Lecture 8

Question 1

Relevance of sales models

Answer 1

Looking for the relationship between advertising and sass is somewhat worse than looking for a needle in a haystack aaker and carmen

Business managers need to know how markets respond to the actions they take

One of the primary goals of marketing science is to provide a structural insight of how a brand generates its sales, market share, customer awareness or any other variable of interest

Question 2

Relevant definitions of
Model
Sales response
Sales response model

Answer 2

A model is a generally simplified mathematical representation of real world relations

Sales response is the rate at which sales changes as a result of changes in business activities

Sales response models try to model a sales response as a function of business activities

Question 3

Before modelling…

Answer 3

Before modelling a response function the level of analysis should reflect the heterogeneous nature of a market

Question 4

Data sources of sales data 3 examples

Answer 4

Internal data - comes from company itself

Ad: readily available, intra daily basis

Household scanner data - data coming from retail upc scanners or from company itself

Ads; available for all major retail chains, available on a daily basis

Annual reports/ data coming from annual report of companies
Ads: available for all companies that are listed on a stock market

Question 5

The sales response model

Answer 5

What are the relevant sales drivers

What functional form appropriately represents the manner in which the drivers exter their effect

Sales = f ( quality, price, advertising)

Question 6

Overview of the types of static sales response models

Answer 6

Constant marginal returns - linear model

Decreasing marginal returns - multiplicative model / semi log model

Saturation volume - modified exponential model

S shaped - log reciprocal mode
Logistic model

Market share models - multiplicative interaction model
Multinomial logit model

Question 7

Objective of regression analysis

Answer 7

Quantification of the stops of a regression line

Estimation of the influence of one variable (x) on another variable (y)

R2 expresses the proportion of The explained variance in the dependent variable (y) that is explained by the regression line (value range 0:1)

Question 8

Linear sales response models

Answer 8

Q = a0 + A1 * x + u

Ex = A1 * x/q

S
Q = sales
X = intensity level for an instrument of the marketing mix
Aa A1 = response parameter to be estimated
U= error term

Question 9

Elasticity

Answer 9

E = dQ / dp * p/q

Question 10

Understanding elasticities

Answer 10

How large is the influence of an increase in advertising expenditures on sales

Absolute effect: sales / advertising = sales now - sales before / detailing now - detailing before

Relative eefct (elasticity) = sales / advertising
= sales now - sales before / sales before
All divided by detailing now - detailing before / detailing before

Question 11

Non linear models

Answer 11

Multiplicative model W= a0 * x a * eu

and semi logarithmic models
Q= a0+ A1 ln x + u
DECREASING MARGINAL returns

Question 12

Function of a multiplicative model

Answer 12

S= ap square root of b

If 0<b>0 goes Dowb curves </b>

Question 13

The multiplicative model : advertising

Answer 13

Q= k * OhH -adv^a * TV - ADV^b

Q= sales
TV-ADH = expenditures on tv advertising
OoH-ADV: expenditures on out of home advertising

a = sales effect of OoH advertising 
B= sales effect of tv advertising 
K= scaling parameter 

Decreasing marginal returns.

Question 14

The multiplicative model: elasticities

Answer 14

How can i analyse the sales effect in the multiplicative model using the estimated model parameters

Absolute effect: daw= dQ/ DTv Adv = k * OohAdv^a = B/ Tv Adv *Q

Effect deends on the level of advertising > that determines the sales level

Relative effect : daw / Q / dTvAdv / Tv Adv = daw / dTV Adv * Tv Adv/ Q = B

Elasticity corresponds to the exponent in the multiplicative model

Question 15

The multiplicative model : elasticities

Answer 15

Situation: Q= 0.4*OoHADV^0.1 * TvAdv^0.3

Absolute effect: 0.3/44 25100 = 17%

Relative effect: 0.3*100=30%

Question 16

Modified exponential model

Answer 16

Saturation limit - graph goes up then saturated

Question 17

Logistic model

Answer 17

Takes on the S shaped model

Log reciprocal model
And logistic model

Question 18

Logistic model starting point

Answer 18

Outcome variable can only be predicted in theory of a probability (eg purchase probability, adoption probability)

Event to be explained / predicted is binary (0 or 1)

Response (yes/no) to an offer in direct marketing
Inclusion of a product in the assortment of a retailer
First use of a new product (eg insulin pump) by buyer (adoption)

Question 19

Logistic model implications

Answer 19

Robust specifications required, I.e predictions lie in the interval 0 to 1.

Robustness implies non linear regression function

Special assumptions regarding the distribution of the error term required (logistic)

Question 20

Market share models

Answer 20

Mci model and mnl models

Question 21

Dynamic effects: sales response modes

Answer 21

Dynamic response model:

Q = Bo + B1X1 + Qt-1

Q is sales depend on time t
B1X1 is current effect
Qt-1 is carry over effect

Customers retailers and competitors might need a certain time to react to marketing activities ( delayed response effects, execution delays, noting delays, purchase delays, recording delays)

Especially the impact of advertising is considered to be a dynamic process

Customers, retailers and competitors might even react to marketing activities in advance (if heh anticipate an expected action)

Question 22

Dynamic effect; customer holdout model

Answer 22

Delayed response model (customer holdout model)

Hysteris effect based