The Effects of Social Media Marketing on Online Consumer Behavior

Social media allows customers and prospects to communicate directly to your brand representative or about your brand with their friends. However, the obvious question is: who are the people interacting online and how engaged are they in online activities? This paper aims to answer this question based on a study regarding the online activities of 236 social media users, by identifying different types of users, a segmentation of these users and a linear model to examine how different predictors related to social networking sites have a positive impact on the respondents’ perception of online advertisements. The answer can help discover how to engage with different types of audiences in order to maximize the effect of the online marketing strategy.

"Lucian Blaga" University of Sibiu, Romania.The responses that will be further explored were gathered through field research, namely the information was collected directly from respondents via the internet, from September 17 to November 18, 2011.Finally, in the last section of the paper, we provide the contributions of the research, the managerial implications, and limitations of the research.

The Development of Social Media Marketing
In recent years, social networking sites and social media have increased in popularity, at a global level.For instance, Facebook is said to have more than a billion active users (as of 2012) since its beginning in 2004 (www.facebook.com).Social networking sites can be described as networks of friends for social or professional interactions (Trusov, Bucklin, & Pauwels, 2009).Indeed, online social networks have profoundly changed the propagation of information by making it incredibly easy to share and digest information on the internet (Akrimi & Khemakhem, 2012).
The unique aspects of social media and its immense popularity have revolutionized marketing practices such as advertising and promotion (Hanna, Rohm, & Crittenden, 2011).Social media has also influenced consumer behavior from information acquisition to post-purchase behavior such as dissatisfaction statements or behaviors (Mangold & Faulds, 2009) and patterns of Internet usage (Ross et al., 2009;Laroche et al., 2012).
Social media is ''a group of internet based applications that builds on the ideological and technological foundations of Web 2.0, and it allows the creation and exchange of user-generated content'' (Kaplan & Haenlein, 2010, p.61).Social media has many advantages as it helps connect businesses to consumers, develop relationships and foster those relationships in a timely manner and at a low cost as Kaplan and Haenlein discovered (2010).
Other functions of social media involve affecting and influencing perceptions, attitudes and end behavior (Williams & Cothrell, 2000), while bringing together different like-minded people (Hagel & Armstrong, 1997).In an online environment, Laroche (2012) pointed out that people like the idea of contributing, creating, and joining communities to fulfill needs of belongingness, being socially connected and recognized or simply enjoying interactions with other like-minded members.
The much higher level of efficiency of social media compared to other traditional communication channels prompted industry leaders to state that companies must participate in Facebook, Twitter, MySpace, and others, in order to succeed in online environments (Kaplan & Haenlein, 2010;Laroche et.al. 2012).Thus, more industries try to benefit from social media as they can be used to develop strategy, accept their roles in managing others' strategy or follow others' directions (Williams & Williams, 2008).
Social media websites provide an opportunity for companies to engage and interact with potential and current consumers, to encourage an increased sense of intimacy of the customer relationship, and build all important meaningful relationships with consumers (Mersey, Malthouse, & Calder 2010) especially in today's business environment when consumer loyalty can vanish at the smallest mistake, which can additionally have online propagation of their unfortunate encounter with a particular product, service, brand or company.Some companies are beginning to take notice of the power of social media.A few corporate social networking websites already allow consumers to not only exchange information about products or services, but also engage in co-creating value in online experiences with offline outcomes, with both current and potential consumers.

Segmentation of Social Media Users
Following the general idea that segmentation can leverage a better understanding of consumers' behavior, and therefore a better targeting, in order to obtain the desired effect of any marketing activity, several studies have been employed to achieve a segmentation of consumers who interact online, particularly to examine their online shopping behavior.Vellido et al. (1999) investigated consumers' opinion on online purchasing and online vendors that seem to consist of the underlying dimensions ''control and convenience,'' ''trust and security,'' ''affordability,'' ''ease of use,'' and ''effort/responsiveness.''Using these dimensions as a segmentation base discerns seven segments: ''unconvinced,'' ''security conscious,'' ''undecided,'' ''convinced,'' ''complexity avoiders,'' ''cost conscious,'' and ''customer service wary.''Starting from consumers' motivations to use the Internet, McDonald (1996) segmented the Internet audience as ''avid adventurers,'' ''fact collectors,'' ''entertainment seekers,'' and ''social shoppers.''Also, Brengman et al. (2005) performed a cluster analysis based on seven factors, such as "Internet convenience", "perceived self-inefficacy", "Internet logistics", "Internet distrust", "Internet offer", "Internet window-shopping".

Research Context
The present research starts with the problem definition, and in this case, it refers to a detailed understanding of the customer's perception and customs regarding the usage of social media, namely how the students of "Lucian Blaga" University of Sibiu engage on social networking sites.
Furthermore, the main purpose, the objectives and the hypotheses were established, as follows:


The main purpose: Determining the students' underlying characteristics in terms of social media engagement.


Objective 1: Determining different types of respondents, based on their online activities.
 Hypothesis 1: At least three types of different users of social media will be identified.


Objective 2: Determining different segments of respondents.


Hypothesis 2: At least three clusters of different users of social media will be identified.


Objective 3: Develop a linear model for studying the impact upon positive reaction to online advertisements  Hypothesis 3: Social media user classification has a great impact in forming a favorable opinion for online ads.

Measurement and Research Instrument
Data collection was achieved by using an online consumer survey.The data was collected with an online survey targeted at the students of Lucian Blaga University of Sibiu studying for a Bachelor, Master or Ph.D. Degree.A short description of the survey and a link-address was posted on the discussion board of the University's website.
All measurement items were newly formed and were aimed at studying students' behavior and reactions on social media websites in order to provide a better understanding of their interactions in an online environment.Table 1 provides the psychometric properties of the measures.
Considering the fact that the scale items were newly compiled, we measured the reliability of the scales used to evaluate the internal consistency of the constructs.Reliability is identified by Cronbach's alpha with a minimum of 0.70 (Cronbach, 1970;Nunnally, 1978).The ads that appear on my profile are relevant for my personal interests and I enjoy seeing them.
Strongly Agree -Agree -Indifferent -Disagree -Strongly Disagree .801 Quite often I access the ads that I see on my social media profile.Experience using social media For how long have you been using social media websites?

Data Collection and Sampling
This study implied a primary research which involved getting original data by conducting a field research.In this case, the information was collected directly from respondents via the internet, from September 17 to November 18, 2011, and the data analysis is quantitative (Dumitrescu, Stanciu, Tichindelean, & Vinerean, 2012).Also, this paper is based on an exploratory research whose primary objective is to provide insights into a marketing phenomenon, namely students' pattern of using social media and social networking sites, and particularly in relation to their reaction to advertising in a medium where they decide and choose the information they engage with.The responses were collected from students from "Lucian Blaga" University of Sibiu, a higher education institution in Romania (Dumitrescu, Stanciu, Tichindelean, & Vinerean, 2012).
The sample size was computed by the formula n= z 2 *s 2 / e 2 , where z is a known tabular value for a specific level of significance, s represents the sample standard deviation of the selection variable (number of students by their year of study and study cycle) and e is the standard error.A confidence level of 95% was chosen, therefore the z 0.95 is 1.96 and the sample standard deviation is s = 28.65 students by their year of study and study cycle.The chosen sampling standard error is e = 3.66 students by their year of study and study cycle.By applying the formula, we have determined a sample size of n = 235.51= 236 students.The sampling process was executed by applying an online questionnaire to 236 students (Dumitrescu, Tichindelean, & Vinerean, 2012;Dumitrescu, Stanciu, Tichindelean, & Vinerean, 2012).

Factor Analysis
In the data analysis phase of the research, the data was collected via the Internet (FreeOnlineSurveys), and for the information's preparation and processing the researchers used the statistical analysis program SPSS (IBM SPSS, 2011) and firstly the Factor Analysis method.
Factor analysis represents a class of procedures primarily used for data reduction and summarization by identifying the latent variables.A factor is an underlying dimension that explains the correlations among a set of variables.As an extraction method, the researchers used the Principal Components method.In second step, the factors are rotated to ease interpretation.Varimax is the rotation method used most frequently with survey data, because it is a method of orthogonal rotation, which tries to load a smaller number of variables highly onto each factor resulting in a more interpretable and more relevant factors (Field, 2005, pp.630-636).As a clustering criterion, the Schwartz's Bayesian Criterion (BIC) was used.
Additionally, we used the Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO) in order to examine the appropriateness of factor analysis.High values (between 0.5 and 1.0) indicate that the factor is relevant.After undergoing the Factor Analysis, using the Principal Component Analysis and the Varimax Rotation Method with Kaiser Normalization, four factors emerged.These four factors represent the basis for an understanding of the students' activities on social media sites (Table 3).
The respondents that formed the first factor have been named Expressers and Informers.They get involved in the online environment but they are mostly focused on them, on providing information about themselves through blogging, Twitter and uploading wiki articles.However, Expressers and Informers are individuals who stay current, particularly by using the RSS, and then by staying current with different sources of information.
The second factor is entitled Engagers because they seek and read different forums and reviews, but they also get involved by posting comments and reviews, rate sites, products and services.They always what to know more, but they also want to let others know about their opinions regarding different subjects.
The third factor has been entitled Networkers or Socializers because they are particularly involved in social media sites like Facebook, Myspace.The Networkers are very vocal and engage in actions like updating their profiles regularly, posting comments to their friends and tagging pictures.
The final factor has been named Watchers and Listeners because it consists of internet users who have a minimum activity online.They only choose to engage in online activity that are entertainment-driven, namely watching movies, TV shows, videos, listening to music, and download music or video.
Similarly, factor analysis (using the principal components and varimax rotation methods) was also performed for other variables (Table1., namely "Trust in information from personal sources", "Trust in information from foreign sources", "Positive reactions to online advertisements") included in the survey in order to observe the students' underlying attitude regarding reactions to online ads and trust in information provided on social media websites.Three new factors were obtained and the general information about these new variables is presented in Table 4.

Cluster Analysis
The next stage involved using the TwoStep Cluster method in order to obtain segmentation, based on newly created factors.Clustering, or segmentation, is used in order to explore patterns of similarity or dissimilarity in different types of participants, i.e. to discover and describe groups of participants who may view particular forms of behavior in similar ways when considering different dimensions, which, eventually, concur to the quality of the higher education institution, "Lucian Blaga" University of Sibiu.We used the Two-Step Cluster procedure in SPSS (IBM SPSS, 2011), which incorporates statistical criteria to determine the optimal number of clusters, considering the number of observations from this database.
Also, the method allowed us to incorporate both continuous (the newly formed factors and the scales presented Table 3 and 4), and categorical variable (also, presented in Table 1).The results of the clustering procedure are presented in Table 5. 2 -3 years -55.3% The first cluster formed consists of 36.9% of the total respondents who do not really engage in social media activities, and have a negative reaction towards networking activities, expressing themselves or being a spectator online, perhaps due to their high level of privacy concern.Also, they have social media profiles for more than 3 years, they seem to log in several times a day and just browse around social media websites for 5 to 15 minutes.However, they have a positive outlook on online advertising and have frequently clicked on such advertisements with the intent to purchase or find out more information.
The students forming the second segment are Expressers and Informers on social media websites, who have been using for an average of two to three years, several times a day, for 5 to 15 minutes.They do not worry about confidentiality issues and they trust mostly their personal sources of information, such as friends and family in an online environment.The 32.2% of the total sample do not appreciate the intrusive online advertisements; however, they have clicked on profile targeted ads a few times.
The third cluster comprises the respondents who get involved in all sorts of social media activities, as they can be classified as Engagers, Expressers and Informers, Networkers, Watchers and Listeners.Similarly as the respondents from the previous clusters, they are using social media for more than 3 years and tend to log in several times a day, for an average of 5 -15 minutes per session.Also, even though these respondents do not have a favorable perception regarding online advertising, they seem to have clicked on social media ads.

Automatic Linear Modeling
Furthermore, a SPSS specific procedure was used, namely Automatic Linear Modeling (ALM), in order to forecast and model a continuous target variable based on linear relationships between the target variable and its designated predictors.8 continuous predictors and 4 categorical predictors were chosen in order to observe the impact on the target variable, namely, in this case, the positive reactions to online ads (Figure 1).The model selection method used was Forward Stepwise, which starts with no effects and then adds and removes them at each step, based on the Information Criterion (AICC).Also, the preparation of the data was automatic.The accuracy value is 51% and it represents the adjusted R 2 multiplied by 100.Hence, even though we included twelve predictors in order to study the target variable of positive reactions to social media advertising, the ALM procedure considered only seven variables: Clicking the ad, Experience using social media, Networkers, Watchers and Listeners, Concern for privacy, Log in pattern, Engagers.6 presents only the variables that are statistically significant according to the F test and does not include some predicting variables that would be considered to be insignificant according to this calculated statistic.In the information criterion, the Automatic Linear Modeling procedure used reduction in order to determine which predictors to add and which ones to eliminate from the final model, in order to increase the model's overall significance.
Regarding the seven variables (Clicking the ad, Experience using social media, Networkers, Watchers and Listeners, Concern for privacy, Log in pattern, Engagers) included in the modeling procedure, we can conclude that there is significant interaction between them and the target variable that studies the respondents' positive reactions towards ads displayed on social media websites.The coefficients show the relationship each predictor has with the target variable.Most notably, the students who actually often click on social media advertisements have a positive reaction to such marketing instruments.The model exhibits watchers and listeners who have a long experience with social networking sites, even though they have privacy issues.

Theoretical Contribution
Nowadays, the analysis of consumer behavior is central for marketing success, especially since most potential consumers are using the internet and different online socializing tools.The online audience is a booming market worldwide, however giving its globalized nature a level of segmentation is needed cross-culturally.Regarding the academic implications, our results contribute to the study of the field of Internet marketing.
The conclusions obtained from our research have important implications for the academic research, derived principally from the analysis of four new types of social media consumers, namely Engagers, Expressers and Informers, Networkers, and Watchers and Listeners.We used this new classification, other newly formed variables (Positive reactions to online ads, Trust in information from personal sources, Trust in information from foreign sources), continuous variables (Concern for privacy, Importance of social media), and categorical variables (Experience using social media, Clicking the ad, Log in pattern, Time spent per login session) to achieve a segmentation of social media users and observe different patterns which could be targeted to improve the effectiveness and efficiency of online marketing activities.Therefore, this research presents new ways to classify online consumers, which served as a basis for psychographic segmentation, based on respondents' activities on different online platforms.
Also, this study contributes to the existing knowledge of customer behavior in an online environment, in general, and on social media websites, in particular, by providing insight through an examination of seven influential variables on developing positive reactions to online advertisements.

Implications for Managers
Today, any marketer or business owner understands the importance of internet marketing.Marketing a business on the web implies leveraging social media to create a lot of buzz in relation to a brand.Social media platforms offer immense possibilities for fostering relationships with consumers in an online environment.This study suggests different approaches for online marketers and managers looking to invest in advertising on social networking sites and hence improve their ads' performance regarding clicking the advert and generating positive reactions towards it.One approach implies understanding the sources of trust in online information provided by sources social media users may or may not know, and how their concern for privacy influences their reactions to online advertising.
Managers should be aware of the importance of social media sites in influencing online shopping by identifying and targeting different types of customers and taking initiatives to recognize and highlight customer interests.
In order to be successful in social media marketing, companies need to create a buyer persona and then develop and constantly adjust the online marketing strategy according to the interests of customers for long-term success.
Figuring out what goes best for which particular audience leads to success.Regarding these aspects, companies can use online reputation systems in order to provide the right online incentives to the right online customers (Dellarocas, 2010).
True customer engagement means commitment-focused, not transaction-focused.The companies that master this aspect are the ones that are truly successful.They undergo continuous online marketing research and must be sensitive to changes in consumer behavior patterns and to identify new areas of consumer values and interest.

Limitations and Future Directions
Limitations of this study include those commonly associated with online questionnaires, including unsystematic sampling procedures and low response rates.While representativeness can always be improved, for the present research great efforts have been made in order to have a higher response rate for the sample.
This research is subject to some limitations which may provide fruitful avenues for future research.Firstly, certain limitations arise from the choice of the sample and the measurement of the variables used.Regarding the sample choice, this study presented responses gathered from students of Lucian Blaga University of Sibiu, Romania.The respondents do not, therefore, reflect customer attitudes and behaviors related to social media users in other countries.Moreover, given the sample, the research did not include in its analysis demographic variables, such as sex, age, social class, and ethnicity.Therefore, this is another area in which the research could How many times did you take action based on an advert you saw on social media (in terms of accessing the site or buying the productimportant do you think social media is for your social life?Very important -Important -Indifferent -Somewhat not important -Not important at all NA* *Cronbach's Alpha could not be computed because the variables are either nominal, ordinal or are formed of one single interval scale.

Figure 1 .
Figure 1.Concepts' relationship tested using automatic linear modeling

Figure 2 .
Figure 2. Predictor importance for the target variable in the automatic linear modeling procedure

Figure 3 .
Figure 3. Studentized residuals for automatic linear modeling procedure

Figure 4 .
Figure 4. Effects exhibited for the proposed model in studying the target variable

Table 2 .
Total variance explained for factor analysis

Table 3 .
Rotated component matrix of the newly formed factors

Table 4 .
Information regarding three new factors: positive reactions to online ads, trust in information from personal sources, and trust in information from foreign sources a These factors are redundant.

Table 5 .
Information that differentiate the clusters formed

Table 6 .
Analysis of variance for the proposed modelThe analysis of variance table presents information about the whole model, by examining how the independent variables interact with each other and what effects these interactions have on the dependent variable.Table

Table 7 .
Coefficients determined for the influential predictors for the target variable a This coefficient is set to zero because it is redundant.