Effects of social networking sites pdf

WOM may be endogenous because it not only influences new customer acquisition but is itself affected by the number of new customers. Likewise, traditional marketing activities may stimulate WOM; they should be credited for this indirect effect as well as the direct effect they may have on customer acquisition. We will empirically test for this endogeneity using Granger causality tests. We link variation in the number of newly acquired customers signups with the number of invitations referrals sent by existing members of the network to their friends outside the network.

The proposed model allows us to measure the short and long-run effects of WOM and to compare the effects of WOM with those of other marketing communications. We estimate a long-run elasticity of 0.

This is approximately 2. Hanssens et al For the company we study, WOM has a much stronger impact on new customer acquisition than traditional forms of marketing. In particular, WOM elasticity is about 20 times higher than the elasticity for marketing events 0.

We translate these findings into monetary implications by calculating how much the average newly acquired customer contributes to firm revenues. This provides an upper bound to the financial incentives the firm might consider offering to existing customers to stimulate outbound word-of-mouth. We note that the practice of seeding or stimulating word-of-mouth has grown rapidly in recent use but that quantifying the effectiveness or returns of this activity remains difficult e.

The authors found that WOM was seven times more effective than print advertising in influencing consumers to switch brands. Since the s, word of mouth has been the subject of more than 70 marketing studies Money et al Until recently, WOM research relied primarily on experimental methods versus studying actual consumer actions in the marketplace.

A major challenge in studying actual WOM is obtaining accurate data on interpersonal communications. Examining WOM on the Internet can help address this problem by offering an easy way to track online interactions.

The Internet, of course, gives only a partial view of interpersonal communication; WOM exchange is not limited to the online world. Recent research has begun to study WOM in an Internet setting. De Bruyn and Lilien observed the reactions of 1, recipients after they received an unsolicited email invitation from one of their acquaintances to participate in a survey.

Godes and Mayzlin suggest that online conversations e. In an application to new television shows, they linked the volume and dispersion of conversations across different Usenet groups to offline show ratings.

Chevalier and Mayzlin used book reviews posted by customers at Amazon. In contrast, Liu shows that both negative and positive WOM increase performance box office revenue. In an application to a web hosting company, the authors showed that customers acquired through marketing add more short-term value to the firm, but customers acquired by word-of-mouth added nearly twice as much long-term value.

Thus, the response of customer acquisition to WOM and traditional marketing activity could not be directly estimated. Our paper differs from above studies in research question and application. First, we aim to directly compare the dynamic effects of word-of-mouth referrals with those of traditional marketing efforts. In so doing, we will also attempt to quantify the monetary value of each WOM referral to the firm.

Second, our empirical application is to an Internet social networking site which provides a set of substantive findings in a novel marketing setting.

Internet Social Networking Sites While still a relatively new Internet phenomenon, online social networking has already attracted attention from major Internet corporations.

Microsoft, Google, Yahoo! According to Wikipedia www. Social Networking Sites Ranking] A social networking site is typically initiated by a small group of founders who send out invitations to join the site to the members of their own personal networks.

In turn, new members send invitations to their networks, and so on. Hence, invitations i. WOM referrals have been the foremost driving force for sites to acquire new members. Typical social networking sites allow a user to build and maintain a network of friends for social or professional interaction.

In the core of a social networking site are personalized user profiles. Other forms of relation formation also exist. In contrast to other Internet businesses, online communities rely upon user-generated content to retain users.

Figure 1 shows how two popular social networking sites, Friendster. Referrals Process at Friendster. Some sites offer incentives to make a referral. For example, Netflix. Many subscription-based services offer progressive discounts on monthly fees for each referral made. While the mechanics of social network formation through the WOM referrals process may be straightforward, little is known about the dynamics and sustainability of this process.

Also, as social networking sites mature, they may begin to increase their use of traditional marketing tools. Management therefore may start to question the relative effectiveness of WOM at this stage. Our objective is to contribute a new set of empirical findings on this subject. Modeling Approach A social networking site has several ways to attract new customers, including event marketing directly paid for by the company , media appearances induced by PR and word-of- mouth WOM referrals.

To model the effectiveness of these communication mechanisms, we turn first to a time series regression approach. As a benchmark model, we may regress signups on events, media and WOM, controlling for deterministic components such as a base level constant , a deterministic time trend, seasonality and lags of the dependent variable Box and Jenkins Seasonal patterns may be both high frequency e. Note that equation 1 includes only the immediate effects of marketing actions on signups.

To include dynamic effects, we can add lags of the marketing actions, obtaining the following autoregressive-distributed lag ARDL model e. While model 2 now captures dynamic effects, it does not account for indirect effects of marketing actions on performance.

For example, events may directly increase signups, receive media coverage indirectly benefiting signups , and increase the likelihood that current customers refer others to the site. These new customers may, in turn, invite their friends to join the site WOM. Figure 2 displays this system of plausible interactions, which may occur immediately i.

Modeling Framework] The links represented in Figure 2 can be tested by investigating which variables Granger cause other variables Granger , Hanssens et al.

We perform a series of Granger causality tests on each pair of key variables. We note that a wrong choice for the number of lags in the test may erroneously conclude the absence of Granger causality e. Because we are applying these tests to investigate the need for modeling a full dynamic system, we are not interested in whether variable X causes variable Y at a specific lag, but in whether we can rule out that X Granger causes Y at any lag.

Therefore, we will run the causality tests for lags up to 20 and report the results for the lag that has the highest significance for Granger causality. To this end, we specify and estimate a vector-autoregressive VAR model. Compared to alternative specifications, VAR models are especially well suited to measure dynamic interactions among performance signups and marketing variables and to estimate the dynamic response of signups to both WOM and traditional marketing actions.

Recently, VAR models have been used to analyze a wide variety of long-term marketing effects — including advertising, price promotions and new product introductions e. Note that in equation 3 the vector of endogenous variables -- signups Y , WOM- referrals X , media appearances M and promotional events E -- is also related to its own past, which allows complex dynamic interactions among these variables. In the absence of cointegration i. VAR modeling is commonly applied to quantify short- and long-run market response Dekimpe and Hanssens We note two features of this approach.

First, the endogenous treatment of WOM implies that it also is explained by its own past and the past of the signups variable. In other words, this dynamic system model estimates the baseline of each endogenous variable and forecasts its future values based on the dynamic interactions of all jointly endogenous variables. Permanent effects are possible for evolving performance variables, and statistical criteria such as the Akaike Information Criterion AIC suggest lag lengths J that balance model fit and complexity Lutkepohl Second, dynamic effects are not a priori restricted in time, sign, or magnitude.

The sign and magnitude of any dynamic effect need not follow any particular pattern — such as the imposed exponential decay pattern from Koyck-type models see Pauwels et al. Testing for Evolution or Stationarity: Unit-Root Tests We perform unit root tests to determine whether the endogenous variables are stable fluctuate temporarily around a fixed mean or evolving have no fixed mean and can deviate permanently from previous levels. The results of the unit root analyses will subsequently affect the model estimation procedure.

Convergent conclusions of these two tests yield higher confidence in our variable classification Maddala and Kim In our case, results of both tests confirmed trend stationarity in all series i.

Thus, we conclude that VAR estimations can be performed with the variables in levels. Impulse Response Functions Because it is infeasible to interpret estimated VAR-coefficients directly Sims , researchers use the estimated coefficients to calculate impulse response functions IRFs.

The IRF simulates the over-time impact of a change over its baseline to one variable on the full dynamic system and thus represents the net result of all modeled actions and reactions see Pauwels for an elaborate discussion.

We adopt the generalized IRF, i. This uses information in the residual variance- covariance matrix of the VAR model, instead of requiring the researcher to impose a causal ordering among the endogenous variables Dekimpe and Hanssens In the context of our research questions, we use impulse response functions to disentangle the short and the long-run effects of WOM and traditional marketing on signups.

This also follows the tradition of VAR-related research published in marketing. The dataset contains 36 weeks of daily numbers of signups and referrals provided to us by the company along with marketing events and media activity obtained from 3rd party sources. The data cover the period from February 1 to October 16, Figures 3 and 4 show time plots for all four variables and Table 2 provides descriptive statistics.

Time Series: Signups, Invitations] [Figure 4. Time Series: Media and Marketing events] [Table 2. Descriptive Statistics] During the observation period, the daily signups and WOM-referrals showed a positive trend.

We observed somewhat lower activity in referrals over the summer season as practiced in the U. Over the 36 weeks, the company organized or co-sponsored promotion events.

On some days, multiple events occurred in different locations. Overall, 86 days in the observation period had some promotion event activity. Finally, we identified appearances on days of the company name in the media. We considered different sources, both electronic and traditional media, as provided by Factiva News and Business Information services www.

We did not use the content of these publications; thus, our measure of media activity is coarse. In a more general case it would be important to account for the valence of the message as Godes and Mayzlin report for TVshows.

Direct Effects of Marketing on Signups We first present the results from estimating the benchmark time series regression models given in equations 1 and 2. These models regress signups against WOM referrals, media, and events, while controlling for a time trend, day-of-week, and seasonality. Results from both models are presented in Table 3. The left column of parameter estimates corresponds to the specification for immediate effects only equation 1 while the right column corresponds to the carry-over effects specification equation 2.

The models are estimated in log form, which provides direct results for the elasticities from the estimated coefficients. Moreover, we find similar effect magnitudes for the two models. WOM has the largest elasticity 0. Because all dynamic effects equation 2 and potential interaction effects results available upon request are insignificant, adding them does not change our substantive findings. Each cell gives the minimum p-value obtained from the causality tests conducted from one lag to 20 lags.

As expected, Granger causality is detected for WOM referrals, media and events on signups the direct effects. In addition, Granger causality is also found for many of the other pairings.

For example, signups Granger cause WOM referrals the interdependent effect argued earlier , events indicating management performance feedback, e. Dekimpe and Hanssens , and media indicating that spikes in signups may receive media attention.

Moreover, events Granger cause media indicating that media covers events and media Granger causes events indicating that management may seek to time events to match pending media coverage. On the other hand, WOM-referrals do not Granger cause events or media appearances as the media does not observe referrals directly and media appearances do not Granger cause WOM.

In sum, the results from the Granger causality tests indicate the need to consider the full dynamic system, as in a VAR-model, and account for the indirect effects of marketing actions. Next, we add day of the week effects, and then a holiday effect.

The model fit results are presented in Table 5. The AIC criterion suggests that the best model includes all of the proposed effects. Finally, we note that the AIC criterion selects 2 as the optimal lag length. These enable us to examine the dynamic effects of each activity on signups, fully accounting for the indirect effects of these activities. Figures 5a, 5b and 5c plot these impulse response functions. The graph shows that it takes approximately three weeks for signups to stabilize after a one standard deviation shock on referrals WOM.

In , more than 1. The unlimited gain traction Anonymous, Today, 35 — 44 years old dimension of the use of the social media among the people have increasingly joined the population and youth of today. Social networking media have become counted as joiners, spectators, and critics. Published by Atlantis Press. They have helped many people feel communication. Slaughter said that social as if they belong to a community and make connection networking media have deeply transformed the not only on campus but with friends outside of school.

Results many aspects of their live and form of large proportion showed that social media is a college interest in the of users on social media networks such as Facebook, Wales University. Students consume a lot of time on Twitter, Instagram, MySpace and many more Guy, social networking sites especially the adolescents.

Blogs, Twitter, chatting with friends and watching online movies MySpace, and Linkedln had significantly lower Kanagarathinam, Twitter is one of the most amounts of student users. The students from the important media companies in the world.

It is now one business school had the highest percentage of users of of the top 10 most visited sites. With more than blogs, Twitter, and Linkedln while Liberal Arts million users, the students can find teachers, schools, students were the highest percentage of My Space and perhaps more important professionals, lead users Martin, They become very smart because of the other classes or even schools. Blogs can highly information they get from these sites and it easy to get motivate the students, especially those who otherwise almost any materials for school assignment.

But some might not become participants in the classrooms. It students become very poor academically Egedegbe, gives excellent opportunities for students to read and According to the research of Brubaker , write effective forums for collaboration and discussion, the current generation of college students has been and powerful tools to enable scaffolding learning or exposed to a technology which led them to rely on mentoring to occur Nguyen, According to the social media such as Facebook and Twitter.

It has an study of Head and Eisenberg conducted in , impact on the academic performance when students Wikipedia provides the students with a summary about overuse or multitask while doing their school work. As Wikipedia and they know its limitation. Because of its an effect, most of the users have the good quality of quick way to get started, they use Wikipedia just as communication skills through the use of social most of people do but not deep and credible.

They were able to communicate with networking media is a basic asset that a college student friends and family by posting information and they nowadays must possesses in order to survive in his spent much of the time viewing information. The face- chosen volition. In this regard, a question may be to-face interactions via the computer facilitate asked: Is any use of social networking media beneficial communication which allows users to keep in touch to students?

Unfortunately, studies revealed different with family and friends in a convenient way and to results. Accordingly, the use of technology such as learn about social events and issues. They depend on the type of SNS the student is using, if can share topics online. Study groups and students can student uses the internet for the purpose of leisure use it to connect with each other outside of the activity that interferes with academic, it will affect the classroom while the bad effect is that students may student academic performance negatively Egedegbe, want to add as friends to make trouble or some According to the study of Tayseer, et al.

Also, the students can also access to , result showed that most of the students who a world of knowledge through the use of social spend many hours in using social networks have a high networking media and other forms of digital GPA at the Petroleum University. Specifically, it sought consider giving up social media for a slightly better to answer the following questions: 1 What are the GPA. However, What are the perceived effects of social networking negative effects abound.

Two step provide them a first-hand information on how social flow interactions, student to student and teacher to networking media affects the academic performance of student favored academic learning through social the students in the University, in general, and in the networks.

For clear perspective on how the specific behavior of example, the National Association for College students is affected by social networking media. To the future In sum, we can say that there are benefits and researchers on social networking media, the results risks associated with using any social network even would serve as a baseline data for them to conduct the though there have been reports regarding its effect on same study in order to validate the findings.

Accordingly, some researchers found a poor effect and influence when the Scope and Limitation of the study media is overuse in such a way that do not Due to time and financial constraints, this academically improve learning or its process. There are study recognizes many limitations not only in terms of still other researchers who examined this same problem its scope and focus but also in its statistical tools, time but have found no conclusive data affirming the frame, sampling methods, and others.

On its scope and significant relationship between using social time-frame, it focused only on the effects of social networking and student academic performance Al- networking media to the academic performance of Rahmi and Othman, Moreover, only basic of the effects of social networking media on the statistics were employed since the study is descriptive academic performance of students, it is timely and an in nature.

Finally, the purposive sampling was imperative necessity to study on whether or not the employed in choosing the respondents of the study. It is accessible by land transportation from any of the four Sampling Procedure main cities in Mindanao: kilometers to Cotabato Three types of sampling methods were used in City, kilometers to Davao City, kilometers to this research.

As previously USM is claimed to be the pioneering land- mentioned, the student-respondents are selected due to grant University in southern Philippines. It was their most accessibility. Proportionate stratified random formerly given the name Mindanao Institute of sampling was employed in getting the prescribed total Technology MIT which was founded by the late Bai number of respondents from the two sample academic Hadja Fatima Matabay Plang, the wife of former departments, that is, 81 from the Islamic Studies Senator and retired Brigadier General Salipada K.

Department while only 10 from the International Pendatun, which became operational on October 1, Relations Department. The third and the last type of It achieved a university status on March 13, sampling method is the systematic random sampling by virtue of Presidential Decree No.

At present, the from the list of enrolees available in the Institute. This University is consisting of nine 9 colleges and two is done by selecting the names of the respondents from 2 institutes, including the Institute of Middle East and the list of enrolees by having five interval names for Asian Studies IMEAS. At present, the Institute is every draw until the desired total number of consisting of only two 2 academic departments, viz: respondents from every Department had been chosen. International Relations and Islamic Studies.

In the First Semester, , the former had one-hundred Research Instrument eighteen enrolees while the latter had nine- In gathering the primary data, the only hundred nineteen It consisted on two parts.

Part 1 contains II. In trying to look for the students. The second type of data was gathered from the USM President through the registrar asking written materials available in the different libraries in approval to get the Grade Point Average of the students the University as well as those found on-line.

Moreover, correlation analysis was used to interpret the Second, the researchers conducted administered the relationship of the socio-demographic characteristics of pretesting of the questionnaire for validation of the said the respondents and the effects of social networking instrument. Third is the sampling and actual media to their academic performance. Questionnaires were retrieved right away Respondents of the study after the respondents have answered them all.

The respondents of the study were the ninety- one 91 students from the IMEAS who were taken as Statistical Analysis samples from the one thousand and thirty-seven The statistical tools used in the study were the students who were officially enrolled in the first descriptive statistics such as percentage, frequencies, semester of the school year — They were and averages. Where: III. This sex, the study gathered that female respondents means that much of the time of the respondents was dominated the samples who constituted This is not surprising since nowadays, the activities.

In network to the respondents, the data disclosed the France, eight centenarians in ten are women following in rank order : 1 helps them to become Anonymous, This output may imply quarter of