Amsterdam Business School
Bachelor Economics and Business
Specialization: Business Administration
“To what extent does critic as well as peer review valence influence consumer loyalty?”
BSc Thesis by
Kestutis Zakarauskas
10860983
Supervisor: Frederik Situmeang
Statement of Originality
This document is written by student Kestutis Zakarauskas, 10860983, who declares to take full responsibility for the contents of this document.
I declare that this text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work and its contents.
Table of Contents
I. Introduction ……….……….……… 1
II. Theoretical Framework ………...………...………. 4
2.1 Review Valence ………..………...………. 5 2..1.1 Critic Reviews...………...…………..………...…………. 5 2.1.2 Peer Reviews………...…………..………...……. ……… 10 2.1.3 Prospect Theory………….………...………..………... 12 2.2 Review Volume……….12 2.3 Variability……….14
2.3.1 Macro- level volatility………15
2.3.2 Politics and volatility………..16
2.3.3 Firm- level volatility………..17
III. Methodology ………..………. 19
3.1 Procedure & measures ……….……….… 19
3.2 Dependent and independent variable ………...……… 20
3.3 Moderator variables………22
3.4 Control variables ………....………..23
IV. Results ………....……….24
4.1 Descriptive statistics and correlations…….………. 24
4.2 Regression…………..………. ………. 25
V. Analysis ………....………30
5.2 Theoretical Implications ……….………. 34
5.3 Practical Implications ……….………. 34
5.4 Limitations and Future Research ………....………. 35
5.5 Conclusion……….………36
Abstract
It has been acknowledged in the marketing literature that electronic word of mouth (eWOM) has had a significant effect on consumption decisions, which in turn provides new ways for firms to compete. There is an extant amount of past literature acknowledging the fact that online consumer and critic reviews shape motion pictures’ box office. This paper was
particularly interested in the relation between consumer as well as critic reviews and consumer loyalty. Since loyalty implies consistent box office revenues over time, it may prove vital for firms’ survival.
In order to determine whether such relationship exists, this study also looked at the moderating effect of community size and volatility of user reviews. Even though the interaction effects were not supported, a significant direct effect between user community size and loyalty was revealed, which in turn was supported by the signalling theory. Moreover, critic reviews were found to have no effect on moviegoers’ loyalty, which is in tact with the prevalent movie studios’ interests.
This research was conducted in the Hollywood motion picture industry. The data was retrieved from http://www.metacritic.com and http://www.boxofficemojo.com. In particular, the sample size of 158 motion pictures corresponding to series was analysed by means of multiple regression analysis in SPSS.
Finally, after providing the results, a robust analysis was carried out by means of explaining the findings based on the overview of the previous literature following the implications for future research and limitations of the study.
1
I. Introduction
With the dawn of the internet era, electronic markets came into existence, which resulted in the unbundling of abundant information about goods and services across various markets (Bellman et al., 2006). Consequently, the web has made the impossible possible; by developing new ways of communicating, introducing new possibilities of doing business universally and contributing to the acceleration of globalisation, it has radically changed the rules of the game (O’ Connor, 2008). Due to successful developments in the B2B setting, P2P interactions have gained unprecedented momentum resulting in an improved access to user of information all over the globe (O'Connor, 2008). As a result, online reviews have become increasingly available in a variety of settings, completely changing the conventional nature of consumerism. For instance, instead of physically going to a store to acquire a good, modern- day consumers can just look up the product online and base their purchase decision on the reviews available about the product in question. Similarly, before seeing a movie, moviegoers can look up the reviews online and make a decision accordingly, based on user and/ or critic reviews.
There is an extant amount of research carried out on consumer loyalty and how it is affected by the importance of brand equity, identity salience, emotional satisfaction and even the prominence of institutional image. However, there is very little knowledge about the relationship between online reviews and respective loyalty motives. In particular, the majority of marketing literature controls for the fact that the product is a sequel (Sunde and
2 Brodie, 1993; Karniouchina, 2011). Consequently, this kind of research usually reveals that sequels perform better in the market than their predecessors due to the carry- over effects of the original products (Ho et al., 2009; Karniouchina, 2011; Sood and Dreze, 2006; Sunde and Brodie, 1993). For instance, if the original product is successful, it creates a positive image amongst the consumers and, as a result, this success is carried over to the subsequent sequels (Situmeang et al., 2013).
This research analyses the effect that both consumer and critic reviews have on consumer loyalty. In the context of motion pictures, the key variable in determining movie success is motion picture sales, which is also known as the box office revenue. As a result, consumer loyalty in this research is computed in terms of the change in worldwide box office revenues.
First, this paper provides a salient account of the theoretical knowledge regarding the independent variables: consumer review valence, which predicts success of a sequel based on the rating of a pre- sequel, and the significance of critic scores. Since both variables are expected to have a direct effect on consumer loyalty, this part delves deeper into the analysis of online reviews in the movie industry. In particular, the effect on loyalty of positive and negative consumer reviews is analysed. Then, signalling theory is introduced in order to comprehend how larger communities of movie enthusiasts affect the relationship between consumer review valence and consumer loyalty e.g. whether the moderating effect exists and if so, whether it is positive. Afterwards, an overarching discourse about variability in
different contexts is presented in order to comprehend the significance of variability not only for this research but also to understand its wide implications in a variety of contexts. Moving on, in the methodology section, a detailed procedure of measuring the independent,
3 dependent and moderator variables is provided together with a separate section for control variables. Then, results of the research are discussed with the help of quantitative analyses and descriptive statistics. Further, a comprehensive discussion of findings, practical as well as theoretical implications and limitations and, most importantly, contribution to future research are provided. Finally, the conclusion is drawn in order to fully grasp the research design and the findings’ implications for practice as well as the scientific community.
4
II. Theoretical Framework
Word of mouth (WOM), which is usually defined as mutual communication about products and services between consumers, is one of the most substantial sources of information for consumers (Alreck & Settle, 1999). Its influence stems from the fact that consumers trust each other more than they trust marketers or salesmen (Sen & Lerman, 2007). eWOM occurs on the online platforms such as retailers’ websites, brands’ websites or social networking sites (Bickart & Schindler, 2001). According to McKinsey research, about two-thirds of the consumer- touch points entail consumer-targeted marketing activities, such as internet reviews and eWOM recommendations. Chatterjee (2001) contests that eWOM is anonymous in nature and in turn, it allows consumers to share information honestly leading to enhanced volume of eWOM. Moreover, the scholar reveals that eWOM is available in large quantities as opposed to similar sort of information that is available offline. This increases the chance of consumers finding more experienced consumers of the product or service (Sen & Lerman, 2007). Further, online review valence indicates whether reviews are predominantly positive or negative (Purnawirawan et al., 2015). This is of paramount importance to eWOM of both consumer and critic reviews, as it determines the nature of the respective message being communicated e.g. how it is going to influence consumers’ buying intention. The latter discussion leads us to the following research question: ‘How is consumer loyalty affected by online review valence under the moderation of review volume and variance?’
5 2.1 Review valence
Consumer reviews have only become available with the dawn of the internet era as opposed to expert reviews also known as critic reviews, which have been widely available in media and newspapers for a long time (Situmeang, Leenders & Wijnberg, 2014).
There is a vast amount of evidence indicating that the role of critics is of paramount
importance in different industries. For instance, in finance field critics, defined as financial analysts, play a prominent role in determining profitable portfolios of stocks for many
different investors around the globe. Surprisingly, the role of experts is the most critical in the movie industry (Holbrook, 1999; Basuroy et al., 2003; Eliashberg and Shugan, 1997).
According to the Wall Street Journal (2001), almost 50% of Americans seek a movie expert advice when contemplating a movie decision. By the same token, movie studios engage in a strategic management process when it comes to expert reviews by selectively publicizing favourable critic reviews in their advertising campaigns, signifying the importance of such reviews (Wall Street Journal, 2011).
2.1.1 Critic reviews
Reinstein & Snyder (2005), discussed several motives for delving deeper into the critic reviews. First, the authors established the fact that critic reviews have a significant impact on the demand for experience goods e.g. goods whose quality cannot be determined prior to consumption. Then, the scholars emphasized the feature of non- substitutability of expert reviews as they are perceived as unique and independent from other parties, such as the movie studios or big enterprises. As a result, independent nature of critic reviews signals authentic quality to the moviegoers resulting in the demand stimulus for such critically acclaimed movies.
6 Throughout the research about the critic reviews in the movie industry, it was found that motion picture experts have two primary roles: predictors, if they manage to predict moviegoers’ tastes as well as the preferences and influencers, given that they manage to influence the decision of the moviegoers in the early weeks of the film release (Basuroy et al., 2003; Eliashberg and Shugan,1997; Gemser, Oostrum and Leenders, 2006).
The first role can be viewed as the critics’ ability to predict the box office, and the second role as the capability to influence it. In this research, the main interest lies within the second role- the ability to influence movie sales. Accordingly, Weiman (1991) defines the role of influencer as a person who is viewed by other people as having credible knowledge or expertise in a particular matter. Following this definition, if an influencer of the group sets an opinion in motion, the rest of the group is expected to follow it. Thus, an influencer is
expected to influence the moviegoers’ opinion with respect to seeing the movie or not. In line with Weiman’s (1991) findings, Eliashberg & Shugan (1997) conducted a thorough research on the critic’s role in the motion picture industry. The scholars’ findings indicate that early critic reviews are crucial not only to the success of the movie but also to the survival. In addition, their research results revealed that critics tend to see themselves as influencers and, as a result, signal their importance to the studios, which have to time the pre- release date accurately in order to attain favourable reviews, which in turn determine attendance of the moviegoers that is directly linked to the motion picture box office.
Although this research is mainly interested in the critics’ ability to manoeuvre the box office, critics role as predictor should not be overlooked. Predictors do not play a major role in determining the motion picture sales and, thus are perceived as secondary to influencers (Basuroy et al., 2003; Eliashberg and Shugan,1997; Gemser, Oostrum and Leenders, 2006).
7 Au contraire, the predictor reviews may determine whether the spectators will
ultimately enjoy the film (Eliashberg & Shugan, 1997). The latter finding is pertinent to this research, as consumer loyalty is based on consumers’ satisfaction with the movie e.g. satisfied consumers are more likely to share positive WOM and, as a result attract more viewers. Moreover, critics’ role as predictors is perceived as more honest as compared to the influencer role (Eliashberg & Shugan, 1997). This discrepancy arises because predictors are usually representative of their audiences and consequently provide an average opinion of the community, whereas influencers may change their opinions in order to adapt to dynamic contexts and ultimately survive, as their credibility rests on their popularity among the masses (Eliashberg & Shugan, 1997).
As a result, H1 a) is developed:
H1a): Critic review valence has a positive effect on consumer loyalty.
However, both predictor and influencer role have the strongest effect in the days of pre-release due to the nature of critic reviews, which are usually written in advance to the official motion picture release (Basuroy et al., 2003; Eliashberg & Shugan, 1997; Liu, 2006). In the research conducted by Liu, it was found that moviegoers’ anticipation is usually higher a few days before the official release and, once the movie is released, the audience becomes more critical of it and, as a result, the box office revenue plummets (2006). In addition, the effectiveness of the movie reviews tends to diminish over time because of the word of mouth communicated to potential audiences by the moviegoers who have already seen the movie (Eliashberg & Shugan, 1997). Thus, initial expert reviews are most pertinent when discerning their influence on the motion picture sales.
8 As a result, a great array of movie studios tries to attain favourable critic feedback by means of showing movies to motion picture experts before the official release date, so that expert reviews can be published on social media platforms, news websites and various
advertisements (Brown, Camerer and Lovallo, 2013). Unfortunately, the scholars found that a lot of Hollywood movies are intentionally withheld from the critics before the official
opening, a procedure also known as “cold opening”. The premise behind “cold opening” is that rational spectators trust the studios to know the quality of their movies and therefore, cold opening may send a signal of the mediocre quality of the cold opened motion pictures (Brown et al. 2012). However, Brown, Camerer and Lovallo’s (2013) regression results indicate that cold opened movies result in a 35% to 55% increase in revenue for opening weekend in America as well as the cumulative box office. These results infer that poor quality movies could boost their revenue by almost 60% by purposely withholding movies from the critics. Additionally, the scholars concluded that the overall quality of moviegoers has significantly diminished over the years as the box office of cold opened movies has dramatically increased, inferring the rise in demand for such films. In line with Brown, Camerer and Lovallo’s (2013) findings, Latwel and Shamsie’s empirical results revealed that movie studios engage in the cold opening primarily due to the expectation of negative results (2000). As a consequence, such practices may irritate the critics, which would eventually lead to unreasonably harsh reviews upon the official movie release date (Latwel & Shamsie, 2000). The latter findings suggest that the rise in popularity of cold opened movies may have a detrimental impact on movie studios’ reputation as well as credibility. This discussion emphasises the fact that critic reviews have been increasingly overlooked within the moviegoers nowadays, possibly because of the critics’ tendency to remain biased or easily
9 influenced by the movie studios’ relative influence in the movie industry (Brown, Camerer and Lovallo, 2013).
On the other hand, as discussed by Eliashberg & Shugan, the impact of the review on moviegoers is the most pronounced when the review is encountered for the the first time (1997). Thus, critics reviews may have tremendous impact on the motion pictures
success. However, in this paper the movie cycle is not taken into account when testing for relation between expert reviews and the moviegoer loyalty, as it is beyond the scope of this research and, as a result, it is ignored when considering the methodology. Additionally, this research draws on the review valence instead of the qualitative nature of reviews, because the former is perceived as a more objective criterion given that it is measured on a quantifiable scale e.g. 1 to 5 or 10 to a 100 etc.
So far, the discussion regarding expert reviews was situated in the American film industry e.g. Hollywood motion pictures. However, different findings were obtained in the Dutch film industry. In particular, Gemser, Oostrum and Leenders (2006), discovered that due to the fundamentally different nature of motion pictures, there are significant differences in the influence of expert reviews on the moviegoers. Notably, the scholars revealed that consumption of art house movies is driven by the influence effect e.g. moviegoers rely on expert reviews when making a decision, whereas mainstream motion picture fans are not affected by expert reviews but rather other sorts of information. Therefore, in the latter case the consumer’s decision to attend a movie in question is not influenced by the critic review, which is merely a reflection of the box office fruition (Gemser, Oostrum and Leenders, 2006).
10 2.1.2 Peer reviews
According to a study, conducted by a prominent search agency BrightLocal (2013), almost 80 percent of consumers trust online reviews as much as they trust each other. Equivalently, two different studies conducted by Kelsey group and ComScore reveal that 78 percent of users claim that they trust an online peer review more than a professional review, and that their willingness to spend is greater for five star rated products as opposed to four star or worse rated ones (ComScore, 2007). Therefore, online peer reviews and ratings are of paramount importance for understanding the relationship between eWOM and consumer loyalty.
Only recently online websites such as Metacritic, IMDB, online retailers such as Amazon, Ebay and BN as well as online platforms in diverse industries started collecting consumer reviews, and a fair amount of other studies succeeded in deriving the impact of consumer reviews on sales (Situmeang et al., 2013). For instance, Amazon made consumer created information- online consumer reviews- available already in 1955 (Park & Lee,2006). To this day, Amazon has over 12 million online consumer reviews available for all of its categories, with online consumer reviews being one of the most prominent and profitable features (Park, Lee & Han, 2007). In the hospitality industry, which has one of the largest online community applications, almost 80% of travellers take into account online user reviews when travelling (Gretzel and Yoo, 2008). In addition, just like motion pictures, hospitality and tourism are experience goods as their quality cannot be known before consumption (Ye, Law and Gu, 2009).
Dellarocas et., al (2007) found that online consumer reviews have a significant positive impact on movie sales. For example, Steven Spielberg’s movie “The Terminal”
11 accumulated the box office of only 77$ million, which in relative terms, as compared to the director’s other movies such as “The Green Mile” and “The road to Perdition” grossed over 120$ million dollars each. The main cause for such discrepancy in the box office was identified as unfavourable consumer reviews for “The Terminal” (Dellarocas et., al, 2007).
Because of this, peer evaluations are equally important as expert reviews in the movie
industry (Basuroy et al., 2003; Elberse and Eliashberg, 2003). Consequently, peer reviews are expected to have a positive effect on consumer loyalty, leading to the following hypothesis:
H1b): Peer review valence in the user community has a positive effect on consumer loyalty.
2.1.3 Prospect Theory
Loss aversion is a phenomenon that has been acknowledged in various fields such as psychology, politics and marketing literature. The term “loss aversion” was coined by
Kaheman and Tversky (1979) and could be explained by the following example. Consider an individual who is faced with the possibility of gaining 10 euros or, equivalently losing 10 euros. Loss aversion stipulates that an individual is more severely affected by the loss of 10 euros than by the equal gain e.g. negative utility exceeds the positive utility (Kaheman & Tversky, 1979).
Prospect theory predicts that individuals tend to pay more attention to losses than to comparable gains and, in general, the consumers tend to be risk- averse with respect to gains as opposed to risk- accepting behaviours when it comes to losses (Levy, 1992). In line with the prospect theory, a great amount of past research indicates that negative information is more influential, predictive and valuable than its counterpart for the majority of consumers
12 (Fiske 1980; Skowronski & Carlston, 1989; Taylor, 1991). This phenomenon is also known as the negativity bias. Furthermore, negative information is more diagnostic than the positive information because it is weighted more than the latter when making a buying decision (Levy, 1992). The latter claim is also supported by Lee, Rodgers & Kim’s (2009) study, where extremely negative consumer reviews had a stronger impact on the brand evaluations than moderately negative or extremely positive reviews, whereas extremely positive and moderately negative reviews had a similar level of influence on consumer buying intentions. As a result, negative reviews may disproportionately affect the service or product in question leading to deterioration of consumer loyalty.
Even though loss aversion is of significant importance to consumer loyalty, it is beyond the scope of this research to delineate the relationship between loss aversion and consumer loyalty.
2.2 Review volume
In the movie industry, films are regarded as experience goods; an individual has to
experience the movie himself in order to know whether he derives any joy from it (Caves, 2000). Moreover, consumer evaluations of experience goods reflect the extent to which the users are pleased with the product (Liu,2006). Content consumers are more likely to trust and remain loyal to the brand, which in turn increases the chance of future consumption i.e. consumer loyalty (Selnes, 1993). In addition, in the movie industry it was found that the community size has more explanatory power than the expert review valence on the movie box office revenue (Liu, 2006).
13 Consequently, since motion pictures are experience goods, moviegoers have to count on the feedback generated by the experienced peers e.g. people who have already seen a movie in question. Therefore, the two separate groups of moviegoers possess asymmetric information and, as a result, the movie fans who are contemplating a movie decision have to trust the experienced viewers. In particular, Stiglitz (2000) revealed that there are two major sorts of information where asymmetry is pronounced: information about intent and
information about quality. In the first case, one party is concerned about the other parties’ behaviour whilst in the latter case one party is not entirely aware of the characteristics of another party (Jensen & Meckling, 1976). The Nobel prize laureate Spence (1978) explained quality as the imperceptible ability of the individual, which is signalled to the labour market upon successful graduation. This is known as the signalling theory and has been used in a great array of Economic, Management as well as Marketing literature over decades (Connelly et al., 2011). Particularly, the meeker the ability of the individual consumer to evaluate the product or service in question, the more significant the presence of signals that are not directly derived from the product/ service will be.
Consequently, movies that have a large amount of user reviews send a credible signal to the unfamiliar viewers indicating that a movie in question has a stronger effect on
moviegoers than the movies with a less pronounced volume of reviews.
Furthermore, Duan, Gu & Whinston’s (2008) findings indicate that due to awareness effect, meaning that consumers become more conscious as they gradually accumulate knowledge online-review volume is significantly positively associated with sales in the movie industry. This finding implies that consumers are more likely to attend movies that have a larger number of reviews. However, attending a movie does not imply amelioration in
14 consumer loyalty. Therefore, in the theoretical model, the volume of reviews measures the community size e.g. the number of user reviews for each movie. Thus, the community size is expected to have a moderating effect on the review valence, such that:
H2: The relation between consumer review valence and loyalty is moderated by the peer
community size, such that larger communities tend to have a positive relationship between
review valence and consumer loyalty.
2.3 Variability
Volatility has been applied and robustly analysed in various fields such as ecommerce, economics, politics, marketing etc. In order to fully grasp the significance and possible consequences that volatility may have on consumer loyalty a detailed review of volatility in different contexts is provided below. In this paper, terms volatility and variability are used interchangeably.
A big variability of the reviews of the same object within the same community can be defined as a lack of consensus, among group members, with regards to the object (Situmeang et al., 2013). Variability in the community consensus has received little attention in the eWOM literature thus far. For instance, Sun (2012), conducted a study focusing mainly on the effect that product ratings may have on product quality and consumption. The latter study was conducted in the online book market where quality, defined as a high average rating for the product, and volatility described as variance in product ratings, helped depict the type of product. High product quality was found to be positively related with consumption. On the other hand, products with a high variance in ratings revealed that the product is niche e.g. the
15 product was not sought by en masse consumers and demand for the product in question would only increase if it’s respective average rating was low. Thus, volatility and quality were found to be inversely related, meaning that high variance in ratings implies a mediocre quality, which in turn reduces demand for the product in question.
2.3.1 Macro- level volatility
Kose, Prasad and Terrones (2003) discuss more pertinent consequences of volatility on the economy as a whole. The scholars revealed that severe crises, that had happened in the 80s and 90s, lead to volatility that has had severe ramifications on the developing economies’ welfare, which eventually lead to openness of trade and financial as well as capital inflows in these countries. Consequently, if the increased capital flows into developing countries’ economies is driven by specialization in different industries across countries and shocks caused in these industries are significant in driving business cycles, then it would lead to output volatility, which over time would result in an increase in the volatility of consumption (Kose, Prasad and Terrones, 2003). On the other hand, Razin and Rose (1994) research findings revealed that improved within- industry trade across countries would improve intermediate inputs trade, which would result in the decline of output, thus moderating the volatility of consumption.
According to economic theory, variability is multi-faceted and, nonetheless has an indirect effect on individual- level consumption (Kose et al., 2003; Razin & Roze, 1994). In addition, volatility in the aggregate demand may disproportionately affect consumption volatility in developing countries, which has a negative effect on welfare gains (Loyaza et al.,
16 2007). Therefore, by minimizing welfare volatility, developing countries may profit from growth in consumption, which would eventually lead to GDP growth.
In the latter context the root cause of volatility is uncertainty because future is not easily predictable and therefore may result in different scenarios. Cornelius et al. (2005) described a way, based on the case of oil giant Shell, to deal with uncertainty and therefore reliability- scenario planning. By taking into account the major shifts that could possibly occur in the business environment and developing credible narratives about the future, the scholars tried to challenge the prevailing mind-set. However, since we are faced with uncertainty on a daily basis, it would be rather impractical and thus wasteful to consider every possible outcome for every action we take. Eisenhardt & Sull’s (2001) findings revealed that by developing and implementing simple rules, organizational members may have freedom to execute decisions and, as a result firms may become dynamic and cope with uncertainty spontaneously. For instance, by setting boundary rules, the firm works its way to determine which opportunities should be pursued, and which should be abandoned.
2.3.2 Politics and volatility
Politics may have a severe impact not only on the society as a whole but on the firm level too. Eisenhardt & Zbaracki (1992) found that organization is comprised of various coalitions of people with conflicting goals where decisions are made in favour of the most powerful members. As a result, individuals tend to engage in politics to advance their power and attain their goals (Eisenhardt & Zbaracki, 1992).
17 In politics, electoral volatility is perceived as the opposite of party loyalty with the majority of parties aiming for low electoral volatility, which would result in the betterment of
democracy (Toka, 1998). Moreover, volatility is perceived as the workhorse of better institutions and party stabilisation mechanism. The meeker the volatility is, the more value the established parties have independent from their leaders’ influence, scandals and issue positions (Toka, 1998). Even a small amount of electoral volatility is sufficient to cause dramatic changes in legislative as well as governmental power. Unfortunately, the majority of democracies nowadays have much more than average levels of volatility, which in turn may lead to the weakness of party loyalty and, as a result, cause destabilizing regimes (Toka, 1998).
2.3.3 Firm- level volatility
In the context of B2B interactions, sales variance to a customer represents uncertainty of the revenues from the customer in question and, consequently sends a credible signal to the sales managers (Miller, & Robert, 2006). Businesses that tend to be more volatile are perceived as riskier by consumers and, as a result, such businesses receive less interest from consumers resulting in a fall in sales to the clients, which leads to a boost in sales volatility to the customers (Tuli et al., 2010). Further, the scholars revealed that diversity of customer ties increases sales, which in turn puts downward pressure on volatility in sales resulting in the improvement of overall sales to customers.
Fischer et al., (2014), researched volatility stemming from marketing related activities such as new- product launches, sales promotions and advertising campaigns. The authors reported that marketing- based activities within the firm are usually executed in bursts rather than in a steady manner and, for that reason may have unintended spill over effects on the
18 firm revenue. As the revenue of the firm overarches all the departments within the firm, it may negatively affect the employee loyalty to the firm because they fail to grasp the relation between the effort that they exert and the overall firm performance (Fischer et al., 2014). On the other hand, variability may be even advantageous given that it improves sales. Raju’s (1992) research findings indicate that by engaging in promotional activities more often, firms may enjoy less volatility in sales and, as a consequence meet operations as well as marketing objectives. The reverse outcome described by Raju (1992) occurs due to the competitive scenario, where firm performance is relative to other firms’ performance as opposed to firm- level performance.
By taking the detailed discussion of volatility into account, it is inferred that volatility induces uncertainty and therefore it is usually perceived as a negative occurrence. Since consumers tend to trust each other more than marketers or salespeople (Sen & Lerman, 2007), community consensus or lack of it, is expected to play a prominent role in determining the relationship between peer review valence and consumer loyalty. Specifically, the lack of community consensus is expected to have a negative moderating effect on review valence such that:
H3: The relationship between peer review valence and consumer loyalty is moderated by the
variance in community consensus, such that high discrepancies between consumer reviews
19
III. Methodology
3.1 Procedure & measures
As it has been already hinted in the introduction, this research was carried out in the movie industry. Most of the data for this research was retrieved from http://www.metacritic.com. “Metacritic” is an entertainment platform summarizing critics’ consensus with a single
unbiased rating- “Metascore”. Originally, the platform was set- up in 1999 as an experimental project by two attorneys, but over time it has gained momentum and evolved into informative and credible source of critic as well as user reviews. “Metacritic” covers not only the expert reviews in the movie industry but it also provides accurate user ratings in the games, TV and music segments following robust reports of the latest happenings in those industries.
Then, worldwide motion picture sales were collected from
http://www.boxofficemojo.com. “Box Office Mojo” provides leading online box- office reporting service. Additionally, it is owned by the International Movie Database (IMDB), which is a leading motion picture website in the world responsible for providing analytical movie reviews and scores, celebrity and TV content as well as reporting the most recent happenings in the movie industry. Further, IMDB has approximately 250 million individual users per month and contains more than 3 million movies and TV programmes. Therefore, consumers rely on the great array of movie data provided by IMDB.
In this research, over 5200 data points were initially retrieved from “Metacritic” and Box Office Mojo, and consequently, in the regression analysis only 158 cases were selected corresponding to series matching all the variables utilized in this research. Then, in order to test the model and the underlying hypotheses, Multiple Regression Analysis was applied in
20 the SPSS software. First, the data in the excel file corresponding to pre sequels and sequels in the series was arranged chronologically. Then, it was imported into the SPSS software and the corresponding Linear Regressions were executed accordingly.
3.2 Dependent and independent variable
First, the combination of three to five letters and two numbers is used to code and identify the sequels from the entire data sat. For example, “Harry Potter and The Philosopher's Stone (2001)” is coded as HPT01 and “Harry Potter and The Chamber of Secrets (2002)” as HPT02 etc. Then, motion pictures belonging to series were filtered in chronological order based on the release date in order to ensure that sequels precede the pre- sequels. For instance, in the latter case it was ensured that HPT02 comes after HPT01 and not the other way around.
Then, in order to measure the independent variables- “User Review Valence”, “User Score”, “Critic Review Valence” and “Critic Score”- respective average rating scores for each movie in the series were obtained from “Box Office Mojo”, and the score for pre-sequel was used to predict the loyalty of the sequel. For example, if the “User Score” for HPT01 was 7.4 then this score was used as a predictor of “User Review Valance” for HPT02
etc. Average scores were used due to the fact that each movie has more than one review. Peer review scores were rated on the scale of 1 to 10, whereas the critic reviews were rated on the scale of 0 to 100. However, the scale has no impact on the regression analysis as both variables are continuous and their unstandardized beta (β) coefficients are interpreted
21 The dependent variable- “Consumer Loyalty”- was calculated for each set of movies in the series based on the difference in adjusted worldwide sales of the sequel and pre-sequel. For example, “Consumer Loyalty” for “HPT01” and “HPT02” was calculated as the
difference in adjusted world wide sales between “HPT02” and “HPT01”. Similarly,
“Consumer Loyalty” for “HPT02” and “HPT03” was computed as the difference in adjusted world wide sales between “HPT03” and “HPT02” etc. The world wide sales were adjusted for inflation because 1 euro today has a different purchasing power as compared to 1 euro 10 years ago.
Further, independent as well as outcome variables are summarized in Table 1 below.
Table 1.
Variable Summary
Variable Name Description
Loyalty The difference between previous and current adjusted world- wide sales in dollars. Continuous variable.
Critic_Score Average critic review scores measured on the scale of 0 to 100.Continuous variable.
User_Score Average user review scores measured on the scale of 0 to 10.Continuous variable.
Critic_Valence Average critic review scores of the pre-sequels used to predict the loyalty of the sequel motion pictures in the series.Continuous variable.
User_Valence Average user review scores of the pre-sequels used to predict the loyalty of the sequel motion pictures in the series.Continuous variable.
22
Community_Size Number of online user reviews. Continuous variable.
Variance_Community Variance in the average user review scores. Continuous variable.
Action Action movie genre. Dummy variable, where “1” indicates action genre and “0” other genre.
Comedy Comedy movie genre. Dummy variable, where “1”indicates comedy genre and “0” other genre.
Drama Drama movie genre. Dummy variable, where “1” indicates drama genre and “0” other genre.
Thriller Thriller movie genre. Dummy variable, where “1” indicates thriller genre and “0” other genre.
3.3 Moderator variables
“Peer Community Size” was measured by the number of user reviews for each individual film. Further, average user scores were used to calculate variance in community consensus by means of “Excel Pivot Table”.
In order to test for moderation, first the mean was calculated for each moderator and the respective dependent variable. Namely: “Community_ Size”, “Variance_ in_ Community Consensus” and “User_ Review_ Valence”. Then, three new variables were computed by subtracting respective means from each variable individually. For example, for “Community_ Size”, the mean was found by means of descriptive statistics, and consequently subtracted from the variable “Community_ Size”, which resulted in a new variable-
“CM_Community_Size”. Further, the centred mean variable for “Community_Size” and “Variance_in_Community_Consensus” was multiplied by the “CM_ User_Review_Valence” variable, in order to compute the moderator variables for regression analysis. In the latter
23 example, “CM_Community_Size” was multiplied by “CM_User_Review_Valence” resulting in a new variable- “Interaction_Community_Size”.
There are no interaction or mediation effects expected regarding the critic reviews because experts tend to be more objective than consumers and, thus expert reviews are usually regarded by the scientific community as independent from other variables.
3.4 Control variables
As it has been hinted in the previous sections, there are numerous variables that may affect consumption decisions. For example, drama fans are more likely to see a drama movie than any other type of film. For this reason, movie motion genre was chosen as a control variable with 4 respective genres: thriller, drama, action and comedy.
In order to qualify for control variable, a variable must have an effect on the
dependent variable and not vice- versa. Thus, in this research consumer loyalty does not have effect on the genre but the opposite is true; genre has an effect on consumer loyalty as
24
IV. Results
4.1 Descriptive statistics and correlations
Relevant descriptive statistics as well as the corresponding Pearson correlation coefficients are summarised below in Table 2.
Table 2.
Descriptive statistics and correlations matrix
Mean SD 1 2 3 4 5 6 7 8 9 10 1 Loyalty -39935904 257800328 2 Critic_Score 57.56 14.81 0.017 3 User_Score 6.24 1.48 0.174* 0.14 4 Critic_Valen ce 57.01 16 0.024 0.328 0.062 5 User_Valenc e 6.53 1.25 0.135 0.33** 0.405** 0.07 6 Community_ Size 60.13 95.27 0.070 0.108 0.195 0.128 0.15* 7 Variance_Us er_Communi ty 7.32 3.58 -0.118 -0.10 -0.566** -0.134 -0.19 -0.03 8 Action 0.29 0.45 0.18** 0.025 0.164 -0.017 0.16* 0.078 -0.09
25
9 Comedy 0.29 0.46 -0.029 0.0045 -0.17** 0.058 -0.104 -0.181** 0.187** -0.18**
10 Drama 0.035 0.185 -0.042 -0.004 -0.001 -0.046 -0.008 -0.004 0.165** -0.123** -0.003
11 Thriller 0.050 0.208 -0.22** -0.016 0.049 -0.001 0.09 -0.015 -0.076 -0.006 -0.14** -0.042
*Correlation significant at the 0.05 level (2-tailed).
**Correlation significant at the 0.01 level (2-tailed).
Table 2 indicates that on average, both user scores and user valence exceeded critic review scores and valence. Customer loyalty, however was found to be negative on average indicating a deterioration in the movie box office revenue across series.
The correlation coefficients may be referred to in order to check the preliminary validity of the hypotheses (Situmeang et al., 2014). The Pearson correlation coefficient indicates the direction e.g. positive or negative and strength e.g. strong, medium or weak, of the relationship between variables (Field, 2013). Both user valence (r=0.135) and critic valence (r=0.024) were found to be weakly positively correlated with loyalty. Moreover, this test it was revealed that user review valence and the community size are weakly positively correlated (r=0.15, p<0.05). Similarly, variance in the user reviews was found to be weakly negatively correlated with user review valence (r=0.19).
4.2 Regression
In order to test the hypotheses, and determine the effect that the independent variables as well as the control variables have on consumer loyalty, multiple regression analysis was run. As it has been already discussed in the theoretical framework, hypothesis 1 a) predicted that critic review valence has a positive effect on consumer loyalty, hypothesis 1 b) conjectured that
26 user reviews have a positive effect on consumer loyalty and hypotheses 2 and 3 presumed that user community has a positive moderating effect on the relationship between user review valence and consumer loyalty, and peer review volatility has a negative effect mediating effect on the latter relationship.
In order to test the four hypotheses, two different models were constructed. The first model consisted of the following predictors: critic score, user score, critic valence, user valence, user review volume, user review variance, drama, action, comedy and thriller. This model was statistically significant F= (157, 2.913); p- value < 0.05 and explained 16.5% of the variance in consumer loyalty. Significance level was below 0.05 (p=.002), meaning that the probability of the null hypothesis is small enough to be confirmed.
In the second model the following predictors were entered: critic score, user score, critic valence, user valence, user review volume, user review variance, drama, action, comedy, thriller, interaction volume and interaction valence. The second model was also found to be significant with F= (157, 2.559); p-value <0.05 and explained 17.5% of the variation in consumer loyalty. However, the improvement of 10% in R2 does not necessarily mean that the second model added more explanatory power. On the contrary, the adjusted R Square plummeted from 0.109 to 0.106 and the F Change fell from 2.913 to 0.824 meaning that the interaction variables did not result in the improvement of the first model. This in turn suggests that the interaction effects were ultimately negligent.
Table 3.
Model Summary
Model R Square Adjusted R Square F Change df1 df2
1 0.165 0.109 2.913 10 147
27 Moreover, as revealed in Table 5 below, the interaction variables were found to be insignificant: moderation of variance in user reviews (p > 0.1) and interaction of volume (p >0.1). This confirms the previous analysis of the two models as well as the results in the correlation matrix indicating that there is no moderation effect. Furthermore, only user review valence (p<0.1), review volume (p<0.1) and action movie genre (p< 0.1) resulted in significant results indicating that there is not enough statistical evidence to support H1 a), H2 and H3, thus they are rejected and only H1 b) is accepted. Nonetheless, community size has a direct effect (p <0.1) on loyalty. Surprisingly, community size has the strongest effect on loyalty (Standardized Beta= 0.249). In particular, if the community size increases by one user review, on average, consumer loyalty increases by $531,560.04. As for user review valence, an incremental increase of one unit in the average “Metascore” of a pre- sequel leads to the additional $25,392852.1 in the box office revenue of a sequel. Similarly, the results indicate that the viewers of action movies are loyal to this genre and, as a result improve the revenue for action genre sequels by $53,198973.61 as opposed to comedy, drama and thriller genre, which have no significant effect on loyalty.
Multicollinearity usually occurs when multiple independent variables have high correlations between each other (Field, 2013). That is one predictor variable may not only have an effect on the dependent variable but also on other predictor variables (Field, 2013). This in turn may affect the calculations regarding individual independent variables. For example, it may render coefficient testing problematic and make the individual contribution of independent variables cumbersome to disentangle (Field, 2013). In order to detect the presence of multicollinearity, Variance Inflation Factor (VIF) is usually used. The rule of thumb states that if VIF is below 5 then multicollinearity is absent (Field, 2013). We can see
28 in Table 5 that VIF for every variable is less than 5 indicating that there is no
multicollinearity between the predictor variables.
Table 4.
Linear Regression Model without Interaction variables.
Model 1 Unstandardized B Standardized Coefficients (Beta) Sig. VIF (Constant) -84382582.65 0.508 Critic_Score -1542268.50 -0.115 0.175 1.263 User_Score 6175776.21 0.049 0.618 1.685 Critic_Valence -563229.77 -0.049 0.527 1.051 User_Valence 24373591.81 0.161 0.062 1.295 Community_Size 550062.72 0.257 0.002 1.159 Variance_User_Community -8409286.54 -0.163 0.086 1.566 Action 53198973.61 0.137 0.083 1.085 Comedy 10557287.31 0.023 0.772 1.151 Drama 87708785.18 0.083 0.297 1.110 Thriller -78941069.42 -0.088 0.252 1.028
29 Table 5.
Linear Regression Model with Interaction variables.
Dependent variable: Loyalty
Table 4 on page 28 represents the initial multiple regression model that consists only of predictor as well as control variables, whereas Table 5 above extends the Table 4 by including the moderator variables.
Model 2 Unstandardized B Standardized Coefficients (Beta) Sig. VIF (Constant) -109468641.1 0.398 Critic_Score -1448559.9 -0.108 0.209 1.295 User_Score 5350446.04 0.042 0.667 1.691 Critic_Valence -549012.8 -0.048 0.538 1.051 User_Valence 25392852.1 0.168 0.055 1.316 Community_Size 531560.04 0.249 0.005 1.306 Variance_User_Community -7047073.15 -0.137 0.162 1.657 Action 61072506.4 0.157 0.052 1.136 Comedy 14495197.84 0.032 0.693 1.163 Drama 99702414.05 0.094 0.240 1.126 Thriller -89630314.83 -0.1 0.198 1.044 Interaction_Variance -3519196.88 -0.1 0.233 1.224 Interaction_Volume 61744.49 0.032 0.698 1.185
30
V. Analysis
This part discusses the findings obtained in the previous section. The confirmed as well as rejected hypotheses are robustly analysed by means of referring to previous scientific literature in order to contribute to future research. Then, practical as well as theoretical implications are provided. Finally, this section is concluded by discussing limitations as well as suggestions for future research.
5.1 Discussion of findings
This research delved into the effect that eWOM- user and critic online review valence- has on customer loyalty in the motion picture industry. In particular, it was tested whether online review valence is positively correlated with consumer loyalty and whether the relationship between peer review valence and loyalty is moderated by the user community size and variance in consumer reviews.
Hypothesis 1 a) predicted that online expert review valence has a positive effect on consumer loyalty. This hypothesis was rejected as there was not enough statistical evidence (p>0.1) to infer that the relation between expert review valence and consumer loyalty exists. The rejection of H1 a) could have occurred due to the fact that critics did not assume the role of influencers but the one of predictors (Basuroy et al., 2003; Eliashberg and Shugan,1997). This in turn could have been caused by the fact that critics were not able to influence the moviegoers’ tastes and preferences, as they did not exert enough influence on their respective audiences, and consequently assumed the role of predictors, where they were only successful at predicting the movie box office (Weiman, 1991).On the other hand, the absence of relation between critic review valence and consumer loyalty can also be explained by the fact that
31 critic reviews were not published at the early days of the movie release date and consequently did not meet consumers’ expectations making the effect of the expert reviews on the movie box office negligent (Basuroy et al., 2003; Eliashberg & Shugan, 1997; Liu, 2006).
Equivalently, since the effectiveness of critic reviews tends to recede over time (Eliashberg & Shugan, 1997), it can be inferred that the expert reviews posted by “Metacritic” were
outdated or even inaccurate. However, the latter explanation is not feasible given the impeccable credibility of http://www.metacritic.com. Furthermore, the lack of statistical evidence for accepting H1 a) can also be explained by the fact that movie studios
corresponding to the motion pictures sample used in this research engaged in “cold opening” (Brown et al. 2012). This explanation is viable if the spectators are rational and hence they trust the movie studios more than the film experts (Brown et al. 2012). In addition, cold opened movies were found to be positively linked to box office revenues, meaning that movie studios may profit from such practices (Lovallo, 2013). Finally, the rejection of H1 a)
supports Brown, Camerer and Lovallo’s conclusion, which states that nowadays critic reviews have been loosing significance because of being easily influenced by the movie studios’ self - interest (2013). However, the reasons for rejecting H1 a) as discussed above, are not normative but rather descriptive, meaning that more research needs to be conducted in order to fully grasp the dynamics behind this relationship.
The second hypothesis- H1 b)- predicted that online user review valence has a positive effect on consumer loyalty. This hypothesis was accepted (p<0.1), indicating that there is enough statistical evidence to infer that user review valence has a positive effect on consumer loyalty. This finding may be explained by multiple determinants. First, the confirmation of H1 b) is in line with BrightLocal (2013) research agency’s findings, which indicate that almost 80% of consumers trust online peer reviews as much as they trust each other. Second, in line with Kelsey group and ComScore’s study results, this paper confirms
32 that consumers who take eWOM into account when making a purchase decision tend to trust online peer reviews more as compared to critic reviews. Furthermore, since consumer loyalty was measured as the difference in the adjusted movie box office between sequels, it can be inferred that consumer review valence has a positive effect on motion picture sales, which is indeed in accordance with Dellarocas (2007) research outcome.
Hypothesis 2 predicted that the relation between user review valence and consumer loyalty is moderated by the user community size, such that larger communities lead to a stronger relationship between peer review valence and loyalty. This hypothesis was rejected (p>0.1) indicating that there is not enough statistical evidence to confirm the moderation effect. Nonetheless, direct effect was confirmed (p<0.1) between online peer community size and customer loyalty, such that larger user communities lead to improved user loyalty. In particular, one additional online review by consumers for a pre- sequel ameliorates sales of the sequel by $531,560. Moreover, the relationship between user community and loyalty is the most significant one (p=0.005) in the model, meaning that it is most likely to hold. This finding is in accordance with signalling theory, meaning that experienced moviegoers send a positive signal about the sequel in the series by means of eWOM in the online- user
community, which in turn encourages moviegoers to go and see the sequel film in question. Therefore, the larger the online user community, the stronger the signal is. In addition, Liu (2006) findings were also confirmed, indicating that volume of user reviews is more significant than critic reviews when determining the motion picture box office. Moreover, Duan, Gu & Whinston’s (2008) findings of a positive relation between review volume and motion picture sales were confirmed.
Hypothesis 3 predicted that the relation between online user review valence and consumer loyalty is moderated by volatility in consumer reviews, such that high volatility reduces the effect that peer reviews have on consumer loyalty. This hypothesis was rejected
33 (p> 0.1) indicating that there is not enough statistical evidence to conclude that volatility moderates the relationship between consumer review valence and loyalty, thus not enough evidence to support the moderation effect. Moreover, there was not enough statistical evidence (p>0.1) to support the direct effect that variance in consumer reviews has on consumer loyalty. At first, it was surprising that volatility in user reviews was found to have no effect given the significance that it plays in a great array of settings. The absence of volatility in user reviews means that there are no discrepancies between online review
valence in the user community e.g. community consensus prevails. According to Sun (2012), high volatility in online user reviews implies mediocre quality of a product in question and thus negatively affects sales of the product. As for this research, the latter finding is in tandem with the previous results discussed in this section meaning that the lack of significance in variance of user reviews is perfectly reasonable.
As for control variables, only the action genre was confirmed to have a significant effect (p<0.1) on consumer loyalty. Liu (2006) revealed that movie genre has a distinct link with the patterns of eWOM behaviour. In particular, certain genres tend to have a more pronounced effect on the movie box office (Liu, 2006). For example, Lord of the Rings action movie series is one of the most successful series in the sample used for this research, which grossed almost $ 3 billion worldwide. Moreover, the trilogy received unprecedented ratings on “Metacritic” (92, 87 and 94 respectively). Therefore, the latter finding of the action genre significance on consumer loyalty is in line with the actual trends in the movie industry. As for other genres- thriller, comedy and drama- they seem to be less popular within the moviegoers, hence insignificant results in the regression analysis.
34 5.2 Theoretical implications
The aim of this research was to shed some light on the relation between eWOM and consumer loyalty as this link has not been thoroughly established in the previous scientific literature and to capture the effect that community dynamics may have on the individual decision making.
Despite the findings of moderating effects being inconclusive, this research has established that community size has the most significant effect on consumer loyalty from all the variables utilized in this paper. Consequently, it confirms the signalling theory meaning that individual decision making is highly dependent on the peer feedback. Additionally, signalling theory was also confirmed by the fact that both user scores and user review valence were found to be significantly positively correlated with loyalty, whereas critic reviews were found to be insignificant. Thus, this research has successfully contributed to narrowing the research gap regarding the literature about eWOM and consumer loyalty.
5.3 Practical implications
This research may be used by movie studios as a reference point when contemplating the significance of critic reviews on the potential success of a movie in question. Since both critic review scores and critic review valance were found to have no effect on consumer loyalty, it may be inferred that by engaging in “cold opening”, the motion picture studios may succeed in maximising their revenue (Brown et al., 2012).
35 5.4 Limitations and future research
Both theoretical and practical implications should be considered in line with limitations surrounding this research.
First, this research could have used a larger sample size than 158 movies in the series as the majority of variables had more data points, which had to be disregarded due to the fact that variables such as critic and user reviews had only 158 data points. This in turn leads to less significant power of the study, as the higher the sample size, the more significant the model is (Field, 2013). However, in order to ensure that data is approximately normally distributed, bootstrapping was incorporated into multiple regression analysis. In order to add significance to this study, a more elaborate study with more data cases should be used in the future research.
Second, only quantitative nature of the online critic and user reviews was
incorporated into this research overlooking the qualitative nature of reviews. This may not be necessarily an issue for critic reviews, as they tend to be more objective in nature as opposed to user reviews, which tend to be more personal and subjective (Reinstein & Snyder, 2005). However, since critic review scores and valence were found to be insignificant, the
qualitative aspect of reviews would certainly add more explanatory power to user review valence and help explain the relation between loyalty and online reviews better. For example, by means of Liker- type scale, qualitative consumer reviews could be quantified e.g. rated on the scale of 1 to 7 (Field, 2013). This in turn would allow the findings to be generalized to a wider population, thus improving generalizability of the research.
36 Third, the validity of the study could have been altered by the fact that online
consumer reviews are anonymous in nature, thus biased reviews may have been provided. In order to to overcome this problem, data on consumer reviews could be gathered from private research firms that specialize in data collection and analysis. Alternatively, only data from credible users who have attained credibility over time by contributing to the online user community could be collected in order to overcome the problem of validity.
5.5 Conclusion
The main goal of this study was to contribute to the existing eWOM literature by providing fresh insights into the relationship between critic as well as consumer reviews and consumer loyalty. The results of this research supported prior findings obtained in the movie industry confirming the ever diminishing role of critic reviews in influencing the motion picture box office and the significance of online peer reviews as well as the community size. However, even though a weak correlation was identified between loyalty, user community size and variance, no support was found for moderating effects.
In the light of this research, the scientific community as well as the movie studios could derive some basis for further analysis. Since loyalty is of paramount importance to a great array of businesses, this type of research could equally be conducted in wider settings in order to see if the findings are generalizable.
37
Bibliography
Alreck, P. L., & Settle, R. B. (1999). Strategies for building consumer brand preference.
Journal of Product & Brand Management, 8(2), 130-144.
Basuroy, S., Chatterjee, S. and Ravid, S.A. (2003), “How critical are critical reviews? The box office effects of film critics, star power, and budgets”, Journal of Marketing, Vol. 67 No. 4, pp. 103-117.
Bellman, S., Johnson, E. J., Lohse, G. L., & Mandel, N. (2006). Designing marketplaces of the artificial with consumers in mind: four approaches to understanding consumer behavior in electronic environments. Journal of Interactive Marketing, 20(1), 21-33.
BI Situmeang, F., AAM Leenders, M., & M. Wijnberg, N. (2014). The good, the bad and the variable: How evaluations of past editions influence the success of sequels. European
Journal of Marketing, 48(7/8), 1466-1486.
Bickart, & Schindler. (2001). Internet forums as influential sources of consumer information.
Journal of Interactive Marketing,15(3), 31-40
Brown, A. L., Camerer, C. F., & Lovallo, D. (2013). Estimating structural models of
equilibrium and cognitive hierarchy thinking in the field: The case of withheld movie critic reviews. Management Science, 59(3), 733-747.
38 Brown AL, Camerer CF, Lovallo D (2012). The journal of industrial economics, 53(1),
27-51.
Brown AL, Camerer CF, Lovallo D (2009) To review or not to review? Limited strategic thinking at the movie box office. Working paper, Texas A&M University, College Station.
Caves, R. E. (2000). Creative industries: Contracts between art and commerce (No. 20). Harvard University Press.
Chatterjee, A., & Hambrick, D. (2011). Executive Personality, Capability Cues, and Risk Taking. Administrative Science Quarterly, 56(2), 202-237.
Connelly, B. L., Certo, S. T., Ireland, R. D., & Reutzel, C. R. (2011). Signaling theory: A review and assessment. Journal of management, 37(1), 39-67.
Cornelius, P., A. van de Putte and M. Romani (2005), Three decades of scenario planning in Shell, California Management Review, 48(1), pp. 92-109.
Dellarocas, C., Zhang, X. M., & Awad, N. F. (2007). Exploring the value of online product reviews in forecasting sales: The case of motion pictures. Journal of Interactive
marketing, 21(4), 23-45.
Dens, N., De Pelsmacker, P., & Purnawirawan, N. (2015). “We(b)care”. Journal of Service
39 Duan, Gu, & Whinston. (2008). Do online reviews matter? — An empirical investigation of
panel data. Decision Support Systems, 45(4), 1007-1016.
Eisenhardt, K.M. and D.N. Sull (2001), Strategy as simple rules, Harvard Business Review, (January-February), pp. 107-116.
Eisenhardt, K., Zbaracki, M., & Schendel, Dan. (1992). Strategic decision making. Strategic
Management Journal, 13(S2), 17-37.
Elberse, A. and Eliashberg, J. (2003), “Demand and supply dynamics for sequentially released products in international markets: the case of motion pictures”, Marketing
Science, Vol. 22 No. 3, pp. 329-354.
Eliashberg, J., & Shugan, S. M. (1997). Film critics: Influencers or predictors?. The Journal
of Marketing, 68-78.
Field, A. (2013). Discovering Statistics Using IBM SPSS Statistics. Sage Publications, London, 4th edition, ISBN 9781446249185.
Fischer, M., Shin, H. S., & Hanssens, D. M. (2015). Brand performance volatility from marketing spending. Management Science, 62(1), 197-215.
40 Fiske, S., & Manis, Melvin. (1980). Attention and weight in person perception: The impact of
negative and extreme behavior. Journal of Personality and Social Psychology, 38(6), 889-906.
Gemser, G., Van Oostrum, M., & Leenders, M. A. (2007). The impact of film reviews on the box office performance of art house versus mainstream motion pictures. Journal of
Cultural Economics, 31(1), 43-63.
Ho, J.Y.C., Dhar, T. and Weinberg, C.B. (2009), “Playoff payoff: super bowl advertising for movies”, International Journal of Research in Marketing, Vol. 26 No. 3, pp. 168-179.
Holbrook, M. (1999). Higher Than the Bottom Line: Reflections on Some Recent Macromarketing Literature. Journal of Macromarketing, 19(1), 48-74.
O’Connor, P. (2008). User-generated content and travel: A case study on Tripadvisor. com.
Information and communication technologies in tourism 2008, 47-58.
Jensen, & Meckling. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305-360.
Journal, W. S. (2011, February 26). Best Movies Ever. Retrieved from
41 Karniouchina, E.V. (2011), “Impact of star and movie buzz on motion picture distribution
and box office revenue”, International Journal of Research in Marketing, Vol. 28 No. 1, pp. 62-74.
Kahneman,D.,Tversky, A. (1981). The framing of decisions and the psychology of choice Science, 211 , pp. 453-458.
Kose, M., Prasad, E., & Terrones, M. (2003). Financial integration and macroeconomic volatility. Imf Staff Papers, 50, 119-142.
Lampel, & Shamsie. (2000). Critical push: Strategies for creating momentum in the motion picture industry. Journal of Management, 26(2), 233-257.
Lee, M., Rodgers, S., & Kim, M. (2009). Effects of Valence and Extremity of eWOM on Attitude toward the Brand and Website. Journal of Current Issues & Research in
Advertising,31(2), 1-11.
Levy, J. (1992). AN INTRODUCTION TO PROSPECT-THEORY. Political Psychology,
13(2), 171-186.
Liu, Y. (2006), “Word of mouth for movies: its dynamics and impact on box office revenue”,
Journal of Marketing, Vol. 7 No. 3, pp. 74-89.
Loayza, N., Rancière, R., Servén, L., & Ventura, J. (2007). Macroeconomic Volatility and Welfare in Developing Countries. September 2007.
42 Miller,& Robert (2006), "Taming the Volatile Sales Cycle," Sloan Management Review, 41
(2), 10-13.
Raju, J. (1992). The Effect of Price Promotions on Variability in Product Category Sales.
Marketing Science, 11(3), 207-220.
Razin, A., & Rose, Andrew Kenan. (1994). Business-cycle volatility and openness an
exploratory cross-sectional analysis. In Capital mobility : The impact on consumption,
investment and growth ; [proceedings of a CEPR conference with the Bank of Israel
and the Pinhas Sapir Center for Development, Tel Aviv University] (pp. 48-80).
Reinstein, D. A., & Snyder, C. M. (2005). The influence of expert reviews on consumer demand for experience goods: A case study of movie critics. The journal of industrial
economics, 53(1), 27-51.
Selnes, F., (1993), “An examination of the effect of product performance on brand reputation, satisfaction and loyalty”, European Journal of Marketing, Vol. 27 No. 9, pp. 19-35.
Sen, S., & Lerman, D. (2007). Why are you telling me this? An examination into negative consumer reviews on the Web.
Settle, R., Dillon, T., & Alreck, P. (1999). Acceptance of the phone-based interface for automated call direction. Behaviour & Information Technology, 18(2), 97-107.
