T H E S I S D E F E N C E – 6 T H J U L Y 2 0 1 7 S A R A V A N E S
Customer Response To Push Advertising
Commercials: A Latent Profile Analysis
Introduction
Organisations used to have access to customers through
only a few media channels
The internet led to a growth of communication
technologies new online media channels
Push Advertising Channels
Interrupt the activities of users while they are serving
on the internet.
On-Demand Channels
Actively sought out by users, in order to find more
information about an offering.
Research Question
Do push advertising channels affect the amount of sales
in organisations in a FMCG setting?
… when the response of … when the response
the entire customer of different segments
Research Gap (I)
Earlier studies
focussed on the effect of banner advertising on the buying behaviour of viewers.
Current study
focusses on the effect of video commercial advertising on the buying behaviour of viewers
“push advertising commercials”
Research Gap (II)
Earlier studies
found an indication of heterogeneity across customers in their response towards push advertising activities
Current study
explores this heterogeneity by identifying discrete, homogeneous segments of customers based on their response to push advertising commercials
Elaboration Likelihood Model
Low level of motivation peripheral route of
processing
focus on superficial cues
(colour,
dynamics, etc.) temporary attitude shift
Push advertising activities particularly attractive to
influence customers in a
FMCG setting
, as they are
likely to buy based on superficial cues anyway?
Research Design
Panel dataset consisting of 7.742 households from GfK
panel with information regarding a soft-drink brand.
Containing information about
- demographics of households (education, age, household cycle)
- daily purchase behaviour (purchase value)
Methods: Ordinary Least Squares
To measure the response of the entire customer base to being
exposed to push advertising commercials Ordinary Least Squares (OLS)
with control function approach
to control for possible endogeneity
𝑺𝒊𝒕 = 𝛽0 + 𝛽1(𝑇𝑖𝑡) + 𝛽2(𝑇𝑖𝑡−1) + 𝛽3(𝑅𝑖𝑡) + 𝛽4(𝑅𝑖𝑡−1) +
𝛽5(𝑌𝑖𝑡) + 𝛽6(𝑌𝑖𝑡−1) + 𝛽7(σ + 𝛽8(σ𝑖𝑇) + 𝛽9(σ𝑖𝑅) + 𝜀𝑖𝑌) 𝑖𝑡
Methods: Latent Profile Analysis
To measure the response of discrete customer segments to
being exposed to push advertising commercials Latent Profile Analysis
determines the probability that a respondent belongs to a
latent cluster ‘k’ and places the respondent in the cluster
with the highest probability
𝑺𝒊𝑲𝒕 = 𝛽1𝐾 𝑇𝑖𝑡 + 𝛽2𝐾 𝑇𝑖𝑡−1 + 𝛽3𝐾 𝑅𝑖𝑡 + 𝛽4𝐾 𝑅𝑖𝑡−1 +
𝛽5𝐾 𝑌𝑖𝑡 + 𝛽6𝐾 𝑌𝑖𝑡−1 + 𝛽7𝐾 + 𝛽8σ𝑖𝑇 𝐾 + 𝛽9σ𝑖𝑅 𝐾 + 𝜀σ𝑖𝑌 𝑖𝐾𝑡
Results: Ordinary Least Squares
Check for endogeneity:
Wu-Hausman Test not significant (p=0,58) no endogeneity observed in the dataset
Exposure to a TV commercial at time t-1 significant with p<0,01
Exposure to a YouTube commercial at time t significant with p=0,02 Other push advertising variables had no significant effect on sales
Insights
Placing the customers together in one segment does not
adequately capture the reaction of different customers to push advertising commercials.
The amount of sales of an organisation barely change
when the exposure to push advertising commercials changes.
Organisations who do not have any information about their customer base should refrain from using push
Results: Latent Profile Analysis
The ideal number of clusters, based on Log-Likelihood values,
information criteria, cluster size and interpretability was three clusters.
The push advertising variables (excl. the exposure to TV commercials at
time t) had a significantly negative effect on the sales of cluster 2, but did not change the sales of the other clusters.
In terms of demographics, the three clusters are very similar.
Cluster 1 Cluster 2 Cluster 3
R2 <0,1% 4,3% 1,3%
Insights
The first cluster (‘frozen non-buyers’) is not affected by
the push advertising commercials and not interested in buying products from the brand lost investment.
The second cluster (‘delicate potentials’) has a higher
interest in the brand but is negatively affected by the
exposure to push advertising commercials lost profit.
The third cluster (‘fickle fans’) has a high interest in the
Limitations
Statistical power
less than 350 observations from 92.555 observations from households exposed to RTL.
less than 700 observations from 92.55 observations from households exposed to YouTube.
Brand awareness
dataset containing information about only one company in the soft-drink industry
Model choice
assumptions OLS not met as data were skewed to the right
Other marketing actions
ignoring possible synergy effects between channels
Conclusion
With regards to online push advertising, the general finding in
this study is that “less is more”.
The risks and costs of exposing customers to push advertising
commercials seems to be higher than the added value.
Difficult to target the ‘fickle fans’, as the three segments are
References
Cacioppo, J. T., & Richard, E. P. (1984). The elaboration likelihood model of persuasion.
Advances in Consumer Research, 11, 673-675.
Cho, H. C. (1999). How advertising works on the WWW: Modified elaboration likelihood
model. Journal of Current Issues & Research in Advertising, 21(1), 33-51.
Cho, H. C. (2003). The effectiveness of banner advertising: Involvement and click-through.
Journalism & Mass Communication Quarterly, 80(3), 623-645.
Manchanda, P., Dubé, J. P., Goh, K. Y., & Chintagunta, P. K. (2006). The effect of banner
advertising on internet purchasing. Journal of Marketing Research, 43(1), 98-108.
Naik, P. A., & Peters, K. (2009). A hierarchical marketing communications model of online and
offline media synergies. Journal of Interactive Marketing, 23, 288-299.
Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. Advances
in Experimental Social Psychology, 19, 123-162.
Spilker-Attig, A., & Brettel, M. (2010). Effectiveness of online advertising channels: A
