“Comparison of influence of selected viral advertising attributes on shopping behavior of Millennials – empirical study”

The study aims to evaluate the impact of selected factors of viral campaigns on Millennials customers’ consumer behavior. This goal was achieved in two steps: in the first step, the authors determined the impact of selected attributes on purchasing behavior in general, and in the second step, they compared the impact of the selected research campaigns – the guerrilla campaign of the company 4KA and the viral cam-paign of the company ABSOLUT. The inputs to the analyses were obtained through answers from 360 respondents, which completed the questionnaire on a sample of Millennials customers generation (1975–2000) – social generation, which collaborate and cooperate, expect technology to simply work for adventure and passionate about values (Smith, Nichols, 2015). The survey part of the questionnaire consisted of 8 attributes (Novelty, Relevance, Aesthetics, Clarity, Humor, Emotion arousal, Surprise, Design, Purchase intention). Data were collected based on participants’ availability and their will to participate in the questionnaire and quota selection. The PLS PM method was used to assess the impact, and the bootstrap-based parametric method was used to assess the difference in the impact. One of the most important findings is that attributes such as Novelty, Relevance, Humor, and Surprise significantly affect purchasing behavior. Concerning the company 4KA, significant impacts were seen in Relevance and Surprise, and with the company ABSOLUT, significant impacts were seen in Relevance, Humor, and Surprise. When analyzing the difference in the impact, there were no significant differences between the campaigns.


INTRODUCTION
In recent decades, the development of modern communication technologies has meant a significant change in marketing (Scullin, Fjermestad, & Romano, 2004).To gain a competitive edge and improve their performance, the companies have to develop new products and strategies to attract new customers and satisfy existing ones (Ungerman, Dedková, & Gurinová, 2018).Many people have transferred a substantial part of their activities to social networks and the Internet in general.Thus, companies that want to compete in a competitive environment have to move their marketing activities in this direction (Fong & Yazdanifard, 2014; Khan, Çera, & Netek, 2019).Companies are competitive if they can sell the products they manufacture or otherwise create (Malega, Rudy, & Kovac, 2019).However, it did not take long for the Internet today to feature tons and tons of promotional content.However, this satiety has caused ordinary users to stop seeing this content (Anusha, 2016).Nevertheless, the Internet and social networks are a good channel to reach the desired segment of customers.If a company wants to be successful on the market in highly competitive times, it should try to differentiate itself through its promotion activities (Verlegh, Fransen, & Kirmani, 2015).Both guerilla and viral campaigns, which are still not widely used in Slovakia (Dubcová, Grančičová, & Hrušovská, 2016), have proven to be suitable tools (Fong & Yazdanifard, 2014).However, their relatively rare occurrence may also present a certain advantage as something new and interesting.

LITERATURE REVIEW
According to the "father" of guerilla marketing (Levinson, 2011), guerilla marketing is a set of unconventional ways of achieving goals.It is an advertising strategy that focuses on low-cost marketing tactics, mostly small-scale, which can achieve interesting and unconventional results (Behal & Saaren, 2013).In essence, it is about using the elements of guerrilla fighting in marketing, that is, to achieve success (profit) with minimal resources (Nufer, 2013).Its great advantage is that it does not require much time, energy, and capital, and is therefore particularly suitable for small and medium-sized enterprises (Greco, 1995;Černá & Váňa, 2011).This marketing strategy is made up of changing and complex conditions that often create chaos or contradiction.The basic elements of guerilla marketing may include word of mouth, fan clubs, free samples, PR, and viral marketing (Lingelbach, Patino, & Pitta, 2012).
The concept of viral marketing describes any strategy that encourages individuals to convey a marketing message to others.In this way, viral marketing creates the potential for its exponential growth, dissemination, and influence.Like viruses, such strategies take advantage of rapid multiplication to spread the message to thousands of users (Wilson, 2000) and create an environment for spreading an idea (Dobele, Toleman, & Beverland, 2005).The principles of viral marketing are also very often used in guerilla campaigns (Ferguson, 2008 agree that in order for a campaign to be successful and have a lasting and unforgettable impact, the focus should be on the following factors: Novelty: the first criterion in deciding whether a product is creative, original, and unique.Novelty is also a common descriptor of creativity that has two characteristics: deviation from the norm (which is a very important point for viral marketing) and a sense of uniqueness (originality) (Jackson & Messick, 1965).
Relevance: it reflects how the advertisement's information contributes to or leads to the advertising message (Heckler & Childers, 1992).There are two types of relevance: advertising relevance (perception of whether ads make sense when communicating a product) and brand relevance (perception of product relevance for customer choice and it needs to be aligned with consumers' vision of the advertised brand) ( Clarity: for people to think about a product or brand, they must first understand the advertising message (Hafer, Reynolds, & Obertynski, 1996).Since in the case of guerrilla and viral campaigns, ads are executed and placed in an unusual way, it is important to ensure that your advertising message is easy to understand (Tam & Khuong, 2015).
Humor: draws attention to the product (Pieters, Rosenberg, & Wedel, 1999).Humor in advertising leads to a more positive attitude towards the ad and the brand itself and positively encourages customers to make purchases (Eisend, 2011).
Emotional arousal: each novelty has emotional consequences that determine whether or not advertising is accepted by consumers (Lindenmeier, 2008).Isen and Shalker (1982) and Damasio (2003) suggest that positive emotions can lead to a more favorable assessment of unexpected stimuli, while negative emotions cause less favorable assessments.
Surprise: it is a kind of wow factor that can draw customers' attention to a particular thing (Nufer, 2013).Guerilla marketing is used to attract attention and consumers (Baltes & Leibing, 2008).
Design: only the most creative, engaging, and entertaining advertising campaign will ensure the attention of the target audience (Shakeel & Khan, 2011).
Purchase intention: every consumer has a certain attitude towards advertising.Increasing the customer's purchase intent can be achieved primarily by creating more creative and trustworthy ads.The act of supporting the purchase intention leads to a higher willingness to buy the product (Schiffman & Kanuk, 2009; Vaitkevicius, Mazeikiene, Bilan, Navickas, & Sananeviciene, 2019).
The use of guerilla and viral campaigns is not very widespread in our territory.Here too, however, one can see some examples.In recent years, the guerilla campaign of the mobile operator 4KA and the viral campaign of ABSOLUT have attracted attention in this respect.
4KA, with its guerrilla campaign called "Cheap Calls and Data Resistance!" responded to its competitors' marketing activities.On March 3, 2016, it focused on the offer of a new mobile operator called Juro operating under Slovak Telekom, when it hung Juro in front of the company and posted signs with "the four biggest evils mobile operators do" (Luha, 2016).The next event took place on March 9, 2016, when the couriers entered several stores of the mobile operator O2 and launched balloons that in addition to 4KA's logo carried the words: A single price of 4 cents per minute, 1 SMS or 1 MB of data in Slovakia and throughout the European Union with 4KA.In O2, with the O2 Fér service, one pays up to 13 cents per minute in Slovakia and the Czech Republic (Ivančíkova, 2016).The last event took place on March 23, 2016 in front of Orange's store in Bratislava's Central shopping center.Four activists dressed as an owl, dolphin, panther, and kangaroo (Orange used these animals' names for their packages) appeared with banners to point out the four biggest evils of the telecoms market.Using the black banners of Orange, they pointed out limited offer validity or bigger benefits only for some, while on white banners, they presented their competitive advantages (Bartošová, 2016).
ABSOLUT conducted its guerrilla campaign in the autumn of 2018.The campaign with the slogan "Have fun.Let people live" divided people in Slovakia into two camps.The ad features a video in which the words of the Slovak folk song Slovenské mamy (Slovak Moms) can be heard when a young singer dressed in a folk costume enters a disco club in Orava.The camera picks up her back, and her face is not to be seen.The song is then transformed into a disco rhythm when she greets her friends and dances to the dance floor center.After a while, she turns to the camera, revealing her dark complexion, and says: "I'm Natalie and I'm Slovak.Slovakia has long been rich in all colors, so throw away your prejudices.Have fun, let people live", she says (Mikušovič, 2018).This ad became viral as soon as it was posted (September 13, 2018) and spurred a heated discussion.To this date, the video has more than 620,000 views on YouTube.While some people praised the idea as the right way to fight racism, others were called for canceling the ad and criticized it for indoctrinating the society with multicultural values (Holková, 2018; Naumova, Bilan, & Naumova, 2019).

DATA, METHODS AND HYPOTHESES
The primary objective of the present research is to evaluate the impact of selected factors of viral campaigns on the consumer behavior of the Millennials generation customers.This objective will be achieved with the help of two stages of statistical processing.In the first stage, the impact will be assessed in general, and in the second stage, the differences in coefficients measured when studying the campaigns of ABSOLUT and 4KA, i.e., the individual campaigns, will be tested (4KA -Guerilla; ABSOLUT -Viral generally).Based on these steps, the suitability of the instrument and its versatility will be assessed.Eight latent varia-bles acted as independent variables in the above assumptions (N -Novelty; R -Relevance; A -Aesthetics; C -Clarity; H -Humor; EA -Emotion arousal; S -Surprise; D -Design), and one acted as a dependent variable (PI -Purchase intention).

Two main hypotheses were formulated:
H1: There is an impact of the selected viral advertising attributes on the purchase intention.
H2: There is a significant difference in the impact of the selected viral advertising attributes on the purchase intention between the analyzed marketing campaigns.
The theoretical definitions subject to the constructs and assumptions in question are defined in the theoretical part.Data collection took place in several stages in 2019, where 360 respondents' opinions were collected.
The questionnaire was distributed electronically, through social networks, and via e-mail.Data were collected based on participants' availability and their will to take part in the questionnaire and quota selection (in the first stages of collection) and quota selection (in the final stages of collection), where efforts have focused primarily on gender.The secondary focus was placed on education and social status.Table 2 shows the selection's characteristics.As can be seen from Table 2, there were deviations in terms of population.However, the authors do not consider the deviations in question to be significant, and therefore there should be no deviations of outputs in terms of representativeness, so the selection is considered valid.Data not corresponding to the age range of the surveyed generation (year of birth: 1975-2000) were removed from the obtained data.Figure 1 shows the distribution of respondents in terms of their age.
As can be observed, the respondents' age is disproportionately distributed, so the outcomes of analytical processes can be more generalized to the younger generation -Millennials.It is also necessary to point out that the generations are specific to specific geographical and social conditions.For the area of Slovakia, the interval from 1975 to 2000 (date of birth) seems to be the most appropriate.
Survey variables with a five-point Likert scale were also included in the questionnaire (strongly disagree, disagree, neutral stance, agree, strongly agree).These variables aimed to find out the attitudes of participants to the attributes like N -Novelty; R -Relevance; A -Aesthetics; C -Clarity; H -Humor; EA -Emotion arousal; S -Surprise; D -Design a PI -Purchase intention.The specific questionnaire items are shown in Appendix 1.
The inference and verification of the sessions defined in the above hypotheses consist of two steps.
In the first step, Confirmatory Factor Analysis (CFA) is performed using Maximum Likelihood Estimation (MLE) (Schmitt, 2011).For CFA, ).In the second step, an analysis will be carried out to determine the impact through the Partial Least Squares method -Path Modeling (PLS PM) (Henseler, Hubona, & Ray, 2017).The impact determining part will be divided into sections where the fulfillment of the conditions and the possibility of applying PLS PM will be pointed out.One will look at sufficient degree of reliability (> 0.7), Eigenvalue (1 st > 1; 2 nd < 1) and factor loadings (> 0.7) (Sanchez, 2013).In the other part of the analyses, a parametric analysis of the difference in impact between individual marketing campaigns will be applied.This method will be applied to bootstrap at 500 replicates.The programming language R v. 3.6.1 (Action of the Toes) and ggplot2, lavaan, and plspm libraries were used for analytical processing.

RESULTS
Table 3 shows the outputs of the basic descriptive statistics and the outputs of the Confirmatory Factor Analysis (CFA).Information resulting from the descriptive statistics completes the view of particular latent variables and their parts -manifest variables.The table shows information determining the appropriate position of factors and the possibility of their processing by planned methods such as FL -Factor Loadings, AVE -Average Variance Extracted, CR -Composite Reliability.
Let us focus on the central tendency characteristics, which on the one hand show the degree to which respondents rated the given factors (the higher the number, the more positive the rating) and on the other, they point to possible distortion of the distribution compared to the normal distribution through a comparison of the median and average.The standard deviation is intended to point to fluctuations from the average.It is mostly used when comparing individual items, where the higher its rate, the higher the inconsistency of individual respondents' opinions in a particular item of the questionnaire.A very important function is the skewness and the spiciness, based on which one can approximately determine deviations from the normal distribution.For some items, deviations are identifiable, but there is no significant difference.Finally, the parametric method of determining difference based on bootstrap will be applied, conditioned by approximate normal distribution and absence of significant outliers.Based on the above, it can be concluded that these conditions are met to an acceptable extent.Factor loadings are less than 0.5 in one case and less than 0.7 in another (EA1 and EA4).These manifest variables will be excluded from further investigation.AVE and CR values meet specified criteria.After removing these variables, RMSEA acquires a value of 0.087, which is higher than the acceptable threshold but not by much, so the authors do not consider this deviation to affect the overall outcome of the investigation.Figure 2 shows the FL values required to apply the PLS PM model shown in the following sections.As is evident from Figure 2, none of the manifest variables acquires a value lower than 0.7, so no further adjustments to the model variables are needed, and from the FL perspective, the current state is considered acceptable.
Table 4 shows the evaluation of the application conditions of the PLS PM model.The first column specifies the latent variables.In the second column, Mode, there is an indication of all latent variables A, which gives us information that it is a reflective way of manifest and latent variables.
The third column indicates the number of manifest variables falling under specific latent variables.The fourth and fifth columns point to the reliability, where, as can be seen, no value is less than 0.7; hence, reliability is acceptable.The penultimate column and the last column point to latent roots, or, more precisely, Eigenvalue, where none of these values in the penultimate column acquires a value less than 1 or greater than 1 in the last column; the output is thus acceptable.The following tables show the model's actual outputs in general, the model for the 4KA's campaign, and the model for the ABSOLUT's campaign.Table 5 shows the outcomes of the analysis of the significance of viral campaigns' effects in general.In the first step, let us focus on the value of p, which is higher than 0.05; therefore, it is not possible to speak of significant influence, as is evident in the variables A -Aesthetics, C -Clarity, EA -Emotion arousal, and D -Design.The model itself has a GOF of approximately 0.5723 and an R2 of approximately 0.433.The estimate column points to the given attributes' impact, where it makes sense to focus only on significant attributes.Table 6 shows the outcomes of the analysis of the significance of 4KA's guerilla campaign's impacts.In the first step, let us focus on the value of p, which is higher than 0.05; therefore, it is not possible to speak of significant influence, which is evident in the variables N -Novelty, A -Aesthetics, C -Clarity, H -Humor, EA -Emotion arousal, and D -design.The model itself has a GOF of approximately 0.6421 and R2 of approximately 0.531.The estimate column points to the given attributes' impact, where it makes sense to focus only on significant attributes.Table 7 shows the outcomes of the analysis of the significance of the impact of ABSOLUT's viral campaign.In the first step, let us focus on the value of p, which is higher than 0.05; therefore, it is not possible to speak of significant influence, which is evident in the variables N -Novelty, A -Aesthetics, C -Clarity, EA -Emotion arousal, and D -Design.The model itself has a GOF of approximately 0.5143 and R2 of approximately 0.363.The estimate column points to the given attributes' impact, where it makes sense to focus only on significant attributes.
In the methodological part, the assumption of the existence of the influence of selected attributes of viral advertising on the purchase intention was determined.From Tables 5, 6, and 7, it can be inferred that in the general analysis, PIs are significantly affected by attributes such as Novelty, Relevance, Humor, and Surprise.With 4KA, significant impacts were seen in Relevance and Surprise, and with ABSOLUT these were seen in Relevance, Humor, and Surprise.From the above, it is possible to talk about the selected attributes' influence, so basic hypothesis H1 is accepted.Table 8 is devoted to the analysis resulting from hypothesis H2.In the first step of the assessment, let us focus on the last column where, as can be seen in any case, the value of p is not less than α = 0.05, so it is not possible to speak of a significant difference in campaigns.Therefore, hypothesis H2 is rejected.The first three digits of the table show the models' coefficients in question, followed by the difference between the 4KA and ABSOLUT models, and the penultimate row shows the statistics resulting from the applied test.
Figure 3 visualizes the impact of the investigated attributes on the PI, and the line thickness defines the degree of impact.The biggest impact was found in attributes like Relevance and Surprise, followed by Humor and Novelty.The dashed line symbolizes an effect that cannot be considered significant at α 0.05.

DISCUSSION
The concept of guerilla marketing and viral marketing is no longer a novelty.However, the main problem lies in the lack of understanding g (Tam & Kuong, 2015).In these conditions, this lack of understanding is even more pronounced, as the use of these tools is still in its initial phase and the vast majority of companies still prefer traditional forms of marketing

CONCLUSION
In the Slovak Republic, the examined marketing tools are still relatively unused, and so far, there is no relevant number of researches that would address this issue.The present article attempts to fill this gap, at least partially.Of course, the measured results will need to be verified on other samples, but they indicate the direction taken in our geopolitical space when using these marketing tools.It turns out that when combining the right factors, their gradual use in practice could have a positive response from consumers.As the chosen issue is still relatively marginal in this country, the problem is also the lack of quality resources in this area, which would relevantly describe the current situation in these geopolitical conditions.The elaboration of this contribution could partially fill the gap in this direction and help carry out further necessary research.

Purchase intention
After looking at the advertising, I would recommend the promoted product to friends who I know might be interested PI1 I will buy the advertised product PI2 If I ever accidentally come across a promoted product, I will probably buy it PI3 I would like to know more about the promoted products PI4 Smith, MacKenzie, Yang, Buchholz, & Darley, 2007; Kim & Chao, 2019).Aesthetics: in this context, aesthetics refers to the harmony of the ad's structure.It is about choosing the right way to combine characters and create complex interactions using the aesthetic factor (Wallend, G. Zinkhan, & L. Zinkhan, 1981; Besemer & O'Quin, 1986).

Figure 1 .
Figure 1.Distribution of age of the examined sample

Figure 2 .
Figure 2. Factor loadings PLS PM model

NoveltyI
was fascinated by advertising N1 Advertising has a good idea N2 The advertising is original N3 I perceive advertising as interesting N4 Relevance I perceive advertising as alive -"fresh" R1 The advertising made me think R2 Advertising has a deeper meaning R3 Advertising clearly distinguished the promoted product from competing R4 Aesthetics I found the advertising fun A1 The advertising was designed carefully A2 I perceive advertising as unique A3 I perceive advertising as unusual A4 Clarity I understood the message in the ad very well C1 I understood the ideas of advertising quickly C2 It was easy to understand the messages in the advertising C3 The essence of advertising was presented very clearly C4 Humor Advertising has a playful form H1 I find the advertising humorous H2 I find the advertising funny H3 The advertising made me laugh H4 Emotion arousal I find the advertising exciting EA1 Advertising made me nervous EA2 The advertising upset me EA3 The advertisement made me feel curious EA4 Surprise I was amazed after watching the advertising S1 The advertising surprised me S2 I perceive advertising as "crazy" S3 Design The design of the advertisement is imaginative D1 I was fascinated by the visual side of the advertising D2 I consider the technical processing of advertising to be of high quality D3 I consider the display's advertising page to be high quality D4 At first glance, I perceive advertising as attractive D5 ).

Table 1 .
Latent research variables -references Note: LV -latent variable, N MV -frequency of manifest variable.

Table 2 .
Identification characteristics of the sample

Table 5 .
PLS PM model output -whole

Table 4 .
PLS PM conditions

Table 8 .
PLS PM estimate differences test Note: A -whole model, B -4KA model, C -ABSOLUT model.