35.00 7.00 38.00 13.00 50.00 15.00 35.00 14.00 45.00 15.00 50.00 Years Sales 2.00 3.00 3.00 6.00 4.00 8.00 6.00 4.00 7.00 10.00 8.00 14.00 9.00 8.00 10.00 12.00 11.00 14.00 12.00 12.00 13.00 16.00

Multiply Regression (Example 1)

M o d el Su m m ary A djusted R Square .955 Std. Error of the Estimate .90670 Model 1 R .982 a R Square .964

a. Predictors: (Constant), Clients called per day , Y ears w ith the f irm

a Coe fficients

Model 1

(Cons tant) Years w ith the firm Clients c alled per day

Unstandardiz ed Coefficients B Std. Error -1.034 .783 .373 .111 .735 .097

Standardized Coefficients Beta .321 .727

t -1.320 3.350 7.590

Sig. .223 .010 .000

a. Dependent Variable: Sales for Period 1

Multiply Regression (Example 1)

Years with the firm .010<.05 significant and can be used Clients called per day .000<.05 significant and can be used

a Coe fficients

Model 1

(Cons tant) Years w ith the firm Clients c alled per day

Unstandardiz ed Coefficients B Std. Error -1.034 .783 .373 .111 .735 .097

Standardized Coefficients Beta .321 .727

t -1.320 3.350 7.590

Sig. .223 .010 .000

a. Dependent Variable: Sales for Period 1

Multiply Regression (Example 1)

M o d el Su m m ary A djusted R Square .955 Std. Error of the Estimate .90670 Model 1 R .982 a R Square .964

a. Predictors: (Constant), Clients called per day , Y ears w ith the f irm

Adjusted R Square equals .955

Multiply Regression (Example 1)

Sales for Period 1= -1.034 + .373(B1) + .735(B2)
Predict sales for the following values Years & Clients 1 5 5 7 10 15
a Coe fficients

Model 1

(Cons tant) Y ears w ith the f irm Clients c alled per day

Unstandardiz ed Coef f icients B Std. Error -1.034 .783 .373 .111 .735 .097

Standardized Coef f icients Beta .321 .727

t -1.320 3.350 7.590

Sig. .223 .010 .000

a. Dependent V ariable: Sales f or Period 1

Multiply Regression (Example 1)

Sales for Period 1= -1.034 + .373(B1) + .735(B2)

Predict sales for the following values

Years 1 5 10 & Clients = 5 7 15 Predicted Sales 3.014 5.976 13.721

Multiply Regression (Example 2)

M o d el Su m m ary A djusted R Square .951 Std. Error of the Estimate .95364 Model 1 R .983 a R Square .965

a. Predictors: (Constant), Hours w orked per w eek, Clients called per day , Y ears w ith the firm

a Coe fficients

Model 1

Unstandardiz ed Coef f icients B Std. Error (Cons tant) .161 2.615 Y ears w ith the f irm .408 .138 Clients called per day .715 .109 Hours w orked per w eek -3.06E-02 .064

Standardized Coef f icients Beta .351 .708 -.040

t .062 2.956 6.548 -.481

Sig. .953 .021 .000 .645

a. Dependent Variable: Sales f or Period 1

Multiply Regression (Example 2)

Years with the firm .021<.05 significant and can be used Clients called per day .000<.05 significant and can be used Hours worked per week

.645<.05 non-significant and model needs to be changed

a Coe fficients

Model 1

Unstandardiz ed Coef f icients B Std. Error (Cons tant) .161 2.615 Y ears w ith the f irm .408 .138 Clients called per day .715 .109 Hours w orked per w eek -3.06E-02 .064

Standardized Coef f icients Beta .351 .708 -.040

t .062 2.956 6.548 -.481

Sig. .953 .021 .000 .645

a. Dependent Variable: Sales f or Period 1

Regression (Summary)

Need significant variables

Looking for a high adjusted R^2

If variables are non-significant we need to modify the model

Quantitative Sales Forecasting

Econometrics

Combines a series of multiply regression equations Reflect the historical relationships among factors in the economy What if statements Can be very expensive Generally used to forecast industry sales and/or trends in the total economy

Summary

Qualitative vs. quantitative Why is forecasting important? Why what you choose to use different forecasting methods?