Timing decisions and their influences on product quality : an investigation in the videogame industry

(1)

Timing decisions and their influences on product quality:

An investigation in the Videogame industry

Daniel Maier 10826971 MSc in Business Administration Entrepreneurship and Management in the Creative Industries Supervisor: dr. F.B. Situmeang Publication date: 29.08.2015

UNIVERSITY OF AMSTERDAM

(2)

2

Statement of Originality

This document is written by Daniel Maier 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)

3

One of the main goals, firms are striving for, is to make profit and offer high quality products. The latter is said to be largely determined by expert reviews, who are expected to be related to consumer online reviews. Additionally, high quality products are said to positively affect sales. In fact, the aforesaid relations are influenced by many more variables. One of these variables, that was not included in previous research, is the time difference between product announcement and product release. Therefore, the purpose of this study is to investigate the effects of timing decisions on product quality. The empirical setting of this research was the videogame industry. A database consisting of information on 832 games and the variables of interest, such as global sales, announcement- and release dates of games, amongst others, made it possible to test the above-mentioned relations. Contrary to expectations, results of analyses showed that neither expert reviews, nor consumer online reviews were related to the variable time difference. Only global sales showed a positive and significant relation to time difference, although it was assumed to be indirect, as it may be the marketing expenditures that directly affect the relationship and therefore sales. The findings of this study, including its limitations, lead to implications for future research and the need to further investigate this topic.

(5)

5

1. Introduction

In recent years the importance of consumer online reviews raised steadily. Due to the fact that more and more people have access to internet, electronic Word-of-Mouth is becoming more popular every year. Consumers are more sophisticated than ever before, in this so called age of criticism, where they are able to make use of online platforms such as Amazon.com, Metacritics.com, to give their opinions on products (Labrecque, vor dem Esche, Mathwick, Novak & Hofacker, 2013). But why shall the industry as a whole and marketers particularly care about online reviews? Easily because these online reviews not only affect the sales of a product (e.g., Chevalier and Mayzlin 2003, Godes and Mayzlin 2004) but they also have very important implications for different kinds of management activities, including brand building, customer acquisition and retention, product development and quality assurance, as noted by Hu, Pavlou and Zhang (2006). Hu et al. (2006) further stated that Online Reviews are written, only if consumers are highly satisfied or highly dissatisfied, which leads to a U-shaped curve. As a result, they found that “[…] unless all the consumers write product review[s] or there is symmetric impact of bragging and moaning for consumers leaving reviews, the average review score does not converge to the true product quality“ (Hu et al., 2006, pp.329). However the product quality is said to be largely determined by expert reviews, even though it is hard to measure product quality by solely looking at expert reviews, as there are other influences, according to Reinstein and Snyder (2005).

One of these influences that recent studies did not consider, are the effects of timing decisions of new product preannouncements on the quality of a product and consumer’s perceived quality. To give a current example of the videogame industry, the game Assassins Creed Unity, which was released to the new generation of consoles PS4 and Xbox One, received the most negative Metacritic score in the history of the franchise. The Publisher, Ubisoft, is releasing this brand yearly, with announcements taking place early, more than 6 months prior

(6)

6 to launch. In 2014 they were developing for the new consoles, XBOX One and PS 4, for the first time. The final product was released thus not being finished, including bugs and a hardly working multiplayer. The consequences were negative expert reviews, as well as dissatisfied users and negative comments all over the online communities. This is only one of many cases in the industry, where the period between announcement date and release date may have affected the product quality. In order to avoid negative expert and consumer reviews and therefore negative WOM, the firm could have pushed back the release date. Even though the final product may offer more value when being delayed, this may harm the firm’s reputation and not only lead to disappointed consumers but the company itself may suffer from being perceived as less credible (Su and Rao, 2007).

Concerning the problematic mentioned above, with this thesis I would like to dig deeper in the field of timing decisions, with the goal to find relations between new product preannouncements and the quality of a product (expert reviews) as well as the perceived quality of a product (consumer reviews). Therefore my research question looks as follows:

“Timing decisions and their influences on product quality”.

The empirical setting of this study is the Videogame industry. My database consists of 832 games released on platforms Xbox 360 and PS3, in the time between 2005 and 2012, including variables such as timespan, global sales and Metacritic scores of both consumer online reviews and expert reviews, amongst others.

This study will further contribute to theory and practice. First with respect to theory it will add the variable timespan and test for possible relations to both, product quality and sales. Second, with respect to practice it will guide marketers to plan their timing decisions accordingly, giving advices about whether, when and what to preannounce, as well as how to use timing decisions in regard to a firm’s financial resources and size of the Publisher.

(7)

7 This thesis is structured as follows: first literature about the effects of expert and consumer online reviews is introduced, followed by a passage about timing decisions concerning new product preannouncements, which will lead to the introduction of the hypotheses and create a link to literature. Second, the research and the methodology are described. Third, the results are presented, followed by a discussion. Finally the conclusion will round off this thesis.

2. Literature Review

Time has always been an important companion, not only in our course of the day- the time we spend on work, with family, and other activities; moreover time is an important factor when we look at the industry itself. Through continuous advances in technology, firms are forced to instantaneously adapt to these developments in order to stay competitive and up to date. On top of that, consumers have never been as empowered as they are nowadays. This is due to the development of new media and internet, which leads to more sophisticated consumers that are involved in the co-production of products and therefore having a high impact on the value of the product, but also on the firm itself (Labrecque et al., 2013).

2.1 Expert reviews and consumer online reviews

However consumers are now also able to use the internet as a platform for criticism, be it through publishing on their own homepages or by using specific websites like Amazon.com or Metacritics.com to give their opinions on specific products. These online reviews vary in size and quality and are mostly based on rather subjective opinions than objective ones, thus involving emotional content. Oppositional to consumer online reviews, there are expert reviews, which are generally based on objective opinions, with reviewers providing sufficient knowledge and background information in the specific field of interest. Both, consumer online

(8)

8 reviews and expert reviews are said to have potential effects on the perceived product quality of consumers, as well as on overall sales performances of a product (Park, Lee & Han, 2007; Zhu and Zhang, 2010). However studies about this topic do not share a common view, even research conducted in the same industry gathered different findings.

In the movie industry for example, Eliashberg and Shugan (1997) state that expert reviews may have a predicting and/or influencing effect, while the influencing effect was statistically insignificant, they did find support in their assumption of expert reviews predicting the performance of a movie. Additionally, effects of expert reviews can further be dependent on the different types of audiences those reviews are addressed to.

Shrum (1991) suggests that effects of expert reviews differ, regarding the type of society, in his case he differentiates between highbrow and lowbrow communities. So there are more factors involved in measuring the effects of online reviews.

Another point is made by Gemser et al. (2007), who takes into account the valence and salience of a review, claiming that a specific product type (here it is differentiated between blockbuster movies and independent movies) is being influenced by either the size/quantity of the review (salience) or by the content of the review (valence).

Other authors, just like Zhu & Zhang (2010) investigated in the impact of online consumer reviews on sales and considered the popularity of a product as being part of the interaction between those variables. They found that in the videogame industry, consumer online reviews are more influential, the less popular a game is.

The Latter papers are only looking at one product and don’t involve sequential releases of products, so called sequels. Taking the videogame industry as an example, things look differently if the game is a new invention or a sequel of a franchise, which already has its fan base. This special case of dealing with sequels was discussed in the paper by Situmeang, Leenders, Wijnberg (2013), which was about the impact of reviews and sales of earlier versions

(9)

9 of a product on consumer and expert reviews of new editions. Their basic finding shows that both consumers and experts are influenced by reviews of past editions by their respective communities.

A different approach to this topic is done by Lee, Park and Han (2008), who tries to examine the relation between the proportions of negative online consumer reviews, the quality of the reviews and the involvement of consumers. Lee et al. (2008) concluded that the more involved consumers are, the greater the expected effect of the proportion of negative online consumer reviews on consumer’s product attitude.

This said, it is obvious that online reviews, be it by experts or by consumers themselves, are connected to many different factors, further depending on the industry itself. Moreover it is difficult to determine, which factors are having what kind of impact on the product quality and sales. Besides of the characteristics of online reviews, there are two more important elements, that is the quality of the product on the one hand, which is more or less determined through expert reviews, and the expectations created by the firm on the other hand.

Regarding the product quality, Sacranie (2011) was investigating in the videogame industry about the quality of videogames and its influences on the consumer, and thus the overall sales of the product. His findings showed that high expert reviewed games showed higher sales, so according to him, the quality of a game might be a predictor of its sales. Overall he did not find a causality between the reviews and sales. In addition he mentions that for future studies, the contribution of marketing expenditures as well as the development time and costs shall be taken into account, as he did not include them in his studies. So basically it can be assumed that the quality of a product contributes somehow to the process of a consumers’ purchase decision and again that time may play a major role in the whole process. Additionally it is the consumer expectations that are somehow related to the quality of a product as well as

(10)

10 its performance. The following papers try to connect the characteristics involved with the product, to the expectations created at the consumer side.

In the research conducted by Olshavsky and Mill (1972), the authors contributed to existing literature by focusing on the positive disconfirmation theory, linking the relationship between consumer’s expectations and perceptions to the nature of their evaluation, taking into account the over- or understatement of a firms’ product quality. Their findings however included that their suggestion, both overstatement and understatement should result in unfavourable product evaluation, was not supported. As this research was conducted back in 1972, where the internet was no yet made available to consumers, the results may not be sufficient for the current situation. Even though overstatement and understatement may not lead to significant bad online reviews, consumers got more sensible and sophisticated through opportunities offered to them by new media and social networks, which may recently rather lead to bad WOM and result in a bad brand image, if over-or understatement is done excessively by the firm.

This assumption is supported by another author, Anderson (1973), who claims that people do have a threshold point of rejection, concerning the difference between what the firm claims the product has to offer, compared to the actual performance. As a consequence, consumer will perceive the product as less favourable and in the worst case, just abandon it. Here it is to add that the author only refers to relatively simple or easily understood products, for more complex products, Anderson (1973) expects different results, due to a higher level of ambiguity and the dependency of consumers on more information.

Taking into account the gap between a firm’s statement of later product features and its actual performance in regard to ratings by expert reviews, as well as the expectations created on the consumer side, firms need to adapt their strategies and be aware of several variables, in order to still be able to offer high quality products that satisfy their customers and generate high

(11)

11 sales. One of these variables that should be included in future research is time difference. It is concerned with timing decisions about new product announcements and release dates. Only few studies have been dealing with this topic in the last couple of years. While past studies, such as Kohli (1999) and Lilly & Walters (1997) mainly focus on providing a guiding framework for firms to take better decisions concerning the timing of new product announcements, recent studies such as Su and Rao (2010), try to include the competitors reactions (simultaneous, sequential moves), cannibalization effects and product quality into their optimal decision making processes. What Su and Rao (2010) indicate as future implications, has partly lead to the research of this master thesis. According to them it would be valuable to examine from the demand side, how consumers’ perceptions, preferences, and choices are affected by firms’ preannouncement.

Therefore, in the next part of the thesis, first timing decisions of early and late product preannouncements are discussed, second the link between the discussed literature and research question is explained and hypotheses are stated.

2.2 Timing decisions of new product preannouncements

In the marketplace, product preannouncements are generally seen as an important signalling tool, a “formal, deliberate communication before a firm actually undertakes a particular marketing action such as a price change, a new advertising campaign, or a product line change”, as defined by Eliashberg and Robertson (1988, p.282).

It is important to note that there is a difference between new product announcement (NPA) and new product preannouncement (NPP). According to Koku Koku, Jagpal, and Viswanath (1997), NPA are more certain as they are made closer to the release date of a product, whereas NPP usually concern announcements that are less certain to both the firm and customer, as they are made further away from the release date. Even though literature does not agree about the

(12)

12 distinction, as it depends on the industry itself, Su and Rao (2010) suggest that any product that is announced less than one month prior to its release can be considered a NPA. Oppositional NPPs are used if a product is announced further than one month to its release date. Due to the fact that my research is done in the videogame industry, where products (games) are mostly announced more than one month prior to their release date, I will use the term NPP, whenever I talk about product preannouncements.

According to the Literature, there are different goals involved regarding new product preannouncements (NPP). Hereby it is to differentiate between goals that concern the target groups, for instance customers or competitors and also between goals that concern the methods that are used. In practice, companies use preannouncements to inform about changes in pricing and distribution channels, market entry and earnings, even though some managers might preannounce in order to create a first mover advantage. In addition preannouncements can be used to prevent customers from buying similar, currently existing products of competitors (Su and Rao, 2010).

In the following part the questions about whether, when and what to preannounce will be discussed.

2.2.1 Whether to preannounce

Before formally introducing a product, a firm has to decide about whether to preannounce it or not. Therefore Su and Rao (2010) suggest that a company should preannounce a product if it expects the benefits of announcing it to be greater than the costs that may occur. Thus further points have to be taken into account when deciding to preannounce. According to Eliashberg and Robertson (1988), there is a negative correlation between the variables market dominance and company size, however they also found a positive correlation between factors such as competitive environment and customer switching costs to the likelihood of preannouncements.

(13)

13 In another research done by Robertson, Eliashberg and Rymon (1995), results showed a positive relation between competitive reactions and the perceived hostility and credibility of a NPP signal. In fact the decision whether a firm is going to preannounce its product or not is mostly being influenced by the ability of the firm to fulfil the commitment of NPP to the specified future date. Furthermore it is important that firms conduct an integrative evaluation of all target audiences in order to make the right decisions to either preannounce or not. This evaluation is vital due to the fact that there exists information asymmetry between firms and audiences, while firms might perceive uncertainty differently than target audiences (Su and Rao, 2010).

2.2.2 When to preannounce

Generally it is to say that the timing of a preannouncement is playing a crucial role regarding the success of a new product. It is to note that preannouncements can take place at any of the five stages of the process of new product development, which are: concept generation; product design; engineering analysis; process analysis and design; and prototype production and testing. Beside the fact that the launch date and the timing of the preannouncement are defined together, it is the extent to which information is available to the firm that influences decisions about the timing of preannouncements (Su and Rao, 2010). For instance firms expect an increase of competitive reactions the earlier a preannouncement is done, in other words, the bigger the timespan between preannouncement and release date, the higher the probability of competitors to show competitive actions (Lilly and Walters, 1997). In case a firm has to delay the launch of a product, it is usually based on firms’ abilities, firm’s motivations and resource constraints (Wu, Balasubramanian, and Mahajan 2004). However, in the latter case, the timing of a NPP, either early or late, might have affected the decisions regarding the delay of the product launch. Therefore it is important for a firm to consider the advantages and disadvantages of early and late NPP.

(14)

14

2.2.2.1 Advantages and disadvantages of early NPP

Firms that engage in early NPPs have to consider its advantages and disadvantages, as they are not only affecting the firm itself but also the customer. If a company chooses to go for an early NPP, the main advantage is to be perceived as a recognized innovator in the industry, which may lead to first mover advantages (Rabino, Samuel and Moore 1995). On the other hand firms may also encounter disadvantages, which are a potential harming by competitive reactions, as well as cannibalization of sales of a firm’s own products.

Buyers on the other hand may benefit from early NPP, because they have more time to plan for the new product purchase. However early NPP may also lead to disadvantages for buyers, such as it may strengthen confusion and anger, for instance if the date of introduction of the product is not in line with the buyers’ expectations (Lilly and Walters, 1997).

2.2.2.1 Advantages and disadvantages of late NPP

Late NPPs are usually done few months before the actual release date of a product but they can be done even one month prior to launch. Up to the time the product launches to the marketplace, marketers are able to hold back information from potential competitors. Another advantage is that late NPPs can be cost-saving, due to already greatly defined markets and products that are targeted. This means that few or no extra announcements have to be done to answer potential questions buyers may come up with. Furthermore, by using late NPP’s, managers are able to downsize possible expectations buyers may create on forthcoming new product introductions. As a result, managers feel less pressured during the process of launching a new product (Lilly and Walters 1997).

However late NPPs can also lead to higher costs and effort after launch, due to the lack of time that firms are facing, it makes it difficult for them to effectively convince buyers of the

(15)

15 innovativeness and freshness of the new product. Furthermore, due to the lack of time for planning an introduction of a new product, the coordination between manufacturer and distributor may further end up insufficient (Lilly and Walters 1997).

Generally, both early and late NPPs can be seen as advantageous or disadvantageous, but it is up to the firm itself to make the right decision on which method to use.

2.3 What to preannounce

Equally important for a firm to decide on whether to announce early or late, it is to decide over what information to include in the announcement. According to Sorescu, Shankar and Kushwaha (2007), it is the specific details that are given to a targeted audience, which affect the effects of NPPs. For instance, being specific about product details in a very early NPP, it is even harder to meet these criteria for the desired launch date. As a result, NPPs that include more detailed information are usually made closer to the launch date, in other words, late NPPs will be used (Su and Rao, 2010).

Another aspect of NPP is that it can further be seen as a Signalling Instrument. Michael Porter (1980, p.75) defines Market Signals as “[…] any action by a competitor that provides a direct or indirect indication of its intentions, motives, goals or internal situation.” In fact, specific signals are send by companies in order to change perceptions of receivers, mostly consumers, on some parameters like true quality of a product for instance, which can’t be directly observed (Su and Rao, 2010). According to research in economics and marketing, the product quality is signalled by advertising and pricing (Curry and Riesz, 1988; Gerstner, 1985; Kihlstrom and Riordan, 1984; Prabhu and Stewart, 2001). Customers may then change their perceptions, make their own interpretations, change prior beliefs, draw conclusion about these parameters or adjust their actions, all based on how they have received the market signals (Su and Rao, 2010).

(16)

16 Therefore, NPP can also be a risk for the company. The sender, the firm itself, is not always certain about various parameters of its new product, whether that is about the launch time, the quality or price. In the videogame industry, the parameter price is not of importance, as new games are priced very similar. Regarding the launch date and the quality of a product, firms are more or less uncertain and by using NPPs during an early stage of the development it is almost impossible for them to determine all these factors at that point in time. This uncertainty adds even more complexity to NPP decisions. One of the possible consequences are product delays, which in turn results in the risk of losing credibility, in case the firm fails to deliver what it has promised to its consumers (Su and Rao 2010).

2.3 Hypotheses and link to literature

As discussed in the timing decisions part, early NPPs offer the opportunity for firms to advertise their products more efficient and to prepare better for the launch of a new product release. In addition, more information can be provided to consumers and stakeholders, enabling them to plan their purchase better (Lilly and Walters, 1997). Anderson (1973) further concludes that if a firm provides more information to the customer, a higher evaluation can be obtained.

Taking into account that this research is done in the videogame industry, products can further be divided into new products or sequels. According to a research done by Situmeang, Leenders, Wijnberg (2013), both consumers and experts are influenced by reviews of past editions by their respective communities. Additionally, Metacritic scores of game sequels are further ranging on the same or similar score (product quality) as their predecessors. Thus expectations may be high and lead to dissatisfaction if the gap between the features of the final product and consumer’s expectation is too big (Anderson, 1973), if the game is a sequel, consumer mostly know what they can expect, as the products’ core elements stay the same with every new entry in the series. Therefore, in line with the findings of Situmeang et.al. (2013), I

(17)

17 suggest that sequels are kind of guaranteeing consumers a high product quality. Concerning early NPPs, the positive effect on the Metacritic score, alias the product quality, will be stronger, as firms may have more time to use market signals to influence consumers’ perception of the true quality of a product (Su and Rao, 2010).

In opposition to early NPPs, I suggest that there is a negative relation between late NPPs and the Metacritic score. The closer the preannouncement is made to the actual release date, the less time is available to the firm to convince consumers of the true product quality (Lilly and Walters, 1997). Thus this might help firms to downsize possible expectations of consumers, it is the lack of time for providing sufficient information about the product features and quality through market signals that causes customers to either misinterpret signals or change prior beliefs about the product (Su and Rao, 2010). Here it is to add that especially sports games are announced late, although Metacritic scores for those games mostly range on high levels. This is due to the fact that sports games like FIFA or NHL first of all aren’t games, where producers may be able to innovate and influence the products’ core elements, so that consumers know what to expect. Second, sports games are sequels, based on yearly releases, with the product quality being stable over time. Stability of quality of sequels was also confirmed by Situmeang (2013).

In general, sequels are expected to lessen the negative relation between late NPPs and the Metacritic score. Due to the fact that consumers and experts orientate their perception of product quality on prior releases, lack of information is not as important as it would be if the game was not a sequel. The negative relation between late NPPs and Metacritic scores will thus be less strong if the game is a sequel.

In addition to sequels, the size of the publisher is expected to mediate the relation between NPPs and the Metacrtitic score of expert reviews. On the website Metacritic.com, a yearly list of major and mid-sized publishers is made available, ranked according to the number

(18)

18 of high quality games released. The list shows that major publishers release more games with high quality (high Metacritic score of expert reviews) than mid-sized ones do. As product quality is said to be determined by expert reviews (Reinstein and Snyder, 2005), it is suggested that major Publishers are mediating the relationship between NPPs and the Metacritic score of expert reviews.

A similar effect is expected for the relationship between NPPs and the Metacritic score of consumer online reviews. In this relation, size of Publisher may send signals about the product quality as well, as major Publishers may be seen as more reliable and trustworthy, due to qualitative stable releases in the past and an already created reputation of the firm itself. Although a bad reputation of a major Publisher may backfire and besides suspicion, lead to more negative evaluations by consumers. However, in this research, reputation was not concerned.

What is more, it can be expected that Early NPPs are rather done by major Publishers, as they possess more financial resources than mid-sized or smaller Publishers do. Therefore the following hypotheses are suggested:

H1a: NPPs are influencing (related to) the Metacritic score of expert reviews H1b: Sequels moderate the relation above

H1c: Size of Publisher mediates the relation above

H2a: NPPs are influencing (related to) the Metacritic score of consumer online reviews H2b: Sequels moderate the relation above

H2c: Size of Publisher mediates the relation above

Considering the relation between expert reviews and consumer online reviews, literature provides several suggestions. As discussed earlier in the review part, both, expert reviews and consumer online reviews are having different kind of effects on other variables and relations,

(19)

19 for instance, both are said to influence sales of a product (Park et.al, 2007; Zhu and Zhang, 2010). Although the question remains how timing decisions may influence sales. In his research, Anderson (1973) talks about the sales-oriented view, which concludes that “the higher the level of consumer product expectations stimulated by promotion the better because the primary objective of the promotional mix is to sell the product.” This view, in connection with the findings of several researchers (Curry and Riesz, 1988; Gerstner, 1985; Kihlstrom and Riordan, 1984; Prabhu and Stewart, 2001), who found that product quality is signalled by advertising, combined with the fact that early NPPs may provide more time to firms to send market signals about true quality of a product, leads to the assumption that sales are higher (lower) if early (late) NPPs are used.

Further I suggest that both, the Metacritic scores of expert reviews and consumer online reviews are mediating the relation between NPPs and sales, because both scores show relations to NPPs as well as sales. As discussed earlier in H1a, H2a, NPPs and Metacritic scores are expected to be related. According to Sacranie (2011), product quality is related to sales, hence a higher product quality, in this context a higher Metacritic score of expert reviews, may lead to higher sales.

Considering the mediating effect of Metacritic score of online consumer reviews on sales, it is to say that this effect may be stronger, the less popular a game is, as proposed by Zhu et.al. (2010). Following the discussion above, three more hypotheses are added:

H3a: NPPs are influencing (related to) sales

H3b: The above relation is mediated by the Metacritic score of expert reviews

(20)

20

3. Data Collection and Methodology

In the following part of this thesis, first the data collection is described. Subsequently, each of the variables are operationalized, such as dependent, independent, moderator, mediator and control variables. The latter is followed by the description of the method.

3.1 Data collection

The empirical setting of this research is the videogame industry. Furthermore this study is a quantitative research that is based on a videogame database available at the University of Amsterdam. It already includes information for the following elements: sales numbers of past Videogame releases, release dates of games, publisher, genre, platform and both, Metacritic scores for expert reviews and consumer online reviews.

In order to be able to answer my research question, the database was extended by the following elements: announcement date, major publisher or mid-size publisher, timespan between announcement date and release date, measured in months, the channel of announcement (Event or Homepage) and the information about the game being a sequel or a new product release. The information about the announcement dates of past videogame releases, as well as information about the channel of announcement and if the game was a new product release or a sequel, were taken from the website IGN.com. Subsequently, the timespan between the announcement date and release date was calculated and stated in months.

However, the differentiation between major publisher and mid-size publisher turned out to be rather problematic, as hardly any information could be found. Thus on the website Metacritics.com, yearly lists (2010-2014) are provided, ranking the major and mid-size publishers according to the number of quality games released. Their first list was done in 2010 and due to the fact that my database consists of games released in between the years of 2005-2012, it got chosen for being the most adequate one for my research, as Publishers amount of

(21)

21 game releases per year were quite stable in that time. Regarding the details about the list, the ranking of publishers will only be used for the variable size of publisher, for my database though, only the names of the publishers and their differentiation in either major or mid-sized Publisher will be taken into account and serve as reference.

What is more, my research only concerned game releases on platforms Xbox 360 and PS 3, which represent one very similar generation of hardware in the years of 2005-2012. By focusing on only two platforms that are similar in its product features, a higher comparability was guaranteed. This lead to the decision to exclude any other platforms such as Nintendo’s Wii or even platforms of older generations, such as Xbox and PS2.

Another reason for focusing on the generation of consoles in the time from 2005-2012 and not on the new generation of consoles, Xbox One and PS4, was due to global sales of games released. Global sales had to be final, with no additional sales taking place, as this would have biased the study. Games released for the new generation of consoles are still selling, therefore Xbox One and PS4 games were excluded from this study.

Furthermore this specific generation, Xbox 360 and PS3, saw an increase of consumer online reviews, thanks to the emerging availability of the internet. Regarding the importance of consumer online reviews for this research, choosing an older generation would have been inadequate, as the internet was not as widespread as in 2005-2012 and consumers have not been as sophisticated as in more recent years like 2005-2012 either (Labrecque et.al , 2013).

3.2 Operationalization of Variables

The independent variable that was used in this study is NPP, which consists of the time difference between the date of announcement of a game and its release date, measured in months. To be more precisely, the independent variable NPP can be divided into early and late NPPs, whereby early NPPs were defined as having a timespan of more than six months and late

(22)

22 NPP on the other hand were defined as having a time span of six or less months. The classification was based on the suggestion made by Su and Rao (2010) about NPPs, even though I had to adapt and create a more specific definition concerning the video game industry. Based on the data gathered, it was possible to differentiate early and late NPPs as discussed above, as games usually get announced at E3 press conference, the biggest and most important videogame exhibition worldwide, taking place yearly around May/June. There, games are either announced and released the same year or only announced but released at a later date. If the games was announced and released in the same year after E3, then late NPPs took place. In opposition, early NPPs took place if the game was indeed announced at E3 but the release followed at a later date, more than 6 months away from the announcement. Regarding announcements aside of the E3, early and late NPPs were defined in the same manner.

Next to the independent variable NPP, also dependant variables were used. These variables are Metacritic score of expert reviews and consumer online reviews as well as global sales.

The Metacritic scores of expert reviews and consumer online reviews were taken from the website Metacritics.com. There, opinions and scores of the most respected critics writing online and in print are collected and a final score is calculated, which is called Metacritic score. Similar to the expert reviews, they also provide consumers an option to leave their opinions and rate the games themselves, and in the end they calculate an average score out of the user scores that were generated. I further divided the Metacritic scores of expert reviews into five categories. This categorization is based on the information provided on metacritics.com. There games are divided as follows: a Metacritic score of 90-100 (coloured green) is defined as Universal Acclaim; 75-89 (coloured green) means that reviews were generally favourable; if a game scores between 50 and 74 (coloured yellow), it has average reviews; scoring between 20-49 (coloured red) means that the game had generally unfavourable reviews and Metacritic

(23)

23 scores beneath 20 (coloured red) are facing overwhelming dislike. In my database, none of the games had a score beneath 20, so this category was left out.

For Metacritic scores of consumer online reviews the same colour coding as for expert reviews is used (green, yellow and red). Therefore the same categorization can be applied, thus it is to add that scores are stated from 0-10 and not from 0-100, as expert reviews are.

Sales numbers of games of different regions worldwide were included in the database I was working with. These numbers were based on the information to be found on NPD.com. For this study however only global sales of each of the consoles, Xbox 360 and PS3 were taken into account, in order to have meaningful, internationally comparable numbers.

Out of the mentioned variables above, the Metacritic scores of expert reviews and consumer online reviews are also used as a mediators, because they are expected to mediate the relationship between the independent variable NPP and the sales of released games (H3b, H3c). In addition the variable size of Publisher (Major Publisher, no Major Publisher) also serves as mediator, as it is assumed to affect the relationship between NPPs and the Metacritic scores of expert and consumer online reviews (H1c, H2c). Major Publishers are expected to be linked to early NPPs, because they have more financial resources than smaller Publishers and therefore are able to announce early and use the extended time span to send signals about the product quality of a game via advertisement and other marketing activities, such as beta tests. This should result in high Metacritic scores of expert reviews, which are in turn said to lead to higher sales (Sacranie, 2011).

The variable Sequel on the other hand is used as moderator in this study, because it is expected to influence the relationship between the independent variable NPP and both Metacritic scores of expert and consumer online reviews. Sequels can be defined as continuations of already released products, in this study it is games. So whenever a new released

(24)

24 game is connected to an original one by an indication of a name or number (Tomb Raider, Rise of the Tomb Raider/ FIFA 10, FIFA 11), continuing the franchise, it is defined as a sequel.

3.3 Method

In order to prepare the database for the analysis, several steps had to be taken. The variables Metacritic score of expert reviews as well as consumer reviews included 0 as values in the database, even though the 0 was not meant as having a 0 score, but as a missing value. Therefore the zeros in each of the two columns were deleted so that spaces were left empty, which in this case prevented a biased outcome.

What is more I had to deal with missing values in my database, as nearly every variable was affected. Overall the percentage of missing values for the variables were between a minimum of 6% and maximum of 16%. Due to the extensive database and the fact that my data was missing at random (MCAR), which means that data was collected randomly and is not dependant on any other variable in the data set, the use of case deletion was valid (Rubin, 1976). All analyses were done with IBM SPSS Statistics 23.

4. Results

This section includes the results of the analyses. First of all, the descriptive statistics of the sample are discussed. As a next step, correlations will be presented. This is followed by testing

Table 1: Descriptive Statistics, Frequencies

metaScor e userScor e Time Difference (rounded in months) Major P or not (1/0) event (1=event, 0=Homep age) Seque l Platfor m Globa l early NPP (1) late NPP(0) valueMeta EX (0-4) valueMeta CO (0-4) N Valid 776 689 734 734 734 826 826 724 733 776 690 Missin g 50 137 92 92 92 0 0 102 93 50 136

(25)

25 the hypothesis through several regression analyses. Last but not least, additional analyses will be presented and explained.

4.1 Descriptive statistics

The first table presented a summary of frequencies of the sample, including missing values. Taking into account the missing values, the percental distribution of the variables of interest was as follows: 63.6% of games were published by major Publishers and 25.3% were published by minor Publishers. In addition, 31.7% of games were announced at an event. In most cases it was the E3 (Electronic Entertainment Expo). Though most of the games of my sample, 57.1%, were announced through the Publishers Homepage. The distribution of frequency of the variable sequel was almost equal with 45% of games being sequels and 55% of games being non-sequels. A similar distribution was present with the variable platform, 56.8% have been Xbox 360 games, whereas 43.2% have been PS3 games. Concerning the independent variable time difference, the differentiation between late and early new product preannouncement showed that 50.2% of games were announced early and 38.5% of games were announced late. Furthermore, by dividing the Metacritic scores of expert reviews into different categories, frequency distribution showed that average scores (50-74) and favourable reviews (75-89) turned out to be the majority, with 41.2% and 38.4% respectively. Similar to expert reviews, Metacritic scores of consumer online reviews mostly ranged between average and favourable reviews, while average reviews made up 37.3%, whereas favourable reviews made up 39.1%. Table 2 shows descriptive statistics including the mean, standard deviation, skewness and kurtosis. Here it is to mention that the means of Metacritic scores of expert reviews and consumer online reviews were almost equal (70.6 vs. 7.168). This is interesting, as it implies that experts, who are said to determine product quality (Reinstein and Snyder, 2005), share the same opinion about quality as consumers do, who perceive product quality. Furthermore,

(26)

26 the mean for the independent variable time difference was 10.90 and the standard deviation 9.540, meaning that on average, games are announced early, about 11 months prior to release.

4.2 Correlations

On the following page, table 3 shows the Correlations between each of the variables. In order to be able to differentiate the independent variable time difference between late and early NPPs, a dummy variable (early NPP (1) late NPP (0)) was added. As a result correlations slightly differed between these two independent variables. Time difference had a low positive, but insignificant correlation with the dependant variables Metacritc scores of expert reviews (r = .051, p > .05) and consumer online reviews ((r = .025, p > .05).

Table 2: Descriptive Statistics

N Minimu m Maximu m Mean Std.

Deviation Skewness Kurtosis

Statistic Statistic Statistic Statistic Statistic Statistic Std. Error Statistic Std. Error metaScore 776 26 98 70,60 14,390 -,594 ,088 -,237 ,175 userScore 689 2,0 9,4 7,168 1,2581 -1,226 ,093 1,867 ,186 Major P or not (1/0) 734 0 1 ,72 ,452 -,956 ,090 -1,089 ,180 event (1=event, 0=Homepage) 734 0 1 ,36 ,479 ,598 ,090 -1,646 ,180 Sequel 826 0 1 ,45 ,498 ,200 ,085 -1,965 ,170 Time Difference (rounded in months) 734 0 64 10,90 9,540 1,837 ,090 4,271 ,180 Global 724 ,01 11,73 ,8068 1,21676 3,969 ,091 22,380 ,181 early NPP (1) late NPP(0) 733 0 1 ,57 ,496 -,268 ,090 -1,934 ,180 valueMetaEX (0-4) 776 1 4 2,43 ,743 -,036 ,088 -,339 ,175 valueMetaCO (0-4) 690 1 4 2,45 ,647 -,272 ,093 -,320 ,186 Valid N (listwise) 591

(27)

27 Table 3: Correlations metaScore userScor e Time Difference (rounded in months) Major P or not (1/0) event (1=event, 0=Homep

age) Sequel Global

early NPP (1) late NPP(0) metaScore Pearson Correlation 1 ,586** _,051 _,280** _,148**_,333** _,491** _,085* Sig. (2-tailed) ,000 ,178 ,000 ,000 ,000 ,000 ,023 N 776 683 711 711 711 776 692 710 userScore Pearson Correlation ,586** ₁ _,025 _,076 _-,013 _,063 _,196** _,067 Sig. (2-tailed) ,000 ,528 ,053 ,738 ,099 ,000 ,089 N 683 689 644 644 644 689 626 644 Time Difference (rounded in months) Pearson Correlation ,051 ,025 1 -,056 ,140**_-,216** _,187** _,644** Sig. (2-tailed) ,178 ,528 ,130 ,000 ,000 ,000 ,000 N 711 644 734 734 734 734 666 733 Major P or not (1/0) Pearson Correlation ,280** _,076 _-,056 ₁ _,168**_,198** _,185** _-,120** Sig. (2-tailed) ,000 ,053 ,130 ,000 ,000 ,000 ,001 N 711 644 734 734 734 734 666 733 event (1=event, 0=Homepage) Pearson Correlation ,148** _-,013 _,140** _,168** ₁ _,043 _,221** _,073* Sig. (2-tailed) ,000 ,738 ,000 ,000 ,243 ,000 ,049 N 711 644 734 734 734 734 666 733 Sequel Pearson Correlation ,333** _,063 _-,216** _,198** _,043 ₁ _,187** _-,222** Sig. (2-tailed) ,000 ,099 ,000 ,000 ,243 ,000 ,000 N 776 689 734 734 734 826 724 733 Global Pearson Correlation ,491** _,196** _,187** _,185** _,221**_,187** ₁ _,134** Sig. (2-tailed) ,000 ,000 ,000 ,000 ,000 ,000 ,001 N 692 626 666 666 666 724 724 666 early NPP (1) late NPP(0) Pearson Correlation ,085* _,067 _,644** _-,120** _,073*_-,222** _,134** ₁ Sig. (2-tailed) ,023 ,089 ,000 ,001 ,049 ,000 ,001 N 710 644 733 733 733 733 666 733

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

(28)

28 However the categorical independent variable early/late NPP showed a significant slightly positive correlation with the dependent variable metaScore of expert reviews (r = .085, p < .05). The correlation between early/late NPP and userScore was slightly positive but insignificant (r = .067, p > .05). Other significant positive correlations of the independent variable time difference included the variables global (sales) (r = .187, p < .05) and event (r = .140, p < .05). A significant negative correlation existed with the variable sequel (r = -.216, p < .05).

The dependent variables metaScore and userScore had a significant and positive relation (r = .586, p < .05). This could suggest that expert reviews may indeed influence consumers. Furthermore, the dependant variable metaScore showed positive and significant correlations to several other variables. The strongest correlations, beside of userScore, were found with the variables global (r = .491, p < .05) and sequel (r = .333, p < .05). In opposition to metaScore, userScore did not show any positive significant correlation with any variable other than global (r = .196, p < .05).

Regarding the variable global, it is remarkable that it was correlated to every other variable in a positive and significant way, while metaScore showed the strongest correlation. This may not be that surprising as sales are dependent or related to many factors and variables.

4.3 Regression analyses

In this section, results of regression analyses are shown and hypotheses are explained. First of all, hypotheses H1a was tested. It suggests that NPPs are influencing or are related to the Metacritic score of expert reviews. As mentioned earlier, the independent variable time difference (NPPs) could be divided into another more detailed categorical variable early and late NPP. Therefore two separate regression analyses were done, the basic independent variable time difference did not show any significant relation to the dependant variable metaScore (p > .05). However, the categorical independent variable early/late NPP showed a significant

(29)

29 positive relation to metaScore of expert reviews (p < .05). Here it is to add that only 0.6% of the total variability of metaScore is explained by the categorical independent variable early/late NPP (R² = .006). This said, it can be concluded that H1a is only partly supported, taking into account the low impact on metaScore, it can even be suggested that H1a is not supported. Thus no direct relation was found, it was assumed that the relation could be indirect, adding the variable marketing expenditures as possible direct influence.

The variable sequel is expected to moderate the relation between the independent variable time difference and the dependant variable metaScore, according to H1b. In order to find out if moderation is taking place between these variables of interest, the two independent variables had to be standardized and then multiplied with each other. The result showed that the moderator is not significant (p < .05). This could be explained by the presence of multicollinearity between the variables. Therefore it can be concluded that the variable sequel is not moderating the relationship between the variables metaScore and time difference, hence hypothesis H1b can be rejected.

For the following hypothesis H1c, it was assumed that size of Publisher (Major Publisher or no major Publisher) is mediating the relation between the independent variable time difference and the dependent variable metaScore. Due to the absence of a significant relation between the independent variable time difference and the dependent variable metaScore, size of Publisher was not considered to mediate the above relation, so that H1c is not supported.

The second group of hypotheses was about the relation of the independent variable time difference and the dependent variable userScore (Metacritic score of consumer online reviews). For the first hypothesis, H2a, it was assumed that NPPs are influencing the Metacritic score of consumer online reviews, in other words, the independent variable time difference is expected to be related to the dependent variable userScore. Therefore a linear regression was done with

(30)

30 the two variables of interest. Table 4 shows that there is no significant relation between time difference and userScore (p < .05). Therefore it can be concluded that H2a is not supported.

Table 4: Regression Analysis H2a

Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 7,141 ,077 92,422 ,000

Time Difference (rounded in months)

,003 ,005 ,025 ,631 ,528

a. Dependent Variable: userScore

For Hypothesis H2b it was expected that sequels moderate the relation between the independent variable time difference and the dependent variable userScore. Hypothesis H2b was tested in the same manner as Hypothesis H1b. First the independent variables sequel and time difference got standardized and were multiplied with each other after. As seen in table 5, both models are insignificant. Therefore it can be concluded that the variable sequel is not moderating the relationship between the independent variable time difference and the dependent variable userScore. This result leads to the rejection of hypothesis H2b.

The following hypothesis, H2c, was about the mediating effect of the variable Publisher (Major or no Major Publisher) on the relation between the independent variable time difference and the dependent variable userScore. In this case, the same results occurred as for hypothesis H1c. The relation between time difference and userScore was found to be insignificant. As this basic requirement to test for mediating effects was not fulfilled, hypothesis H2c was rejected.

(31)

31 Table 5: Moderation H2b Model R R Square Adjusted R Square Std. Error of the Estimate Change Statistics R Square Change F Change df1 df2 Sig. F Change 1 ,089a _,008 _,005 _1,2605 _,008 _2,547 ₂ ₆₄₁ _,079 2 ,103b _,011 _,006 _1,2597 _,003 _1,825 ₁ ₆₄₀ _,177

a. Predictors: (Constant), Sequel, Time Difference (rounded in months)

b. Predictors: (Constant), Sequel, Time Difference (rounded in months), moderator1

Block 3 of hypotheses is dealing with the relation of NPPs and their influence on sales, taking into account possible mediating effects of Metacritic scores of expert reviews and consumer online reviews. First, hypothesis H3a was tested. It was suggested that the independent variable time difference is related to the dependent variable global (sales). In order to test the hypothesis, a regression analysis was conducted. Results in table 6 show that there was a significant relation between the two variables of interest (p < .05), which means that hypothesis H3a is supported.

Table 6: Model Summary H3a

Model R R Square Adjusted R Square Std. Error of the Estimate Change Statistics R Square Change F Change df1 df2 Sig. F Change 1 ,187a _,035 _,033 _1,23286 _,035 _23,956 ₁ ₆₆₄ _,000

a. Predictors: (Constant), Time Difference (rounded in months)

Several steps had to be made and basic assumptions to be confirmed to test hypothesis H3b and find out if the variable metaScore, the metacritic score of expert reviews, is mediating the relation between the variables time difference and global. First of all the relationship between the initial independent variable time difference and the dependent variable metaScore had to be confirmed as significant. Results showed that the relation was significant (beta = .187; p < .05). As a next step, the suspected mediator metaScore was added and the regression was

(32)

32 re-ran. If the significant beta between the original independent variable time difference and the dependent variable global changes into an insignificant beta and the beta between the suspected mediator metaScore and the dependent variable global is significant, then metaScore is a mediator. Table 7 shows that mediation is not taking place, as both predictors/independent variables are significant and no change in significance could be determined. Therefore H3b is not supported.

Table 7: Regression analysis: H3b

Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) -2,408 ,218 -11,023 ,000

Time Difference (rounded in months)

,021 ,004 ,162 4,805 ,000

metaScore ,043 ,003 ,487 14,482 ,000

a. Dependent Variable: Global

Last but not least, hypothesis H3c was tested. This time it was controlled for possible mediating effects of Metacritic scores of consumer online reviews on the relationship between NPPs and global sales. As already tested in H3b, the relation between the independent variable time difference and the dependent variable global was significant. Instead of metaScore, now the variable userScore was added to the regression analysis. Again, results showed that both variables, time difference and userScore, were significant and therefore not fulfilling the requirements for mediation to take place. As a consequence, hypothesis H3c was also rejected.

(33)

33

4.4 Additional Analyses

For additional analyses it was interesting to dig deeper into the relationship between NPPs and global sales, as global sales showed several correlations to other variables of the dataset. Moreover it was tested if Metacritic scores of expert reviews and consumer online reviews were related to global sales. The latter was done to investigate if findings differ from other researches, such as the one done by Sacranie (2010), who stated that product quality, alias expert reviews, are related to sales.

Regarding hypotheses H3b and H3c, it was shown that Metacritic scores of expert and consumer online reviews did not mediate the relationship between the independent variable time difference and the dependent variable global. However there are two more variables that may mediate the relationship above, which are sequel and size of the Publisher.

First it was assumed that the variable sequel is mediating the relationship between the independent variable time difference and the dependent variable global. According to results gathered when testing for hypothesis H3a, H3b and H3c, time difference was significantly related to global. This said, another regression was run and again, no mediating effect could be determined. Both variables, time difference and sequel were significant. In order for mediation to take place, the relationship between time difference and global would have to turn insignificant, while the relationship between sequel and global would have to be significant.

Second, it was investigated if the variable MajorP (Major Publisher or no Major Publisher) may mediate the relation between the independent variable time difference and the dependent variable global. However this was not the case, both variables were significant and as a consequence, no mediating effect was found.

In another regression analysis it was investigated for possible relations between Metacritic scores of expert reviews and consumer online reviews and global sales. Therefore two separate regression analyses were run. Results are shown in table 8, whereby Model 1

(34)

34 consists of the relation between metaScore and global and Model 2 consists of userScore and global.

Table 8: Additional analysis, coefficients

Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) -2,164 ,230 -9,398 ,000 metaScore ,043 ,003 ,481 13,640 ,000 2 (Constant) -1,702 ,270 -6,294 ,000 metaScore ,050 ,004 ,564 12,907 ,000 userScore -,140 ,044 -,140 -3,191 ,001

a. Dependent Variable: Global

Both variables, metaScore and userScore are significant. Therefore it may be concluded that product quality does indeed affect the sales of a product. This result further contributes to literature and supports the findings of Sacranie (2010), who claimed that high expert reviews may be a predictor for higher sales. In addition, his research was also done in the videogame industry.

5. Discussion

In the following part, first empirical findings are summarized and being discussed in regard to the literature review and theoretical contributions of this thesis are stated. What follows is a discussion of practical implications of this research. Last but not least, limitations of this study are disputed and possible implications for future research are stated.

5.1 Empirical findings and theoretical contributions

This study explored the effects of timing decisions of new product preannouncements on the product quality and it was further investigated in relations between timing decisions and sales. The empirical setting of this research was the videogame industry. Furthermore, previous

(35)

35 studies were either investigating in effects of timing decisions itself (Lilly & Walters, 1997; Su & Rao, 2007; Eliashberg & Robertson, 1988) or did research on the role of product quality (expert reviews) in regard to other variables, such as sales (Sacranie, 2010; Eliashberg & Shugan, 1997; Park, Lee & Han, 2007; Zhu and Zhang, 2010). Yet no research was done that investigated the relationship between timing decisions, when to announce a new product, and its influences on product quality. With this thesis I tried to bridge the gap and come up with new findings that should contribute to literature theoretically and give practical implications for managers.

Based on the literature discussed, I came up with three basic hypotheses including two sub hypotheses that further explored possible effects and relations of the three main hypotheses. Hypothesis H1 was about the relation between NPPs and the product quality, alias metacritic scores of expert reviews, as well as possible moderating effects of the variable sequel and mediating effects of the variable size of Publisher. For hypothesis H1a a positive and significant relation was expected, although results showed an insignificant one. One possible explanation for this finding may be that other variables are stronger related to product quality, than NPPs are. According to literature, a longer time span between the announcement of a product and its release, allows a firm to use market signals to influence consumers’ perception of the true quality of the product and advertise the product more efficient (Su and Rao, 2010; Lilly and Walters, 1997). Therefore it can be assumed that product quality is strongly related to marketing expenditures and advertising in general and that high financial resources may lead to high product quality. Unfortunately, no information was available about firms’ marketing expenditures or financial resources.

In order to dig deeper into the relationship between NPPs and product quality, hypothesis H1b was testing for possible moderating effects of the variable sequel. The findings showed that the variable sequel did not have any significant moderating effects on the relation

(36)

36 between NPPs and product quality. One possible explanation for this result may be the presence of multicollinearity between the variables of interest, which means that sequel is related to NPPs, as well as to product quality. The latter relation was also confirmed by Situmeang (2013), who stated that sequels offer a stable quality over time.

Regarding hypothesis H1c, it was expected that size of the Publisher, either major Publisher or no major Publisher, may mediate the relation between NPPs and product quality. Due to an insignificant relation between the basic independent variable time difference and the dependent variable metaScore (product quality), which is a basic requirement for mediation to occur, possible mediating effects of the variable Major Publisher could be disclosed. Even though the variable major Publisher is positively and significantly related to both, the independent variable time difference and the dependent variable metaScore, it does not show any mediating effect. Although it can be assumed that the size of the Publisher is of importance concerning the financial support of the developer, be it for marketing expenditures, advertising, etc.

Similar to hypothesis H1, hypothesis H2 also explored the relation between NPPs and product quality, only this time it was about the perceived product quality, alias the metacritic score of consumer online reviews. In addition, the same variables were used when investigating for possible moderating effects of the variable sequel and mediating effects of the variable major Publisher.

Regarding hypothesis H2a, a positive and significant relation between the independent variable time difference and the dependent variable userScore was expected. However, no significant relation was found. As a consequence it can be concluded that customers, be it the buyers or the experts themselves, probably don’t consider time span between a new product announcement and its release date when forming an opinion about the product quality. In other words, neither reviews of experts nor consumers are directly influenced by the amount of time

(37)

37 that passes between product announcement and product release. Thus it is to add that an indirect effect of timespan on the perceived product quality of consumers may occur. According to literature, extra time enables firms to increase the use of market signals to influence consumers’ perception of the true quality of the product (Su and Rao, 2010). Therefore it can be assumed that, if used efficiently, these market signals may influence consumers in their evaluation process, leading to more positive or negative evaluations.

Hypothesis H2a was testing for possible moderating effects of the variable sequel on the relation between the independent variable time difference and the dependent variable userScore. The effects were expected to be positive and significant, although no mediating effects could be found. In this case, compared to H1b, sequel was insignificantly related to userScore and it showed a negative significant relation to time difference. One possible reason for the lack of moderating effects of the variable sequel may be due to the fact that the quality of sequels is said to be stable over time (Situmeang, 2013). Stability of quality may neither strengthen nor weaken any possible effect on the relation between the independent variable time difference and the dependent variable userScore.

Hypothesis H2c, in the same manner as hypothesis H1c, was investigating in possible mediating effects of the variable major Publisher, alias the size of the Publisher; this time it was about the relation between the independent variable time difference and the dependent variable userScore. Contrary to an expected mediating effect, findings showed that the size of the Publisher is not mediating the relationship between NPPs and the metascore of consumer online reviews. The basic requirements for mediation were not fulfilled, just as it occurred when testing for mediating effects on hypothesis H1c. Here it is to add that the variable major Publisher was insignificantly related to the dependent variable userScore and showed a negative but significant relation to the independent variable time difference. Although no mediating effects could be found, size of the Publisher may be important concerning the decision of a firm

Timing decisions and their influences on product quality : an investigation in the videogame industry