• No results found

The Effects of Multi-Channel Advertising Exposures and Payday Cycle in Customer Purchase Behavior

N/A
N/A
Protected

Academic year: 2021

Share "The Effects of Multi-Channel Advertising Exposures and Payday Cycle in Customer Purchase Behavior"

Copied!
14
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)
(2)
(3)
(4)

Conceptual Model

Purchase?

(Y/N)

(1) Channel?

(Online/Offline)

(2) Basket Value

Online ads

exposure

Control variables:

Folder,

Demographic,

Lag effect

Payday Cycle

(Payday/Reg)

Offline ads

exposure

Purchase Decision Stage

PURCHASE (Y)

Transaction Stage

(Masthead, GDN,

banner)

(TV, radio, print)

(5)
(6)

Methodology & Model Selection

Two-part model for purchase incidence (probit) & basket value (log-linear)

Separate binary probit for for channel selection

5

Does not account for endogeneity à does not guarantee better model

10

Robustness check: Tobit type 2 (purchase incidence & basket value), 2 binary

probit for channel selection purchase incidence

Predictive validity: hit rate, TDL, & MAPE

Model

Description

Log Likelihood

AIC

Model Fit

MPurchase1

Original model

-4,796.9

9,635.71

0.0202

MPurchase2

With lag ads

-4,795.3

9,644.56

0.0205

MChannel1

Original model

-123.07

286.15

0.136

MChannel2

With lag ads

-122.28

296.55

0.141

MBasket1

Original model

-

2,242.42

0.0135

MBasket2

With lag ads

-

2,245,65

0.0206

2 models for each stages

All models are

significant with low

model fit

(7)
(8)
(9)

Result: Basket Value

Positive Effect

Google Display

Network

Print

Relevancy: ads show up during browsing for

related information

16

Review longer à want more products

Payday

(moderate)

Increase of spending on payday

7

Lag-masthead

High obstrusiveness à longer recall

Lag-basket

value

Satisfaction à repurchase intention

13

Negative Effect

(10)

Predictive Validity

Predictive Validity

Purchase Incidence

Channel Selection

Basket Value

Estimation Sample

Hit Rate

58.79%

64.87%

-Top Decile Lift

1.368

2.424

-MAPE

-

-

1.35

Holdout Sample

Hit Rate

64.63%

72.92%

-Top Decile Lift

2.5

0

-MAPE

-

-

1.50

Inaccurate forecast à low model fit à need a lot more factors to explain

the dependent variable

More suitable to use the model to gain insight rather than for forecasting

Run model on estimation sample and holdout sample

Better performance on holdout sample à small sample, less variation

(11)
(12)
(13)

References

(14)

Referenties

GERELATEERDE DOCUMENTEN

To conclude, display advertising does influence sales significantly for both brand and category sales and this effect is even enhanced during the Christmas period for category

Do own and competitor intermediate mind-set states extracted using a dynamic hierarchical factor model provide a useful basis to improve the ability to predict sales in

Consumer purchase behaviour Television advertising Online advertising  Skippable  Non-skippable Synergy effect.. H1: TV advertising has a positive

Not only that, but also that it is more effective than some traditional media channels (TV and Radio) and that consumer journeys that consist of both online and offline

Therefore, we think the effect of banner advertising has a positive an maybe larger impact on households with children than TV and print advertising with regard to

Understanding factors that affect customer purchase behavior in a multi-channel environment is an important task for retailers; that includes the effect of

Four different models are constructed to investigate sequence effects on different forms of purchasing behavior: if sequence effects play a role in whether households purchase

If this is the case, it is important to ascertain which combination of cross-media marketing activities might have the greatest influence on the purchase behavior of