and Their Influence on Consumer Behavior

Fractal Characteristics as Advertising Stimuli

and Their Influence on Consumer Behavior

Table of Content

Introduction

• Different visual aspects of ads have been investigated as determinants of aesthetic liking (symmetry, contrast, clarity, complexity)

• Visual aspect which has been overlooked in advertising research  fractal characteristics of an ad

• Nature, art works and architecture share a similar level of fractal characteristics

(Braun, Amirshahi, Denzler, Schweinberger & Redies, 2013)

• Fractal characteristics in ads preferred?

How do the fractal characteristics of an ad influence the aesthetic liking of the ad and eventually the consumer’s willingness to pay and the net promoter score for the

Literature review

(fractal characteristics)

• Mandelbrot (1983) developed the ‘fractal geometry’

• Complex mathematical structures that appear to be self-similar whether viewed on a larger or a smaller scale

(Mandelbrot, 1967)

• Self-similarity

Literature review

(fractal characteristics)

Fractal dimension

• Parameter describes how the patterns build the fractal shape • Quantifies complexity (strong related to self-similarity)

• Smooth line (value of D = 1) + Completely filled area (value of D = 2)

Literature review

(fractal characteristics)

• Scale invariant  same characteristics at every scale of observation (Brown & Liebovitch, 2010, p5)

• can be quantified with the Fourier power spectrum • Linear relation with a particular slope

• measures the spatial frequency power (Menzel, Hayn-Leichsenring, Langner, Wiese & Redies, 2015)

• Different spatial scales, from very fine to very coarse

• Fourier slope of complex natural scenes is characterized by a slope of -2 • scale invariant spatial frequency spectrum

Literature review

(aesthetic liking)

• Ad designer tries to induce a positive affective thought toward the ad (Vakratsas & Ambler, 1999)

• Aesthetic liking  the mental processes that underlie experiences with positive end feelings that would accompany verbal expressions such as 'oh wow', 'that’s

Literature review

(aesthetic liking)

Aesthetic liking and fractal characteristics

• Aesthetic preference across fractal images which contain fractal dimensions in the 1.3–1.5 range (Spehar et al., 2003)

• A shallow slope of -2 is preferred over a steep slope of -3.5 when judging the beauty of human faces or natural scenes (Blickhan, Kaufmann, Denzler, Schweinberger and Redies, 2011)

• Subsets of artworks and other visually pleasing images also share this slope of about -2 (Braun et al., 2013; Graham & Field, 2007; Graham & Redies, 2010; Koch, Denzler & Redies, 2010; Redies et al., 2007)

Literature review

(WTP)

• Marketing managers are constantly seeking opportunities to enhance positive consumer behavior  higher WTP for example

• Willingness to pay (WTP)

• the maximum amount of money a consumer is willing to spend for a product or service (Krishna, 1991)

• Pricing effects can occur due to advertising (Kalra & Goodstein, 1998)

H2: An ad with a high scale invariant fractal characteristic will more positively

Literature review

(Net Promoter Score)

• NPS

• measure of customer satisfaction (Homburg et al., 2005)

• NPS

• ‘how likely is it that you would recommend (brand / product) to a friend or colleague?’ (Reichheld, 2003)

• NPS can influence future consumer behavior (Reichheld, 2003; Van Doorn et al., 2013)

H3: An ad which contains a high scale invariant fractal characteristic will more

Literature review

(aesthetic liking mediator)

• Consumers normally do not switch from being disinterested individuals to convinced purchasers in one instantaneous step

• Attitude toward the ad (affective construct) is a mediator between the ad stimuli and consumer behavior (Mitchell, 2013; Vakratstas and Ambler, 1999)

H4: Aesthetic liking mediates the effect of a high scale invariant fractal characteristic of an ad on the willingness to pay for the advertised product

Literature review

(consumer motivation)

• Ad effectiveness depends also on the particular mental processes that an ad recipient invokes (Meyers-Levy & Malaviya, 1999)

• Consumer motivation is defined as the readiness to process the information in an ad

• High motivated consumers rely their judgment more on cognitive thinking • Low motivated consumers rely their judgment more on affective feelings

Literature review

(self-monitoring)

• Self-monitoring  individuals differ in the extent to which they are able to monitor (observe and control) their own expressive behavior and self-presentation (Snyder & DeBono, 1987)

• High and low self-monitoring people differ in their behavioral responses to image- and quality-oriented advertising appeals (Snyder & DeBono, 1985)

Methodology

• A factorial 2 (“low consumer motivation” vs “high consumer motivation”) by 3 (“high scale invariant Fourier slope -2” vs “low scale invariant Fourier slope 0” vs “low scale invariant Fourier slope -4”) between participants design was set up

Low

consumer motivation

High

consumer motivation High scale invariant ad (slope -2) Condition 1 Condition 2

Low scale invariant ad (slope 0) Condition 3 Condition 4

Methodology

Manipulation fractal characteristics

Methodology

Manipulation consumer motivation

• The consumer motivation manipulation attempted to increase the decision relevance

• The low motivation instructions were: ‘On the following page you will see an

advertisement. You have to answer a number of questions about the advertisement. Do not try to analyze what you are being shown, just relax and answer the questions.’

• The high motivation instructions were: ‘On the following page you will see an

Methodology

Aesthetic liking measurement

• Attractiveness, arousal, interestingness, pleasantness, boredom, and innovativeness rated on a 7-point Likert scale

• Cronbach’s alpha: Aesthetic liking:  = 0.90

WTP measurement

• Gabor-Granger technique

• Different prices were presented, in random order

• The base price was the average price of the product

• 10% more than the average price, 5% less, 15% more, 10% less, 20% more and 5% more • The mean price was adjusted to €1.70, and the other prices were €1.87; €1.62; €1.95; €1.53; €2.04;

€1.79

NPS measurement

• ‘How likely is it that you would recommend this product to a friend or colleague?’ • 0 = ‘not at all likely’ and 10 = ‘extremely likely’

Self-monitoring measurement

• 18-item self-monitoring scale developed by Gangestad and Snyder (2000)

Results

Demographics

• 125 respondents completed the survey

• The total sample consisted of 70 (56%) men and 55 (44%) women • The average age was 27

Manipulation check

• An one-way ANOVA :

• consumer motivation = IV • experimental conditions = DV

• This one-way ANOVA appeared to be not significant, F(1, 123) = 2.06, p = .154

• The manipulation did not work as planned

Results (hypothesis 1 & 6)

• Two-way Anova

• Fourier slope and consumer motivation = IVs • Aesthetic liking = DV

• Fourier slope and aesthetic liking not significant , F (2, 119) = 1.19, p = .309 • Hypothesis 1 not confirmed

• Consumer motivation and aesthetic liking not significant , F(2, 119) = 0.03, p = .866 • Hypothesis 6 not confirmed

• Interaction effect not significant, F (2, 119) = 1.18, p = .310

2,6 3,24 2,5 2,67 2,75 2,8 2,4 2,6 2,8 3 3,2 3,4 0 -2 -4 Aesthetic liking Means Fourier Slope Low motivation High motivation Additional finding

Within the low motivation condition marginally significant difference in aesthetic liking between the Fourier slope of 0 (M = 2.60, SD = 0.96) and the Fourier slope of -2 (M = 3.24, SD = 1.55), p = .087

Within the low motivation condition significant

Results (hypothesis 2)

• Two-way Anova

• Fourier slope and consumer motivation = IVs

• WTP = DV

• Fourier slope and WTP not significant, F (2, 119) = 1.98, p = .142 • Hypothesis 2 not confirmed

• Consumer motivation and WTP not significant , F(2, 119) = 0.587, p = .445 • Interaction effect not significant, F (2, 119) = 0.129, p = .879

Additional finding

Excluding respondents who were not willing to pay something for the advertised product

Results (hypothesis 3)

• Two-way Anova

• Fourier slope and consumer motivation = IVs

• NPS = DV

• Fourier slope and NPS not significant, F (2, 119) = 1.19, p = .309 • Hypothesis 3 not confirmed

• Consumer motivation and NPS not significant , F (2, 119) = 0.19, p = .663 • Interaction effect not significant, F (2, 119) = 0.94, p = .394

Additional finding

marginally significant difference in NPS between the Fourier slope of -2 (M = 3.27 SD = 3.10) and the

Fourier slope of -4 (M = 1.90 SD = 2.30), p = .057 within

the low motivation condition 2,43

Results (hypothesis 4 & 5)

• Mediation analysis via PROCESS macro • WTP = DV

• Fourier slope -2 and -4 = IV

• Aesthetic liking = proposed mediator (M)

• Confidence interval includes zero (-0.9869 to 0.1350), so no statistically interaction • Hypothesis 4 is not supported

Hypothesis 5

• Mediation analysis via PROCESS macro • NPS = DV

• Fourier slope -2 and -4 = IV

• Aesthetic liking = proposed mediator (M)

• Confidence interval includes zero (-1.2178 to 0.2176), so no statistically interaction • Hypothesis 5 is not supported

Additional finding

• Mediation analysis via PROCESS macro (only in the low motivation condition) • NPS = DV

• Fourier slope -2 and -4 = IV

• Aesthetic liking = proposed mediator (M)

• Confidence interval not includes zero (-2.4189 to -0.0652), so statistically interaction

Results (hypothesis 7)

• Effect of self-monitoring on the relationship between Fourier slope and aesthetic liking

• Moderation test via PROCESS macro

• Fourier slope = IV • Aesthetic liking = DV

• Self-monitoring = proposed moderator (M)

Discussion

• Although this research did not confirm clearly that fractal characteristics in ads influence

• the aesthetic liking of the ad

• the WTP for the advertised product • the NPS of the ad

• Important finding: aesthetic liking differs between a high scale invariant ad and a low scale invariant ad, when a consumer is less motivated

• These findings are consistent

• with the conclusions of Mitchell (2013) who found that when consumer motivation is low, the consumers will not actively search for information in memory, but rather will base their judgments on affective information

Implications,

Limitations & Future research

Implications

• Marketers can make difference between their consumers with regard to motivation or specific motivational states

Limitations and future research

• Age is not representative of the general population

• Future: test hypotheses with larger and more representative sample

• Consumer motivation manipulation did not succeed

• Future: Solid manipulation check or measure motivational state

• Large number of respondents did not want to pay anything • Future: conjoint analysis to measure WTP