The high-wire act of being creative in the movie industry : Exploring the influence of successful product formulas on creativity and perceived quality

(1)

Bachelor Thesis Yannick Roosendaal 10439242 Supervisor: F.B. Situmeang 24-6-2016

The high-wire act of being creative in the movie

industry

Exploring the influence of successful product formulas on creativity and

perceived quality

(2)

1

Statement of Originality

This document is written by student Yannick Roosendaal who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is 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, not for the contents.

(3)

2

Abstract

In this day and age, new technologies play a big role in catering to consumer’s preference. The biggest players in creative industries make use of big data analytics to analyse and precisely figure out what content scores best. This development sparks a discussion. Does it lead to the increasing homogeneity of content, seen on in the media, and does this ultimately mean that creativity is diminishing and might be dead in the near future? To find out if consumers still appreciate creativity, this study questions how creative input in a successful formula influences online review ratings. Research is done to find out if and how an input of creativity, a change of genre, in a movie series mediates the effect that review ratings have on the review ratings and box office revenue of their respective sequel. In doing so proposing five hypotheses. The first and second proposing that previous review ratings positively influence sequel review ratings and box office revenue. The third proposing that review ratings show the probability that a sequel will have a different genre. The fourth that a change of genre positively influences review ratings and box office revenues in the sequel. The fifth and final proposing that a change in genre is less well received for successful series compared to unsuccessful series. Regression analyses prove the first two hypotheses p>0.000. The third proves that review ratings can predict whether a sequel will differ in genre. Interesting results on the relationship between review ratings, creativity and box office call for further research on the subject.

(4)

3

1. INTRODUCTION

Netflix’s 2016 first quarter results surpassed all expectations. Their big hit series, such as House of cards, are to a large extent responsible for these successful figures. An important factor to Netflix’s booming success is that Netflix uses big data analytics to predict how likely it is that people are going to enjoy a certain movie. Netflix has years of viewer-data including online reviews of their customers, which they use to develop predictive algorithms. It is clear that online reviews have a big effect on organizational decision-making and as a result on the content that consumers receive.

Netflix is certainly not the only data-driven company. More and more companies base their strategy and other important decisions on big data. Whilst in a lot of industries, this has solely positive effects, in the creative industry, these developments could have some worrying implications. Consumer demand is highly uncertain in the creative industry, which is why producers tend to reuse ideas and concepts that have been known to be successful in the past (Tschang & Szczypula, 2006). In the music industry for instance, the constant quest for big hit records resulted in a great increase in the amount of homogeneity of hit songs. Research on the past five decades of pop music has shown that these songs are almost solely based on the ten most popular chords (Serrá, et al 2012). Considering that more and more data is being stored and analysed to see what makes a product a financial success, it is not farfetched to conclude that in the near future, content will rapidly and increasingly be the same rehashed, homogenous formula. However, financial success is not always in the best interest of the audience. Research has shown that whilst revenues tend to go up when releasing sequels, expert and consumer ratings actually tend to go down (White, 2009). When also considering the fact that audience rating remains to be one of the most understudied variables in movie research, the necessity of further research becomes clear (Chang & Ki, 2005).

Most people will agree that rising revenues and dwindling expert and consumer ratings are not desirable developments. This could very well have a negative influence on creativity. As companies will always try to reduce risk, whilst maximizing profits. As a result, it is more likely that a proven-to-be successful formula will be rehashed, instead of coming up with something wildly creative and risky. Besides, seeing as data gathered online can drive organizational decisions, as a result, if box office data keeps pointing in the same direction, leading organizations to the same, proven to be successful formula, in the future it is expected to see even more homogenous, less creative, mass marketed content.

(6)

5

Although big data can lead to a certain direction, it does not necessarily have to kill creativity, it can also function as mere guidance. In the end, commercial companies will want to cater to consumer preferences. As long as consumers value new creative input in an

existing formula, companies will try to incorporate this in their product.

In current literature, the effect and drivers of previous success and input of creativity in particular, on review ratings has not yet been thoroughly researched, however this is important to know, for society, to see if the fear for creative bankruptcy is justified, and for organizations, to have a better understanding of consumer preferences. In addition, it can an important way to get ideas on how to be competitive in the contemporary analytics driven environment.

Thus, the goal of the paper is to find out if people really do appreciate and highly rate the same formula consistently. Or do they prefer new creative input in an otherwise

homogenous success formula, and is previous success actually a driver of creativity? To find answers, the associated variables need to be researched. First of all, the effect of previous success on online review ratings and box office revenue needs to be researched. Furthermore, the mediating effect of creative input on online review ratings and box office revenue is to be researched. Therefore, the research question in this paper is: how does creative input in a successful formula influence online review ratings and box office performance?

(7)

6

2. LITERATURE REVIEW

In this section, an in-depth review of current literature regarding successful formulas and their influence on creativity and consumer/critic appreciation will be presented. This framework will provide a representation of relevant concepts and their relationships will be explored. The literature reviewed will serve as a foundation on which the conceptual model in the next section is established.

2.1 Successful formulas

Research has shown that because of the uncertainty of consumer demand in the creative industry, producers tend to reuse successful formulas, and due to this fact, the strategy of releasing sequels has become more popular (Rosen, 2011). This is not the only article that stresses that the creative industry relies more and more on successful formulas, which leads to more homogenous product releases. A recent article on western pop music shows that 10% of the songs generate 90% of the revenue (Serra, 2012). When we look at other research data we see that sequel movies in a successful formula tend to have higher sales (Moon Bergey & Iacobucci). This provides an explanation for the phenomenon of increased homogeneity. Rehashing the same successful formula is a fairly riskless venture, as revenues are almost guaranteed to increase, even with decreasing review ratings (Moon, Bergey & Iacobucci, 2010). Guaranteed financial success is of course an attractive incentive for producers.

From the consumer’s point of view however, at first, this fact looks strange and counterintuitive. Why would sales increase, when the quality seems to decrease? There are a few important concepts that contribute to this phenomenon. First off, according to branding theories, a sequel is a form of a brand extension. A brand extension takes place when

producers use established brands to introduce new products. Brand extension evaluations are influenced by the perceived similarity to the parent brand. When perceived similarity is high, associations from the parent brand are transferred to the extension (Sood & Dreze, 2006). When making use of this tactic, producers can take advantage of the established parent brand, and attract customer attention more easily, whilst reducing marketing costs during the

introduction of the new product (Chang & Ki, 2005). Which uncovers another financial incentive for producers to rehash successful formulas.

However, not only producers are risk averse. According to signalling theory, when two sides have asymmetric information, the first party, the sender can choose how and when to communicate signals to the receiving party. In the movie industry, the producers would be the sending party, and the consumers the receiving party. The consumers then have to

(8)

7

interpret the information signalled, which can contain implicit information about the movies in order for them to reduce the information asymmetry and decide on what movie they would like to watch (Connelly, Certo, Ireland & Reutzel, 2011). Signalling is most advantageous for experience goods such as movies, where the quality is relatively unknown before the

purchase. Signals from the producers are proven to be positively related to movie

performance (Basuroy & Chatterjee, 2008). This explains why signals that have to do with for example brand extension, lure more consumers to the movies.

2.3 Creative input

When taking a look at creativity in other creative industries, such as the music industry and the gaming industry we see worrying developments. For instance, in the music industry, the large majority of hit songs are produced by just four producers, making it increasingly hard for competitors, and as a result crowding out creativity (Serra, 2012). White describes a similar phenomenon in the gaming industry (2009). An increasing lack of creativity occurs due to rising financial and logistic barriers, results in repeating series of slightly altered games, the successful football series Fifa 98-Fifa-2016 exemplifies this. White goes as far as stating that the shift away from risky new ideas can potentially be disastrous for the future of this industry.

Interestingly, White’s paper argues that a lot of sales can be generated when a product is based on a license, a successful formula, even though the review ratings are mediocre (White, 2009). This finding can have a significantly negative effect on creativity, as it diminishes producers’ incentive to spend money on risky new ideas, when there is practical proof that rehashing a successful product can generate a lot of sales, even though the reviews are not overwhelmingly positive.

An explanation for these not so overwhelmingly positive reviews can be the satiation effect. Satiation is discovered to be a mechanism that results in variety-seeking behaviour. Satiation results in the preference to experience something new in sequences of experiential goods (Ratner Kahn, & Kahneman, 1997). The fact that sequels with descriptive titles

received higher ratings compared to movies that were titled with their sequential number, can hint at a lack of creativity and an occurrence of the satiation effect (Sood & Drèze, 2006).

Consumers desire new experiences as they optimize their level of stimulation, variety seeking therefore shows the preference of dissimilar goods over similar goods when

(9)

8

movies, Sood and Dreze describe that although high similarity results in a close connection to a previous movie, it is likely that high similarity will result in satiation of the experience (2006). Subsequently, they describe that consumers can become satiated by certain attributes of experiential goods, when these attributes exceed a specific level. When this happens, consumers prefer a different product the next time. From these findings it is possible to deduct that creative input can significantly influence consumer reviews, although producers are risk averse and try to limit creativity.

2.2 Consumer and critic reviews

The concept of online reviews originates from word-of-mouth, one of the most important ways to exchange and transmit information. Exactly how important is defined by the reach and the impact of the message (Lang & Lawson, 2013). In contemporary society, the internet has removed the physical boundaries involved in word-of-mouth. Thus, greatly increasing the reach of word-of-mouth. Due to the fast speed and wide range of the internet, the availability of information about products and services has widely improved (Liu Y.

2006). Subsequently, the introduction of online review systems has made it possible for consumers to freely access and exchange information on products, services and so on (Duan, Gu & Whinston, 2008).

Reviews can be expressed in words, as sentiments and they can also be expressed as a numeric representation in the form of ratings. These ratings are found to have an indirect impact on sales through sentiments, Textual sentiments however, impact sales directly (Hu, Ko & Reddy, 2014). A research conducted in 2013, found that when an online review has a similar manner of articulating and communicative style compared to the reader, the review would be valued more and perceived to be more credible (Ludwig, de Ruyter, Friedman, Brüggen, Wetzels, & Pfann, 2013).

In addition, an article by Moon, Bergey & Iacobucci showed evidence that sequel movies usually have higher revenue but receive lower ratings than originals (2010). An explanation for these lower ratings might lie in the fact that the viewers are less satisfied because they are already used to the experience due to a lack of the sequels innovation and creativity, which results in lower ratings (Moon, Bergey & Iacobucci, 2010).

(10)

9

3. CONCEPTUAL MODEL

In the following section, a conceptual model will be conceptualized according to the literature reviewed and analysed in the aforementioned literature review. Important variables in this paper will be described and their interrelationships will be hypothesized. Starting with perceived quality, divided in critic and consumer ratings. Subsequently, box office performance, previous success and finally, what happens when a sequel changes genre.

3.1 Perceived quality

To find out whether consumers’ preference differs between creative sequels or rehashed sequels, online review ratings are used as the indicating variable of perceived quality. Review ratings can be seen as the quantified form of online reviews. In the case of ratings, the

evaluation of a certain product is expressed in a single rating. In a recent article Situmeang, Leenders & Wijberg state that the perceived quality of a product is a legitimate indicator of the success of a movie, and consumer and expert reviews can both signal the quality of a product (2014). In addition, Duan, Gu & Whinston state that average user ratings are to be considered a driving force of customers’ product choice (2008).

In many articles, box office revenue is used as the main performance measure, however, as research proved, box office can increase whilst perceived product quality and review ratings decrease (Moon, Bergey & Iacobucci, 2010). This is why it is important to not only look at box office performance, but to do research for both consumer and critic review ratings as well. Furthermore, because literature suggests that when a movie differs from the expectations set by the original, this translates in higher rated movies than those which are similar to the original, this makes the variable online review ratings interesting to look at. When review ratings are low, this seems to have few consequences on the short term. On the long term however, there is evidence that this leads to decreasing revenues in sequels in the long run (Sood & Drèze, 2006).

In this research, both consumer and critic ratings will be analysed. Critics function as both influencers and predictors. They influence consumers in their choice between movies, and their reviews can be interpreted as an estimation of what the consumers’ opinion on a movie is going to be (Basuroy, Chatterjee & Ravid, 2003)

In conclusion, online review ratings, both consumer and critic are directly or indirectly related to perceived product quality and box office performance and as such are used in this research to indicate product appreciation by consumers, and performance in the broad sense

(11)

10

of the word. As such, the influence of creativity on movie performance can be researched. The following is hypothesized:

H1a: Previous average ratings positively influence sequel average ratings

H1b: Previous consumer ratings positively influence sequel consumer ratings

H1c: Previous critic ratings positively influence sequel critic ratings

Figure 1 - Conceptual model of the proposed effect of prequel ratings and genre on sequel review ratings

Change of

genre

Sequel

ratings

(12)

11

3.2 Box office

An important factor to include in this study is the financial side of the story. From a

consumer’s point of view, one could state that when perceived quality of a series of movies is high, this makes or breaks a successful formula. From the producer’s point of view however, financial success is equally, if not more important than successful the movie being positively received in the form of review ratings, because profit must be made to keep things running.

Most studies use total domestic box office as the variable for financial success. Some studies use worldwide box office gross (Chang & Ki, 2005). However, for this paper, it is important that the concept financial success is measured in a fair way. When measuring worldwide gross, uncontrollable variables such as different foreign places of release and cultural differences could affect certain box office performances (Oh, 2001). In this article, the term box office refers to domestic total gross.

It is clear from previously mentioned literature that box office is related to review ratings, be it a direct effect, or an indirect effect (Hu, Ko & Reddy, 2014). Therefore, it is also presumable, that previous review ratings influence box office of the sequel movie. By also researching sequel box office, it can be verified whether there is a financial incentive for the producers to add creative input to a series. According, the following is hypothesized:

H2a: Previous average ratings positively influence sequel box office

H2b: Previous consumer ratings positively influence sequel box office

H2c: Previous critic ratings positively influence sequel box office

Figure 2 - Conceptual model of the proposed effect of previous ratings and change of genre on sequel box office

Change of genre Sequel box office Previous ratings

(13)

12

3.3 Change of genre

An important attribute such as the genre of a movie is often featured in promotion efforts of the producers, such as movie trailers. The parent brand associations in movie sequels which are most important are those made by experiential attributes. These experiential attributes, such as the genre and the story-line however, tend to satiate the consumers. According to Sood, and Dreze this results in consumers desiring a different experience in the sequel (2006). Because although high similarity makes for a closer connection to the prequel, satiation is likely to occur when it comes to experience goods like movie sequels. This satiation effect therefore is likely to result in lowered movie ratings in sequels. Sood & Dreze also argue that consumers might be more attracted to movies which include a new theme and genre instead of simply continuing the same theme where original movie has ended.

A similar interesting development in genre is described by Kücklich in the gaming industry (2005). He noted that although big players in the industry rely increasingly on licensing and sequels to ensure profitability, this is counterbalanced by a community of consumers that use modification tools to change the genre of the game and implement their own creativity. This development shows evidence to believe that although creativity in the creative industry suffers from the current dominant business model, consumers actually desire and positively welcome creative input and changes in genre in a successful formula. This leads to believe that when a sequel in a series of movies adds and/or changes the genre, something new happens, due to input of extra creativity which is positively welcomed by consumers. However, when review ratings are high to begin with, it is more likely that producers are reserved and restrictive in changing the genre. This can also work the other way, if previous review ratings were relatively low, producers probably have a bigger incentive to change the genre.

According to the aforementioned, the following is hypothesized:

H3a: Previous average review ratings predict whether a movie’s sequel will change in genre.

H3b: Previous critic review ratings predict whether a movie’s sequel will change in genre.

(14)

13

H4a: When a sequel changes in genre compared to the prequel, this positively influences the sequel’s review ratings

H4b: When a sequel changes in genre compared to the prequel, this positively influences the sequel’s box office

3.4 Successful versus unsuccessful series

The previous success of a certain movie-series can be split in various forms of success. Mentioned before is the perceived quality of a product, which is a legitimate an indicator of the success of a movie because consumer and expert reviews can both signal the quality of a product (Situmeang, Leenders & Wijnberg, 2014). Moreover, financial success is also an important measure for movie success ((Chang & Ki, 2005). Some studies also include Academy Awards nominations as a measure for movie success (Krauss, Nann, Simon, Gloor & Fischbach, 2008).

Current literature however, has not reached full consensus on whether the amount of success of a movie is an indicator of the online review ratings of the sequel. In 2014,

Situmeang, Leenders & Wijnberg found results in their study indicating that expert reviews of sequels are influenced by the expert ratings of the prequel, as well as average sales of the prequels. Furthemore, average consumer reviews on prequels positively influenced the sequel’s consumer reviews. Another study by Chang & Ki, showed that a movie being a sequel is a significant predictor for financial success. However, this was only the case when it came to movie sequels. When the movie was a sequel to other media, such as a book, the result was not significant (2005). In addition, another study argues that previous success can also build high expectations, which can lead to lower levels of satisfaction (Oliver, 2009).

Whilst there might not be full consensus on the issue, most research does point in one direction. When also taking the development into account that producers tend to reuse successful formulas more and more, it is not unthinkable to hypothesize that consumers appreciate and positively receive a sequel based on the same successful formula (Tschang & Szczypula, 2006). However, it is important to take into account that there might be a

significant difference in how these variables interact for successful versus unsuccessful movie series. For successful series, consumers might not appreciate creative input and change in the

(15)

14

series, and producers are likely less inclined to change the winning formula. For unsuccessful series however, producers are more likely to try and change the formula for the better, with new creative input and consumers are might be more appreciative of change. Hence it is hypothesized that:

H5a: When the sequel of a successful movie changes of genre, this negatively influence sequel’s review ratings and box office

H5b: When the sequel of an unsuccessful movie changes of genre, this negatively influences sequel’s review ratings and box office

(16)

15

4. METHODOLOGY

The following paragraph will be used to explain the research methods used to conduct the analyses. The specific scale of the variables that are being used are illustrated. Furthermore, an explanation of how the hypotheses are tested will be stated. The sample and the field are determined, and the procedure and predictions are stated, to enable reproduction of this research.

4.1 Research design

Research is to be conducted in multiple ways, starting with existing literature. Furthermore, the hypotheses will be tested by using quantitative research methods. The dataset to be researched contains data of approximately 6000 movies including sequels. The focus will lie on these series. A series of movies consists of at least two movies. In this dataset, the data that will be used are, online expert review data from Metacritic and online consumer review data from Metacritic, as well as the specified genre of these sequels and the domestic box office gross of these movies. The data is collected previously by people at the University of Amsterdam.

4.2 Measurements

Measurements will be done on distinguishable variables. The first of these variables is online review ratings. Which are split out in critic review ratings and consumer review ratings. These two ratings combined will form the average review rating. Online review ratings of sequel’s will form a dependent variable. Previous online review ratings will form an independent variable. Only review ratings on movies that are part of a series of at least two movies are to be considered. This means that only movies that have another movie as a prequel or sequel in the dataset qualify for the research. The review ratings will be divided between critic and consumer ratings.

The online critic review ratings are derived from the website Metacritic.com. These ratings are compiled from the reviews of a group of selected critics. Each movie has a score based on the weighted average review rating of these reviews by the selected critics. These scores on Metacritic.com range from 0 to 100. Zero being the worst possible score and 100 the highest score achievable.

(17)

16

Online consumer review ratings are collected from the same website. The website Metacritic compiles weighted average scores from votes of their users which are invited to rate the movies on a scale from 1.0 to 10.0. Furthermore, the website Metacritic also has a consumer review system, where users write a review and are subsequently asked to give the product a rating. In this research the consumer review ratings on movies that are hosted on the website Metacritic are also included. These consumer review ratings are also on a scale from 1.0 to 10. Therefore, these ratings will be multiplied by a factor of 10.

An independent variable in this research is previous ratings. Previous ratings will be measured by comparing the online review ratings of sequel movies to the rating of their corresponding previous movie in the series. A sequel movie is a movie that is released to be the follow-up of an earlier released movie. For example, Spider-Man 1 is the earlier released movie, and Spider-Man 2 is the sequel. Previous ratings will be measured according to the consumer ratings and critic ratings. Only movies that have a sequel in the form of another movie will qualify for research. These ratings are also based on Metacritic critics and

consumer’s ratings, on a scale from 1.0 to 10.0, ten being the most positive score and one the most negative. The consumer ratings on Metacritic will be multiplied by a factor of 10, to match the Metacritic critics ratings, which are on a scale from one to hundred, where one is the most negative score and one hundred is the most positive score.

When a sequel diverts from the original ‘genre’ a series of movies has, this will be noted as change in genre which indicates a creative input. Genre indications are gathered from the website Metacritic.com. To find whether changing genre has an influence on the average review rating by consumers and critics, all sequels in the data set are analysed. When it is found that in a series, a sequel changes the genre compared to the previous movie, a dummy variable is made. When a movie does not change, this is noted as 0, whilst a change is noted as 1. For clarity, a 0 indicates that an extra genre has not been added nor changed compared to the previous movie in the series. 1 means that the genre differs, be it an extra added genre, a removed genre or a change in genre. These movies will get labelled in the dataset for analysis of the mediator effect of changing the genre.

Furthermore, research will be done on series that are successful versus series that are unsuccessful. Successful movies are those that are above the review rating of 60.

Unsuccessful movies are those that are 60 or below this score. This is based on the Metacritic scoring system, where movies rated above 60 are above average, and beneath 60 are average or worse (2012).

(18)

17

4.3 Procedure

Data was acquired from an excel file previously used for research at the University of

Amsterdam. This file contained movie data for over 6000 movies, including Metacritic scores, box office and sequels. All movies that were not part of a series of at least two movies, were deleted from the file. In some cases, data was incomplete. For instance; a (sequel) movie was missing or a rating was absent. In these cases, the data-set was completed using the website Metacritic.com.

4.4 Analyses & Predictions

Since the conceptual model being tested is a mediator model, the analyses used will be based on Baron and Kenny’s steps for mediation (1986). Hence, a regression of sequel ratings on previous movie ratings will be performed, to check whether previous ratings is a significant predictor for sequel ratings. Next, the same will be done but now for sequel box office on previous review ratings. The second step is to perform a logistic regression, since the

mediator is a dichotomous variable. Change of genre is to be regressed on previous ratings, to check to association between previous ratings and change of genre. The following step is to regress sequel ratings on both change of genre and previous success, for ratings and for box office. Finally, the same procedures will be done for the movies that belong to the group of successful movies and for the movies that belong to the group for unsuccessful movies, to check whether there is a significant difference.

A significant positive effect of previous ratings on sequel ratings is predicted.

Furthermore, it is expected that previous ratings are a significant predictor of change of genre. Finally, it is to expected that when controlling for previous ratings, change of genre

significantly and positively affects sequel success. This mediating effect of change of genre on sequel success is predicted to be stronger than the direct effect of previous ratings on sequel success, both for measured by ratings and by box office. Finally, it is predicted that this influence is stronger for unsuccessful movies than for successful movies.

(19)

18

5. RESULTS

This section contains the results from the statistical analyses. First, the sample will be described. Subsequently, an overview of the used main variables and their correlations is presented. The aforementioned conceptual model is added including marked

hypotheses, betas and the significance levels according to the applied regressions.

5.1 Sample

The total amount of series in the data set add up to (N = 101). The average series in the dataset contains just under three movies, the total amount of movies in the data set is (N = 277). The critic ratings of said movies ranged from 9 being the lowest, to a score of 100 being the highest. The average being (M = 52,64, SD = 17,14). The consumer ratings ranged from 26 to 92 and averaged slightly higher (M = 65,64, SD = 13,737). This makes for an average online review rating of (M = 59,14, SD = 14,25).

Each movie was paired with its corresponding sequel. This resulted in (N = 176) pairs. In 74 of these pairs, the movie’s genre was the same as in the corresponding sequel. In 102 cases, the sequel changed in genre compared to its predecessor. The box office for average sequel is in US dollar (M = 133235320, SD = 102404676).

5.2 Correlations

In table 1, the correlations of the main variables are presented. Pearson correlations were performed to assess whether sequel average, consumer and critic ratings could be predicted from their predecessing average, consumer and critic ratings with genre change as a possible mediating variable. Average critic ratings have a large positive correlation with average sequel ratings r = .543, p = .000 and box office sequels r = .412, p = .000 Critic ratings also have a large positive correlation with critic sequel ratings r = .614, p = .000 and a medium positive correlation with consumer sequel ratings r = .388, p = .000. Consumer ratings have medium positive effect size on consumer sequel ratings r = 354, p = .000 and a medium positive relation with critic sequel ratings r = 378, p = .000. Change of genre has a small negative effect of correlation for average ratings r = .248, p = .001, critic ratings r = -.172, p = .022 and consumer ratings r = -.292, p = .000. The correlation between change of genre and the according sequel ratings are nonsignificant. Respectively p = .058 for average sequel ratings p = .061 for critic sequel ratings and p = .102 for consumer sequel ratings. The same goes for the correlation between change of genre and sequel box office, r = -.103, p

(20)

19

=.176. These correlations did not provide evidence of a linear relationship between genre change and sequel ratings nor for genre change and sequel box office.

Table 1. Descriptive statistics and correlations between the variables

*p<.05. **p.<.01

Variable name Mean S.D. 1 2 3 4 5 6 7

1. Average rating 62.79 12.56

2. Critic rating 56.28 16.18 .927**

3. Consumer rating 69.30 11.80 .858** .603**

4. Average rating sequel 56.20 14.56 .543** .553** .396**

5. Critic rating sequel 50.04 17.13 .560** .614** .

351** .946**

6. Consumer rating sequel 62.37 14.06 .441** .398** .393** .918** 739**

7. Change of genre .58 .50 -.248* -.172* -.292** -.143 -.142 -.124

(21)

20

5.3 Results

The first hypothesis was that previous ratings positively influences sequel ratings. Thus, first a linear regression is conducted to establish that predictor previous success is statistically significant related to the outcome variable sequel success. The average sequel review scores are regressed on the average score of their respective predecessor. The exact test results can be found in table 2. The regression is statistically significant F (1,174) = 72.579, p = .000 and the relationship in this sample is positive, as predicted: B = .629. This provides support for H1a: Previous average ratings positively influences average sequel success.

The relationship being significant also means that the first criterion for mediation is met. Furthermore, when looking at the adjusted R2 = .29, it indicates that 29% of the

variability of average sequel ratings can be explained by the previous average ratings. For an increase of 1 in a previous rating an increase of .629 in the respective sequel rating is

expected. When regressing critic and consumer ratings separately, both are significant

positive predictors of their respective sequel rating. For critic ratings F (1,174) = 105.046, p = .000, B = .650 and for consumer ratings F (1,174) = 31.827, p = .000, B = .330. This provides support for H1b: Previous consumer ratings positively influence sequel consumer ratings and H1c: Previous critic ratings positively influence sequel critic ratings.

(22)

21

Table 2

Linear regressions of previous review ratings on their respective sequel’s ratings

Average sequel ratings (DV) t Sig. 95,0% Confidence Interval for B B Std. Error Beta Lower Bound Upper Bound (Constant) 16,726 4,725 3,540 ,001 7,401 26,051 Average rating ,629 ,074 ,543 8,520 ,000 ,483 ,774 Critic sequel ratings (DV) (Constant) 13,479 3,711 3,632 ,000 6,155 20,803 Critic rating ,650 ,063 ,614 10,249 ,000 ,525 ,775 Consumer sequel ratings (DV) (Constant) 29,892 5,839 5,119 ,000 18,368 41,417 Consumer rating ,469 ,083 ,393 5,642 ,000 ,305 ,633 Note. N= 176 p<.05.

(23)

22

The second hypothesis was that previous review ratings also positively influence sequel’s box office. The same procedure is conducted as for the first hypothesis.

The regression is statistically significant for average, p = .000, B = 3368595, critic p = .000, B = 2943217 and consumer p = .001, B = 2108138 review ratings as expected. Thus, this provides support for H2a, H2b and H2c. An increase of 1 in average previous ratings seems to relate to an increase of approximately ~30 million in box office of the sequel. See table 3 below for raw regression results.

Table 3

Linear regressions of previous review ratings on their sequel’s box office Box office sequel (DV) t Sig. 95,0% Confidence Interval for B B Std. Error Beta Lower Bound Upper Bound (Constant) -78241974, 114 36266806, 398 2,157 ,032 -149824357 ,062 -6659591,1 66 Average rating 3368595,1 26 566304,85 5 ,412 5,948 ,000 2250838,8 36 4486351,4 15 (Constant) -32465801, 474 25064017, 333 1,295 ,197 -81936440, 474 17004837, 526 Critic rating 2943216,9 16 427742,43 5 ,464 6,881 ,000 2098951,1 58 3787482,6 73 (Constant) -12743565,3 09 45133843,7 25 ,282 ,778 -101827452, 376 76340321,7 58 Consumer rating 2108137,80 3 642282,404 ,242 3,282 ,001 840419,204 3375856,40 3 Note. N= 176 p<.05.

(24)

23

Following the steps discussed by Baron and Kenny (1986) to find a mediator, the proposed mediator, change of genre is now regressed on the independent variable previous success. Since the mediator in this research is dichotomous, a binary logistic regression analysis was performed to predict a possible change in genre of a sequel movie based on whether the relative scores of the previous review ratings. In the dataset, the outcome variable change of genre was coded 0 = does not differ in genre and 1 = does differ in genre. Data from 176 cases were included in these analyses.

For both average ratings as well as critic and consumer ratings as predictor variables, the chi-square test, compared with a null model / constant-only, proved to be statistically significant. For respectively average ratings, x2 (1)= 11.210, p = .001, critic ratings: x2 (1)=

5.285, p =.022 and for consumer ratings: x2 (1)= 15.964, p = .000. This provides support for

H3a: Previous average review ratings predict whether a movie changes in genre, whilst also providing support for H3b and H3c. The association between change of genre was stronger for consumer ratings with Cox and Snell’s R2=0.87 and Nagelkerke’s R2 = .117 than for critic ratings with Cox and Snell’s R2 = 0.30 and Nagelkerke’s R2 = 0.40. Table 4 shows the logistic regression coefficients, Wald statistics, and the estimated odds of a change of genre.

Interestingly, the tables show that for both ratings, when they increase, the odds that the sequel changes in genre, decrease. Important for the mediation model researched is that the logistic regression is statistically significant for average ratings p = .001 and both critic ratings p = .024 and consumer ratings p = .000.

(25)

24

Table 4

Binary logistic regression analysis: Prediction of change of genre in sequel from previous review ratings

Change of genre

(DV) B S.E. Wald df Sig. Exp(B)

95% C.I.for EXP(B) Lower Upper Average rating -,043 ,013 10,219 1 ,001 ,958 ,933 ,984 Constant 3,037 ,870 12,175 1 ,000 20,835 CriticRating -,022 ,010 5,077 1 ,024 ,978 ,959 ,997 Constant 1,579 ,584 7,327 1 ,007 4,852 ConsumerRating -,056 ,015 13,959 1 ,000 ,945 ,918 ,974 Constant 4,262 1,077 15,658 1 ,000 70,920 Note. N= 176 p<.05.

Next, a final regression was conducted to determine whether change of genre mediates the relationship between previous review ratings and sequel success, for sequel’s review ratings and/or sequel’s box office. The dependent variable sequel box office and sequel review ratings was regressed on both change of genre and previous review ratings.

In these regressions, average ratings, critic ratings and consumer ratings, change of genre were nonsignificant. The effect mediating effect of genre change on box office of the sequels was also statistically not significant. This result was not expected based on the predictions. However, it seems that there is not any statistically significant support for the hypothesis that a change in genre positively influences the sequel’s review ratings and/or sequel’s box office. Logically, this does not provide support for H4a and H4b, that when a sequel changes in genre compared to the prequel, this change positively influences the sequel’s review ratings and the sequel’s box office. For the raw regression results see table 7 and table 8.

(26)

25

Table 7

Regressions to explore whether change of genre is a significant predictor of the sequel review ratings, while controlling for the previous review ratings.

Average sequel Rating (DV) Unstandardized Coefficients Standardiz ed Coefficient s t Sig. 95,0% Confidence Interval for B B Std. Error Beta Lower Bound Upper Bound (Constant) 17,040 5,273 3,232 ,001 6,632 27,447 Average rating ,626 ,076 ,540 8,197 ,000 ,475 ,777 Genre Change -,263 1,938 -,009 -,135 ,892 -4,089 3,563 Critic rating Sequel (DV) (Constant) 14,600 4,151 3,517 ,001 6,407 22,792 Genre Change -1,279 2,107 -,037 -,607 ,545 -5,438 2,880 Critic rating ,643 ,064 ,607 9,973 ,000 ,516 ,770 Consumer sequel Rating (DV) (Constant) 30,281 6,565 4,613 ,000 17,324 43,239 Genre Change -,272 2,076 -,010 -,131 ,896 -4,370 3,826 Consumer rating ,465 ,087 ,390 5,342 ,000 ,293 ,637 Note. N= 176 p<.05

(27)

26

Table 8

Regressions to explore whether change of genre is a significant predictor of the sequel’s box office, while controlling for the previous review ratings.

Box office sequel (DV) t Sig. 95,0% Confidence Interval for B B Std. Error Beta Lower Bound Upper Bound (Constant) -78112045, 442 40468227, 260 1,930 ,055 -157990343 ,614 1766252,7 31 Genre change -109177,11 3 14907650, 350 -,001 ,007 ,994 -29534675, 097 29316320, 872 Average rating 3367529,5 06 586291,33 3 ,412 5,744 ,000 2210277,0 86 4524781,9 25 (Constant) -2515555,1 51 50771869, 699 ,050 ,961 -102731720 ,305 97700610, 002 GenreChange Previous -7137747,1 99 16084467, 403 -,034 ,444 ,658 -38886108, 078 24610613, 679 ConsumerRat ing 2019959,1 92 673746,43 1 ,232 2,998 ,003 690083,33 8 3349835,0 46 (Constant) -2515555,1 51 50771869, 699 ,050 ,961 -102731720 ,305 97700610, 002 Genre change -7137747,1 99 16084467, 403 -,034 ,444 ,658 -38886108, 078 24610613, 679 Consumer rating 2019959,1 92 673746,43 1 ,232 2,998 ,003 690083,33 8 3349835,0 46 Note. N= 175 p<.05

(28)

27

β = -.043** β = -.143

β = . 543** (.540)

Figure 3. Conceptual model for sequel review ratings, including marked hypotheses, significance levels and betas. Numbers in brackets are regression weights after the mediator has been controlled for. Success is measured by average review ratings.

β = -.043** β = -.001

β = . 412** (.412)

Figure 4. Conceptual model sequel box office, including marked hypotheses, significance levels and betas. Numbers in brackets are regression weights after the mediator has been controlled for. Success is measured by average review ratings.

Previous review ratings Sequel review ratings Change of genre Previous review ratings Sequel box office Change of genre

(29)

28

When distinguishing between successful and unsuccessful movie series, the results are as follows.

The chi-square test, compared with a null model / constant-only, proved to be

statistically significant only for successful movies. For unsuccessful previous review ratings: x2 (1)= .025, p = .875 and p = .876, B = -.008. This sample does not provide support for H5a.

For successful review ratings: x2 (1)= 4.247, p = .039 and p = .043, B = -.041. Indicating that

when ratings of successful movies go up by 1, the likeliness that the sequel has a different genre goes down. However, when looking further support was not found for H5b either. As for successful movies both the sequel’s review ratings and the sequel’s box office are not statistically significantly different when the genre changes. See table 9 below for raw regression results

Table 9

Regressions to explore whether change of genre in successful movies is a significant predictor of the sequel’s box office, while controlling for the previous review ratings.

Average rating sequel (DV) t Sig. 95,0% Confidence Interval for B B Std. Error Beta Lower Bound Upper Bound (Constant) 8,829 8,292 1,065 ,289 -7,580 25,239 Genre Change -,679 2,103 -,025 -,323 ,747 -4,841 3,482 Average rating ,753 ,116 ,504 6,466 ,000 ,522 ,983 Box office sequel (DV) (Constant) -144429178 ,759 68865629, 344 2,097 ,038 -280722806 ,105 -8135551,4 13 GenreChang ePrevious -5732472,2 48 17489958, 526 -,027 ,328 ,744 -40347270, 840 28882326, 343 Average rating 4371049,6 66 967030,07 8 ,379 4,520 ,000 2457177,1 24 6284922,2 08 Note. N= 175 p<.05

(30)

29

6. DISCUSSION

In this part, insight is created on a theoretical level. An overview of the concepts studied in this paper is presented along with the answers to the hypotheses and the research question. Furthermore, implications for practice and further research are highlighted. The effects and insights described are based on existing literature.

6.1 Summary

Factors that make a movie series a success are different for the producers of said movies than for the consumers of these movies. Consumers of course, want to be entertained by good movies, for which they would rate the viewing experience highly. When the experience is rated highly, a movie is a success, in the consumer’s eye. Producers would obviously like to deliver quality products, that in turn receive high ratings. However, producing a movie is an enormously expensive endeavour. Thus, producers face a lot of pressure and have incentives to generate as much financial profit as possible. As a result, producers tend to fall back on ideas and concepts that have proven to be successful in the past (Tschang & Szczypula, 2006). This development has been going on not just the movie industry but in different creative industries, such as the gaming industry and the music industry (Serrá, et al 2012). The fact that rehashing proven concepts is a popular practice amongst producers is no surprise,

research has shown that sequels tend to have higher sales, even when their review ratings tend to be lower on average (Moon Bergey & Iacobucci). The danger lurks that these

developments lead to creative bankruptcy. This paper researches whether this fear is well-grounded.

The main goal of the paper is to find out whether consumers appreciate and highly rate the same homogenous formula structurally and consistently. If this is the case, it is no wonder that producers capitalize on rehashing the same concepts. After all, why would a producer be creative when the consumers do not appreciate the efforts and risks taken in doing so. In other words, if the perceived quality remains the same for consumers, there is less incentive to go the extra mile for the sake of creativity. To find out whether creative bankruptcy is truly something to be aware of, the research question was; how does creative input in a successful formula influence online review ratings?

(31)

30

This research looked at the perceived quality of movies, expressed in both critic and consumer online review ratings, because these ratings are a driver of consumer’s product choice (Duan, Gu & Whinston, 2008). By making use of these online review ratings, it was possible to research whether previous success, measured by these ratings, can predict the success of a sequel. In other words, if a movie is successful, can we expect the sequel to be a success as well? According, this study hypothesized that previous ratings positively

influences sequel success. The concept of success is operationalised from the viewpoint of the consumer, measured by the respective critic and consumer online review ratings of the

according movies. The concept of success is also operationalised from the viewpoint of the producer, measured by box office. The results showed support for the first two hypotheses. First, average review ratings proved to predict and positively influence the average ratings of sequel, and thus the perceived success by the consumer. This relationship proved to be true for both critic ratings and consumer ratings. Critic ratings showed to predict and positively influence sequel critic ratings, and consumer ratings proved to predict and positively influence the sequel’s consumer review ratings. In addition, average, consumer and critic review ratings predicted box office success of the sequel. So far, this suggests that when a movie is

successful, most likely, the sequel of this movie is going to be successful as well. However, this does not yet explain what the role of creativity is in perceived quality and success.

To research the role of creativity, this research looked at the genre of the movies that the research was conducted on. More specifically, the research was focused on what happens when a movie changes genre, because if the genre within a series of movies would suddenly change, this would indicate input of creativity.

The results showed that review ratings predict whether or not the sequel of a movie will change. These results suggest that the lower the initial review ratings were, the higher the chance is that the sequel movie would have a change in its genre. This of course works both ways, higher initial movie ratings suggest that the probability that the sequel has a different genre decreases. These results do not support the view that previous success is a driver of creativity. Quite the contrary, these results hint that a change in genre is more likely to be a desperate measure, for when a movie series is in dire need to try a different approach, due to disappointing review ratings.

Furthermore, the results of the analyses did not prove that critics, nor consumers, rate sequels that differ in genre from its predecessor higher than sequels that are exactly the same in their genre, compared to the previous movie in the series. The sequel’s also show no

(32)

31

proven difference in box office when they change in genre compared to their predecessor. These results suggest that there is no evidence that previous success of movies act as a driver of creativity. According to these results it is more likely that successful formulas negatively influence future creativity. Moreover, sequels on average received a lower rating than original movies, and the average sequel that had a different genre compared to its predecessor was on average even lower.

When looking for an alternative explanation for this phenomenon, one might argue that these effect differ for successful series versus unsuccessful series, as producers of unsuccessful would be more inclined to radically change their formula and consumers would appreciate this more than when a successful series changes radically. Nonetheless, evidence to support these theories was not found. The outcome of these results did not match the

predictions and this leads to several discussion points.

6.2 Limitations & Discussion points

Several factors might have influenced the results. There might be an alternative explanation as to why successful formulas could actually be a driver of creativity, taking the results of this paper into account. For instance, rehashing of successful formulas might be a more common practice to squeeze every bit of profit out of a series, however these profits could actually be used to develop all kinds of new creative movies. When one of them turns out to be a success, they are likely to be transformed in a series to capitalize on the initial success. This is a possibility that this research did not account for. A future research would have to look at how different and creative each series is, compared to other series.

Moreover, this paper took genre as the main measure for creativity. A change of genre clearly points at an input of creativity that is different from the previous movie. However, most of the movies in the dataset were assigned multiple genres, and a movie has on average multiple genres. When only one of these changed, it was noted as a change in the dataset, however, maybe the change is not significant enough if only one genre changed. Also, there are a lot more measures that can be taken into account when looking for creativity. One could look at the storylines for example.

Despite the fact that the first and second hypotheses were accepted, there were limitations to this research. With more power, a larger sample, the mediation effect of genre might actually be confirmed, when testing for critic ratings and, the effect was on the edge of

(33)

32

being significant. However, it would be hard to collect more data, because Metacritic does not have data on every existing series. Furthermore, a series of two movies only gives one

comparison, so one would need to filter through upwards of ten thousand movies, simply because most movies are not part of a series.

6.3 Interpretation of results, contributions, practical

implications & future research

Interpreting the results, it seems that in general, movie ratings can show if the sequel will be successful both in terms of review ratings and in box office revenue. Thus, producers can use this as a tool to decide if they should make another sequel. Whether or not the sequel has a different genre, does not seem to have a big impact on the amount of success measured in review ratings and in box office revenue. However, review ratings do seem to predict the probability of a genre change in a sequel. The effects of a change of genre however, do not seem to result in better review ratings or more box office revenue. Looking at the models, it seems that there is not significant difference for these effects between successful and unsuccessful movies.

From these findings, that for the sake of profit, if producers have made a good movie, financially speaking it would be a very safe bet to make a sequel, as chances are it will a success again. For unsuccessful movies, it is most likely not worth the effort to add a lot of creativity and change the genre, to try and make the sequel a success, because the results of this study advise against it. A better idea would be to keep trying to find a formula that works, and then make sequels.

This paper contributes to current literature from a new angle, creativity. It gives an adds insight in the debate whether current technologies of consumer preference lead to more or less creativity, or even creative bankruptcy as Rosen puts it (2011). Success does not seem to spark a lot of creativity within the same series.

Future research could however focus on if the ones exploiting these series, invest a lot of this money back on other creative movies. They might as well be producing all kinds of creative products and use their successful formula as a cash cow. Future research should also focus on more factors that determine creativity, and more in more creative industries. Another advice would be to triangulate the findings with other research designs.

(34)

33

6.4 Conclusion

It seems that creative input in successful formula’s is deliberately limited, and although consumers do seem to perceive this as a sign of less quality, changing the genre in a series does not do the trick to raise perceived quality. Producers however, do not seem to care that much, because box office revenues increase with sequels, irrespectively of a possible genre change. To fully fathom how the relationship between success and creativity works, more research is required.

Although the phenomenon of increasing homogeneity in the creative industry is taking place, a part of this is logically due to globalisation and increase of worldwide communication methods. The long tail of the internet however, is still full of creativity. To state that this means the end for creativity is bit blunt, the large reach of global companies with successful formulas does add to an increase of homogeneity, as more people get reached with the same content. And although this paper does not find evidence that creative input in a series leads to higher perceived quality, it is clear that sequels are not appreciated better than original

movies.

In conclusion, the earning model of producers of movies seems to be based on utilizing and exploiting successful movie series, and being too creative seems to be a risky venture in ironically, this creative industry. This might mean that more and more sequels will be made, but does not mean that Netflix’s tactics of data analytics will kill all creativity, and even if it did, Netflix would probably still manage to keep people glued to the tube.

(35)

34

REFERENCES

Basuroy, S., & Chatterjee, S. (2008). Fast and frequent: Investigating box office revenues of motion picture sequels. Journal of Business Research,61(7), 798-803.

Basuroy, S., Chatterjee, S., & Ravid, S. A. (2003). How critical are critical reviews? The box office effects of film critics, star power, and budgets. Journal of marketing, 67(4), 103-117.

Chang, B. H., & Ki, E. J. (2005). Devising a practical model for predicting theatrical movie success: Focusing on the experience good property. Journal of Media Economics,

18(4), 247-269.

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.

Duan, W., Gu, B., & Whinston, A. B. (2008). Do online reviews matter? —An empirical investigation of panel data. Decision support systems, 45(4), 1007-1016.

Duan, W., Gu, B., & Whinston, A. B. (2008). The dynamics of online word-of-mouth and product sales—An empirical investigation of the movie industry.Journal of retailing,

84(2), 233-242.

Hu, N., Koh, N. S., & Reddy, S. K. (2014). Ratings lead you to the product, reviews help you clinch it? The mediating role of online review sentiments on product sales. Decision

support systems, 57, 42-53.

Krauss, J., Nann, S., Simon, D., Gloor, P. A., & Fischbach, K. (2008, June). Predicting Movie Success and Academy Awards through Sentiment and Social Network Analysis. In

ECIS (pp. 2026-2037).

Kücklich, J. (2005). Precarious playbour: Modders and the digital games industry. fibreculture, 5(1).

(36)

35

Lang, B., & Lawson, R. (2013). Dissecting Word-of-Mouth's Effectiveness and How to Use It as a Proconsumer Tool. Journal of Nonprofit & Public Sector Marketing, 25(4), 374-399.

Liu, Y. (2006). Word of mouth for movies: Its dynamics and impact on box office revenue. Journal of marketing, 70(3), 74-89.

Ludwig, S., De Ruyter, K., Friedman, M., Brüggen, E. C., Wetzels, M., & Pfann, G. (2013). More than words: The influence of affective content and linguistic style matches in online reviews on conversion rates. Journal of Marketing, 77(1), 87-103.

Metacritic. How We Create the Metascore Magic. (2012). Retrieved from http://www.metacritic.com/about-metascores

Moon, S., Bergey, P. K., & Iacobucci, D. (2010). Dynamic effects among movie ratings, movie revenues, and viewer satisfaction. Journal of Marketing,74(1), 108-121.

Oliver, R. L. (2014). Satisfaction: A behavioral perspective on the consumer. Routledge.

Ratner, R. K., Kahn, B. E., & Kahneman, D. (1999). Choosing less-preferred experiences for the sake of variety. Journal of Consumer Research, 26(1), 1-15.

Rosen, D. (2011). The other side: Creative bankruptcy.

Serrà, J., Corral, Á., Boguña, M., Haro, M., & Arcos, J. L. (2012). Measuring the evolution of contemporary western popular music. Scientific reports, 2.

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.

Situmeang, F. B., Leenders, M. A., & Wijnberg, N. M. (2014). History matters: The impact of reviews and sales of earlier versions of a product on consumer and expert reviews of new editions. European management journal, 32(1), 73-83.

(37)

36

Sood, S., & Drèze, X. (2006). Brand extensions of experiential goods: Movie sequel evaluations. Journal of Consumer Research, 33(3), 352-360.

Tschang, F. T., & Szczypula, J. (2006). Idea creation, constructivism and evolution as key characteristics in the videogame artifact design process.European Management

Journal, 24(4), 270-287.

White, M. M. (2009). The senescence of creativity: How market forces are killing digital games. Journal of the Canadian Games Studies Association, 3.

The high-wire act of being creative in the movie industry : Exploring the influence of successful product formulas on creativity and perceived quality