Team recombination within the videogame industry : a good or bad decision to be made?

(1)

Thesis

23-06-2015

Team recombination within the videogame industry: A good or bad decision to be made?

Ruben Houtenbos (10317252)

University of Amsterdam Amsterdam Business School

Executive Programme Management Studies Strategy Track

Supervisor: Dr. F.B. (Frederik) Situmeang Second Supervisor: Dr. C.K. (Claudia) Buengeler

(2)

Never Change a winning Team - Sir Alfred Ernest Ramsey

(3)

Acknowledgement

For my thesis I have had a wonderful supervisor who helped me whenever I got stuck. Besides helping with the content of the thesis, dr. F.B. (Frederik) Situmeang also helped by being so enthusiastic during the whole thesis process. Because his enthusiasm was so enormous, it also gave me a great feeling about the thesis and getting me more enthusiastic as well. Another thing dr. F.B.

(Frederik) Situmeang helped me with, is the fact that he provided me a lot of data needed for the analysis and also being structured in meetings and

communication.

As mentioned already, dr. F.B. (Frederik) Situmeang provided me with a lot of data, however this was not enough to complete the analysis of this

research, as some data was still missing. For gathering the data both D. Heijmans and S. Bouchaâla helped me in gathering the other part of the data needed. As I do not have the expertise in gathering the data automated, and this would take a really long time to gather it manually, it has been a great help that this has been gathered by the two fellow colleagues of Brunel IT, therefore I would like to thank them as well.

As last but certainly not least, I would like to thank my family and friends for supporting me the couple of times I got stuck. They have helped me in

brainstorming when I got stuck, but also letting me do the research in first place and planning around the tight schedule where possible.

(5)

Abstract

Previous research has suggested that project teams should be kept together, for a small period of time, thereafter problems will arise. Team recombination could be a possible solution for these problems. Team recombination has been research at project team level, but this thesis extends former research by diving deeper to functional team level. For the purpose of checking if team recombination is a solution, the videogame industry has been used. The sample of 226 videogames has been found as secondary data, which all are sequels and therefore show ‘changes’. The searching distance for new employees has also been checked, to find out if Extra-Team (inside the firm) has a larger impact on team performance then Extra-Organizational (outside the firm). The results have shown that a difference exists, as some functional teams has to be kept together while others should be changed to improve the project team’s

(6)

Introduction

Sir Alfred Ernest Ramsey has had a career from soccer player, to trainer and eventually towards a coach of the national soccer team. In 1955 Alfred Ramsey has started with a team of Ipswich which was a small club. Alfred Ramsey has kept the team together for a long period of time, and therefore the team performed better as the team members knew each other’s strengths and used it to perform better. Within seven years the club had become a top club and played soccer in the highest division. Therefore he made a statement to “Never change a winning team”. This thesis will research the area of team recombination, as every firm in any industry has (project) teams with combined knowledge bases. When

recombining teams, new connected knowledge bases will be created which can extend the seizing of potential opportunities, as these opportunities can be found with the knowledge base and experiences of the ‘new’ added (project) team members. By creating this new and broader knowledge base, new outcomes can be created and former products could be improved in quality, which has been defined by Sethi (2000) as “the superiority of the product in dimensions such as appearance,

performance, workmanship, and life/durability”. A research has been done using data from the dynamic market of Videogames. This industry has been used for analysis as it changes a lot due to the further improvement of technology. The current literature have searched the team

(7)

recombination of teams has not been mentioned much. Especially

recombination of project teams consisting of several sub-teams. Therefore this thesis is extending our knowledge about teams, as the focus is set to sub-teams within a project team.

Definitions used within this thesis

(Project) team has been defined as “A team is a distinguishable set of two or more individuals who interact interdependently and adaptively to achieve specified, shared, and valued objectives” adopted from Morgan, Glickman, Woodard, Blaiwes and Salas (1993). In this thesis the project team consists of members of three functional teams (Commercial, Artistic and Technical operation teams).

Team recombination has been defined as “creating a new team by changing team members from the (project) team by employees from in- and outside the firm and thereby changing the composition of the team and the knowledge bases of these teams”.

Not Invented Here syndrome “The tendency of a project group of stable composition to believe it possesses a monopoly of knowledge of its field, which lead it to reject new ideas from outsiders to the likely

detriment of its performance” adopted from Katz and Allen (1982) Too Good Friends problem “Too-Good Friends which was

characterized by too much harmony. The parties had so much regard for each other, and were so complacent in their relationship, that they did not challenge each other's judgments” adopted from Souder (1980, 1988)

(8)

Commercial team is the team which consists of employees who are fulfilling the function of Producers. The main role of the commercial team is to get the highest sales numbers possible.

Artistic team is the team which consists of employees who are fulfilling the function of Designers. The main role of the artistic team is to get the best artistic value.

Technical operation team is the team which consists of employees who are fulfilling the function of Programmers. The main role of the

technical operation team is to get the best performance of the videogame.

Outline thesis

In this section an introduction has been described about the topic of this thesis, as well as definitions used in this thesis. The following sections consists of (1) the literature review which is the basis for writing this thesis, (2) the hypothesis tested in this thesis, (3) the research design/methods used for analysis, (4) the results and findings of the analysis, (5) the discussion, limitations and future research, (6) the conclusion of the thesis and (7) the reference list used for this thesis.

(9)

1. Literature Review

Teams are made within a firm and teams could cooperate between firms, this thesis will evaluate the team composition and team changes. Both project teams and functional teams can be distinct. Some aspects of team recombination has been researched already. For example aspects like routines (Bresman; 2013), rotating leadership (Davis and Eisenhardt; 2011) and knowledge search (Miller, Fern and Cardinal; 2007) has already been researched. However team recombination in the way that changes occur in the composition of the functional team and therefore the human resources, of the project team as well, has not been researched yet. It has especially not been researched whether these changes will be positively or negatively related to the performance of the project team.

The article of Huckman, Staats and Upton (2009) describes a reason why a team should be kept together, which is different than routines, rotating leadership and knowledge. The reason is the fact that team members will get familiar with each other and therefore will improve performance of the project team. Team familiarity has been described as ‘the degree to which team members have worked with one another in the past (Huckman, Staats and Upton; 2009). As project teams are not stable over time, experience of team members will change as well. It has also been mentioned that team members will not gain the same experience as everybody learns something else (beside the common lessons learned of

(10)

the project). Therefore Huckman, Staats and Upton (2009) controlled the research in the fact that a difference has been made in the role

experience, which is described as ‘the experience that individuals have operating in their current roles of either team member or team manager’. The findings of the article gave a clear view that team familiarity and role experience are positively related to project team performance, where the manager role experience has not been supported. This would therefore suggest that project team members should be kept together for a longer period of time.

Other articles have a slightly different perspective then Huckman, Staats and Upton (2009). For example Katz and Allen (1982), which has also been mentioned in the article by Huckman, Staats and Upton, explain that keeping the team together will only be beneficial for a small period of time, but then the Not Invented Here syndrome will diminish or even make the relationship towards the project team performance negatively. The article by Souder (1980, 1988) also explain that keeping the team together is beneficial for a small period of time, as the Too Good Friends problem will make the relationship towards the project team performance negatively. Therefore it can be said that problems will arise when team composition has not be recombined to form a new team.

(11)

Problems when not recombining the (project) team composition Team recombination could be of a huge impact as has been

described by Katz and Allen (1982). The article of Katz and Allen (1982) describes the Not Invented Here (NIH) syndrome and defined it as “the tendency of a group of stable composition to believe it possesses a

monopoly of knowledge of its field, which lead it to reject new ideas from outsiders to the likely detriment of its performance”. Therefore the NIH syndrome can be explained as a result of the knowledge process / flows within and from outside the project team. Within the project team, the knowledge will first be boosted as all experiences and knowledge from the project team members will be combined and lead to synergies and a more extended knowledge base. This also means that project team members will be more likely to follow a learning sequence as described by Bingham and Davis (2012). However, as the NIH syndrome explains, after a period of time the project team members believe that they possess all knowledge available in their field and reject potential knowledge from outside the project team. This does also mean that new employees from outside the project team could have great ideas but will still be rejected because of the ‘power’ of the current project team members. As an addition to the problem described by Katz and Allen (1982), another problem than the rejection of ‘new’ knowledge from employees outside the project team will exist when the project team members stay together for a long period. Souder (1980, 1988), has found this problem which will lead to the problem of not seizing potential opportunities. The problem as described

(12)

by Souder (1980,1988) is about the project team members’ social

relationship among each other. When the social relationships have been created, this would lead to a better project team performance. This is because of the ties a developer has (Claussen, Falck and Grohsjeanwhen; 2011), which will influence the revenues and critics’ score positively and this same network would also provide advantages over competitors if the network is strong as discussed by Shankar and Bayus (2003). However this social relationship will become a problem once a period of time has passed. When the social relationship has become strong, the feeling about each other’s expertise among the project team members will be strong as well. In the article made by Souder (1980, 1988) the problem has been described as the Too Good Friends (TGF) problem. Souder (1980, 1988) has explained this problem as “characterized by too much harmony. The parties had so much regard for each other, and were so complacent in their relationship, that they did not challenge each other's judgments”. Therefore the TGF problem will arise when project team members will become too much attached in their social network within the other project team members and will no longer discuss each other’s proceedings as much as needed. By not having discussions and conflicts, potential

opportunities will not be sensed and therefore also not be seized (O’Reilly and Tushman; 2008).

By combining the article of Katz and Allen (1982) and Souder (1980, 1988) it can be argued that a project team will lose the potential benefit of a better knowledge base when both problems are in place. When these

(13)

problems are in place (NIH syndrome and the TGF problem), then

potential opportunities for a larger knowledge base will be rejected from employees outside the project team and knowledge inside the project team will not be used to its optimum. The problems as discussed by Katz and Allen (1982) and Souder (1980,1988) will arise when the project team members stay together for a longer period and therefore will believe that they possess all knowledge possible and will become too good friends because of the social network which has been established after a small period of time.

Possible solution by using knowledge search

To solve (or at least diminish) the aforementioned problems, changes should be made within the project team. The current team members cause the problems, because they believe that nobody outside the team can contribute to the current knowledge of the team and

because they do not challenge each other anymore. Therefore in this thesis it has been argued that the team members has to be changed which leads to the team recombination. To find new team members , a knowledge search (Miller, Fern and Cardinal; 2007) could be used, to make sure that new potential opportunities could be sensed and seized (O’reilly and Tushman; 2008). As has been stated by Miller, Fern and Cardinal (2007) the recombination of teams is not limited only to the firm because the organization can also benefit from distant external

(14)

search is distinct into local and distant knowledge search. The local

knowledge search can be explained as a reshuffling of current knowledge from inside the firm, which can be split into “intra divisional” and “extra divisional”. The distant knowledge search can be explained as a searching of knowledge which is totally new for the firm and therefore comes from outside the firm and has been described as the “extra-organizational”. Miller, Fern and Cardinal (2007) has explained their findings on the basis of divisions, however it can be argued that their findings are also possible outcomes for a lower level of teams. When this concept has been taken to this lower level of teams, it suggests that recombining (project) team members should boost the knowledge base of the (project) team. In this thesis the research level is set to sub-teams, therefore the definition is slightly different. Intra-divisional has been used in terms of ‘intra-team’, which means that knowledge has been searched within the same team. Intra-team specifies that an employee was already part of the project team of the former product development cycle and therefore the last videogame made in the videogame series. As this thesis is about team recombination and therefore ‘changes’, the ‘intra-team’ members of the project team can be seen as employees which have not been changed from the former project team to the new project team. Extra-divisional has been used in terms of ‘extra-team’, which means that knowledge has been searched outside the project team, but at the same time within the same firm. When comparing these two aspects ‘inside the firm’ then it has been argued by Miller, Fern and Cardinal (2007) that the ‘extra-team’ is

(15)

slightly more distance then the ‘intra-team’ knowledge search. From the distant knowledge search perspective, the only possible view has been determined as ‘extra-organizational’. ‘Extra-organizational’ knowledge has been termed the same as Miller, Fern and Cardinal’s (2007) explanation which means that knowledge search will be done outside the firm and therefore is the most distant search. As ‘extra-organizational’ knowledge search will lead to knowledge found outside the firm, it could lead to “path-breaking innovations” and therefore is interesting for the R&D

department. When a project team is only looking for knowledge within the current team, it will not benefit from the best possible knowledge base. This is because knowledge which has been searched at a further distance could lead to knowledge which will be less likely to be the same as current knowledge within the current project team composition. As a result of searching for ‘new’ knowledge, the more distant the knowledge search will be, the larger impact it will have to solve the problems discussed by Katz and Allen (1982) and Souder (1980,1988). At the same time the more distant the knowledge search has been done, the harder it will be to integrate into the current team’s knowledge base. The reason for this would be that the current project team members has already accepted local knowledge, which is the knowledge inside the former project team’s knowledge base (intra-team), but will not be pleased to integrate

knowledge outside the former project team’s knowledge base, but within the same firm (extra-team). As knowledge search from outside the firm (extra-organizational) will have an even further distance then extra-team,

(16)

this will be less likely to be adopted by the project team. Therefore the potential opportunities for a larger knowledge base could be rejected by current project team members for ‘extra-team’ and ‘extra-organizational’ knowledge search. As this opportunity for extending the knowledge base will then not be taken, the project team will not perform in the best possible way and therefore suffer in the project team’s performance as well. The impact from a change in project team will be best when having a knowledge search which is not too local and at the same time not too distant. This relationship would therefore have an inverted U-Shape, as has been stated by Miller, Fern and Cardinal (2007) that it is “sufficient distance, not greatest distance that matters”.

Another article, which has been written by Bingham and Davis (2012) explains the learning sequence, which could be seen as an

extension to the article of Miller, Fern and Cardinal (2007). Within a firm learning sequences can be stimulated, when this is in place, then it could lead to a less distant knowledge search and therefore more local

knowledge search as all employees are learning and thereby knowledge from outside the firm could be possessed by employees inside the firm. The reason for this is that recombination inside the firm (both intra- and extra-team) would be more beneficial because knowledge which wasn’t available first has now become available inside the firm which will be adapted more easily then knowledge from outside the firm (extra-organizational). The better adaptation can be explained as the ‘new’

(17)

an extension of former knowledge. This will be better integrated than knowledge which is totally different from current knowledge as the ‘extra-organizational’ knowledge search would be. All the aforementioned articles discuss characteristics of teams, as knowledge, relationships among team members, the NIH and TGF problems occur in teams, etc. The way a team could perform best, is to establish a good relationship among the team members as different participants with different strengths and weaknesses are part of the team. For example, when having relationships among

teams which all contribute to a particular new product, then it would be better to have collaboration established between the teams as each team has a specific specialism. By having many different participants with different backgrounds, education level, interest, knowledge bases, etc. more innovativeness would be possible as has been explained by Davis and Eisenhardt (2011) as the open and closed innovation.

Taken together, two problems has been found by Katz and Allen (1982) and Souder (1980, 1988) which has a large impact on the

performance of the (project) team. One possible solution to solve (or at least diminish) these problems, has been described by Miller, Fern and Cardinal (2007) as the knowledge search. The reason for this is that the problems will occur when (project) team members has been kept together for a long period. By staying together for a long period, the (project) team members will believe that they possess all possible knowledge in their field. At the same time, during the same period, the (project) team

(18)

members have established a social relationship and therefore will no longer challenge each other’s judgments, arguments and decisions.

Research gap

After reading the aforementioned articles, the idea of exploring team recombination has been found. The start has been the article of Katz and Allen (1982). As the article argues, teams staying together for a long period is good, but eventually will fail in capturing the best possible value. However, it has not yet been tested whether the argumentation

mentioned in Katz and Allen’s article has truly lead to a better

performance after changing members of a given team. Other articles which cite Katz and Allen’s article have mostly discussed the

recombination of routines and knowledge. However changing team members within the team is slightly different, as has been discussed in the article of Galbraith (1971) about the matrix-organization. Galbraith’s article suggests that two sorts of separation exists within a firm, which are the project teams and the functional teams. The functional teams have to deal with routines, while knowledge will be quite stable within that given team. But recombination in project teams will lead to less routines as the employees will be part of a team for a ‘short time’. At the same time the knowledge base of the project team will be specifically needed to

accomplish the project and therefore will include knowledge which is gathered from over the boundary of a functional separated structure.

(19)

As project teams will be made for many purposes in almost all industries, this area would be of a huge interest for practical reasons as business would like to have the best (overall) performance. By testing whether team recombination influence the team performance or not, it could lead to more evidence for Katz and Allen’s NIH syndrome and the TGF problem by Souder. Huckman, Staats and Upon (2009) also cite the article by Katz and Allen (1982), where it has been suggested that teams should stay together as the strength lies in the familiarity of team members to each other. Huckman, Staats and Upon (2009) only mentioned that the

performance is measured at the level of the project team and individuals, but it has not been mentioned that the project team exists of multiple teams as explained by the matrix organization by Galbraith (1971) and maybe different results could be found at this deeper level. This gap will therefore be researched in this thesis. Huckman, Staats and Upon (2009), Katz and Allen (1982) and Souder (1980, 1988) all describe aspects of keeping teams together, however it should also be tested if changing team composition is good as well or even better than keeping the team together. This other gap will also be researched in this thesis, by using the distance knowledge search article of Miller, Fern and Cardinal (2007), to control if employees hired from outside the firm have a larger impact on performance than employees hired inside the firm. As researched by Huckman, Staats and Upon (2009), the software industry is an industry which best matches investigation at team level. As this thesis focuses on functional teams which are part of the project team and the fact that the

(20)

data had to be available publicly, the research has been executed in the videogame industry. As mentioned by Brandenburger and Stuart (1996) team performance can be measured on the basis of willingness to pay. This can be translated towards the critics’ score of consumers who already bought the videogame and expressing their feelings about it. Another important performance measurement can be seen in the sales in numbers of copies sold. If the videogame is not sold, then the team didn’t perform well (at least the marketing has done not so well). This data is also

available for the videogame industry.

Therefore in the context of the videogame industry and the questions emerged by combining previous literature, the following research question has been made:

“How does team recombination relate to team performance expressed in sales of product and critics’ score?”

To investigate this question the following figure has been made:

============================================== Insert Figure 1 around here

============================================== In sum team performance has been researched extensively while team recombination has not. Huckman, Staats and Upon (2009) has discussed that keeping a team together would be good for performance because of

(21)

team familiarity. However the test of checking if team familiarity boost performance has only been executed at the level of the project team. This thesis will therefore dive deeper and check if team recombination of

functional team members has an impact on team performance. Beside the fact that familiarity has been researched, Katz and Allen (1982) and

Souder (1980, 1988) has described problems which will arise when a team stays together for a longer period. These three articles all describe the ‘keeping team composition the same is good’ story, however it could be argued from the other perspective as well, where ‘changing team

composition is good’. Katz and Allen (1982) saw this already and mentioned that at some point it would be good to change, where this point is the time that the NIH syndrome has occurred.

(22)

2. Hypothesis

In the former sections it has been argued why this thesis has been written and the research question has been described as well. The hypotheses has been made based upon the research question and literature and will be described in this section. As it has been suggested in articles about

management teams, this thesis will follow the research about the impact of changes in a teams. To contribute to former literature, it will be

analyzed whether the project team will continue to have the impact on team performance as expected, when the team has been recombined. The reason for team recombination is to reconfigure the current knowledge resources available within the organization and outside the organization. Recombination of the team composition will also lead to a recombination / reconfiguration of knowledge, which can both be found within and outside the organization as described by Miller, Fern and Cardinal (2007). New knowledge from outside the (project) team could challenge the

conventional, thereby extend the team’s perspectives, and thus lead to higher performance because of a better decision making process. The decision making process will influence the performance, as the final decision has been well thought of. A decision made, can both be bad and good, even while the decision making process is executed very well.

However a better decision making process will most likely lead to a better decision and this process execution can be influenced. As Schwenk (1984) has argued, the decision making process consists of goal formulation,

(23)

alternative generation and evaluation and selection of the most favorable alternative. Eisenhardt (1999) adds the dynamism dimension by arguing that safeguards are available for dynamic markets to create better

decisions. These safeguards are building collective intuition, stimulating quick conflict, maintaining disciplined pace and defusing political behavior. When combining these two articles, it can be concluded that the decision making process would be better if the decisions made, are made by a team instead of one employee, because of the discussion and conflict opportunities. As the decision will be made by the project team, the influences of this team on the project team’s performance will be larger than a decision made by an individual. Therefore as discussed in the introduction the (project) team has to be changed if the NIH syndrome and / or the TGF problem as described by Katz and Allen (1982) and Souder (1980,1988) respectively, has to be solved. When the project team changes, new experiences and thereby knowledge will become part of the project team which would solve the NIH syndrome. At the same time the TGF problem within the project team will become less as the project team members will become challenged to think, out-of-the-box / out of their comfort zone, again. Both of these aspects will have a positive impact on the project team’s performance, as creativity will be stimulated and therefore boost performance because new possibilities and former strengths can be combined. As the two problems described by Katz and Allen (1982) and Souder (1980, 1988) will occur when no changes are made within the team composition, it has been argued that both sales in

(24)

numbers of copies sold and critics’ score (by experts) will be lower if the project team has remained unchanged. While a team which has been changed would lead to better sales in numbers of copies sold and a higher critic’s score. However at the other extreme, when everybody within the project team has been changed, than knowledge and experience found by the former project team will not be used. This will result in a lower sales in numbers of copies sold and critics’ score, because the learning curve has not been used from the former project and therefore a complete ‘restart’ for the product development will be needed, which will include all pitfalls again as well.

To have team recombination, new employees from outside the current project team should be found. Miller, Fern and Cardinal (2007) has discussed the knowledge search based on divisions, when taking the concept to teams, it has been suggested that employees could be found extra-team and extra-organizational. Therefore a difference between local and distance knowledge search exists. As have been argued, “it is not about the most distance, but the most sufficient distance”. Three different types of knowledge search exists for the team level as well. The types are distint from each other as the knowledge can be found at different

distances. Knowledge is found in the former project team’s composition (intra-team), outside the former project team but within the same firm (extra-team) and outside the firm (extra-organizational). path, Fern and Cardinal (2007) also mention that knowledge found at a larger distant will

(25)

be more expensive and harder to integrate, due to resistance of the current project team members. At the other hand, employees from another market will have the strength of not being influenced by the organization’s routines and therefore will be beneficial for the current project team’s knowledge base. It has also been said that knowledge found at a larger distant will be less likely to be the same as the current knowledge and therefore would lead to newness which can be translated to innovativeness in the context of new product development (Olson, Walker and Ruekert; 1995). At the same time, when knowledge is found locally, then the innovativeness will be small and the added value of the ‘new’ project member will therefore be small as well. The relationship, as explained above, exists between a change in the team composition and the quality and sales of the videogame, where the strength of this relationship is determined by the distance of knowledge search. As too few or too many changes in team composition will both not be the best possible solution for gathering new knowledge, it would be best to have a sufficient distance and therefore would be somewhere in between.

As already mentioned, the project team aspect has already been researched by Huckman, Staats and Upon (2009), but it has not been researched if team members which are part of different functional teams have different impacts on the project team’s performance. All employees would like to improve a different aspect of the overall objective of the project, because the functional teams that they are part of are only

(26)

responsible for a part of the objective. At the same time all functional teams have the same objective, which is the creation of the best possible videogame with the best critics’ score and highest number of copies sold. As discussed by Bhansing, Leender and Wijnberg (2012), dual leadership could lead to conflicts, but has the main advantage of complementarity as every leader has an own specialism. The three functional teams discussed in this thesis also have different specialisms and therefore it could be argued that the impact of each functional team can differ to each other. By having combinations of these three functional teams within one project team, multiple specialisms has been put together, which collectively

constitutes the project team’s knowledge base. By combining the functional specialisms in a project organization, an overall boost in the project’s performance should occur, as has been said in the matrix

organization written by Galbraith (1971). However when each function is analyzed individually, this could lead to different outcomes because the focus of each functional team differs, therefore each team will be

discussed individually as well.

Combined as a project team

All three functional teams have different focusses, responsibilities,

budgets and so on and so forth. As the three teams have to come up with one new product by using the new product development process (Veyzer; 1998), conflicts will be common as explained by the iron triangle of

(27)

============================================== This triangle gives an overview of aspects which will affect each other and thereby the performance of the project team. The overall quality of the product will lead to the rise or fall of the critics’ score, which is the first measurement of the project team’s performance used in this thesis. The other measurement can be expressed as discussed by economics, which is the sales of the product, and can be seen as the number of copies sold (expressed in millions). Quality and ‘Money’ are two of the three aspects of the iron triangle. It can therefore be argued that both sales and critics’ score are somehow related to each other. As the triangle suggests, if one part changes it will have an impact on the other aspects. This also shows the problem for the project team, which consist of different functional teams with different responsibilities and thereby different tasks and influences on the project team’s performance. Each functional team will now be discussed which also will be linked towards the sales in numbers of copies sold and the critics’ score. This connection has been made, to determine if changes in the functional team have an impact on the two measurements, which together determine the team’s performance.

(28)

Commercial team (Producers)

The commercial team, which consists of all producers of the specific videogame, is responsible for the expenses and revenues and therefore the sales in numbers of the videogames. Research in economics has suggested that a firm has to be sure to have a strategy which can be exploited by the firm. Porter (1980) has described generic strategies, which can be a determinant of the firm’s performance. For example, when a firm is willing to minimize the costs as much as possible, then it would have a cost leadership strategy, while a firm which is willing to have high quality would exploit the strategy of differentiation. To control for the aspect of lowest costs possible and highest sales possible, the commercial team will make use of a cost-benefit analysis. The commercial team will try to influence these two aspects to make sure that the performance as measured in the sales of the videogame become as high as possible. When commercial team members have been part of the project team which has staid to be a team for a longer period, then this project team will be performing worse than when changes in the team composition has been made to overcome the NIH syndrome and the TGF problem. The reason for this lies in the explanation of the NIH syndrome and the TGF problem, as mentioned before. If the commercial team does not change, then it will reject potential beneficial knowledge from outside the team (NIH). The other aspect will be the awareness of the other teams’ specialism (artistic and technical operational team) and social network

(29)

established with the other project team members (TGF). This would lead to not challenging the conventional of the other teams and therefore not influencing the other teams’ performance.

Changes in the commercial team (Producers) composition

When the composition of the commercial team would be changed, then new knowledge and experiences would be taken inside the project team, which would overcome the NIH syndrome and the TGF problem. The market opportunities will be better sensed and seized by the new

composition of the commercial team, because the former team has not influenced the new team member(s) and therefore challenge the new team members and become challenged by them too. Because these new project team members are still aware of the responsibilities and will try to do the best, lower costs and/or higher sales will have a higher probability. When combining the argumentation for the commercial team which has no changes in the commercial team composition and the argumentation for the commercial team which has changes in the team composition, the following hypothesis has been made:

Hypothesis 1A: Changes in the commercial team composition is positively related to the sales of the videogame sequel.

(30)

Artistic team (Designers and Artists)

The artistic team is responsible for the (artistic) value of the product, which can be expressed as the quality of the product. To improve quality, time and costs has to be changed as discussed in the iron triangle of project management. Therefore it can be argued that the artistic team is willing to sacrifice some economic value to improve the artistic value of the product, which is expressed in quality. This reasoning has been

discussed by Corrigan and Corrigan (2012) and has been described as the ‘tension between art and commerce’ in any creative industry. The quality of the product will be measured by the value appointed to the product, which has been expressed as the critics’ score because the consumers create the willingness to pay which describes the added value made by the firm (Brandenburger and Stuart; 1996). Therefore the second measurement of the (project) team’s performance is measured by the consumers of the product as it will fulfill their needs. The consumers’ evaluation of the (artistic) value of the product has been measured by critics’ score as discussed by Eliashberg and Shugan (1997).

Changes in the artistic team (Designers) composition

Just as changes of the commercial team, the new artistic team members has not been influenced and therefore would be able to achieve a higher quality of the product. As new knowledge and experiences has been taken inside the artistic team, an increase of awareness of quality will emerge.

(31)

Beside the fact that a new team member has new knowledge and

experience, a change could also lead to more creativity as an artist has an ‘own style’ and therefore could create a more innovative product.

To be innovate, designers has to be creative and therefore needs to be controlled for their outcome instead of their behavior (Eisenhardt; 1989). As the new artistic team member would be more aware of quality and potential (artistic) value within the market, a better quality should be achieved when new team members are found outside the organization. However at the same time as quality rises, the costs involved in achieving this higher quality will go up as well. A reason for this could be the costs involved in training designers who need to upgrade their quality level by using (for example) new technological improvements which are coming with costs. As the commercial team is also involved in the project team, they would control for the costs made for these trainings and searching costs for new opportunities and when benefits are larger than the costs, then these improvements in quality can be made. When the quality of the videogame will become higher than former videogames, than the

likelihood of higher sales will emerge as well. Therefore indirectly the number of sales will be influenced by the artistic team as well. The more obvious aspect of the artistic team’s responsibility is the quality of the videogame (in the design perspective), therefore the critics’ score will be influenced by changes in the artistic team. As has been argued as the changes of the commercial team, when changes in team composition are too few or too large, than the impact on team performance will be low.

(32)

Therefore as like the changes in the commercial team, the same can be argued for the artistic team’s changes, which means that a change in the team composition will have a larger impact when at least some changes has been made in the team composition and not everybody has been changed. By combining the argumentation for a team which has no changes in the artistic team composition and the argumentation for the artistic team which has changes in the team composition, the following hypothesis has been made:

Hypothesis 2A: Changes in the artistic team composition is positively related to the quality of the videogame sequel.

Hypothesis 2B: Changes in the artistic team composition is mediated by the quality change of the videogame and is positively related to the sales of the videogame sequel.

(33)

Technical Operation team (Programmers)

The technical operation team is responsible for the working of the videogame, which is the written code behind the videogame. Together with the designers the videogame will be created, therefore these teams together can be seen as the Research and Development (R&D)

department. Indirectly the technical operation team will influence all

aspects of the iron triangle. For example, by investing (costs) in R&D, new products will become available for the longer term because of more

innovativeness (time), which will result in a product with higher quality. The costs involved in being innovative will be higher than not being innovative, however as the quality of the product will go up as well, it cannot be said easily if the profit will be high. As the project team also consists of project team members from the commercial team, costs will be kept under control, even though the quality of the product will not be much improved when no changes has been made in the team

composition.

Changes in the technical operation team (Programmers) composition Just like the changes in the commercial and artistic teams, when team members of the technical operation team has been changed, then new knowledge and experiences will be taken into the project team which has not been influenced. The technical operation team is related to the artistic team because the designers and programmers teams are both teams who

(34)

have responsibilities for the quality of the product. Product innovation as discussed by Lau, Tang and Yam (2010), has a positive boost on the quality of the product and therefore the critics’ score. To be innovative as explained by Miller, Fern and Cardinal (2007), knowledge from outside can lead to path breaking innovations. Therefore changes in the team

composition would be needed. However the innovativeness of

programmers should not be too high as the former gameplay should be seen in the new game, but then improved where possible. Therefore it could be argued that technical operation team members are both needed from in and outside the former, to keep the former way but add new stuff as well. Because the market of technology is changing rapidly, the need for training is high as well. This will come with costs which will be

controlled by the commercial team. The cost-benefit analysis will turn out to be positive because the outcome (product developed) will be

qualitatively higher than without using the potential possibilities made available by the changing environment. Therefore quality of the

videogame (expressed in the performance as playability of the videogame, like high speed and low numbers of bugs in the videogame) will be

improved, which will thereafter (indirectly) improve the number of sales of the videogame. Therefore the programmers improve the quality of the videogame differently than the artistic team, because the focus of designers to improve the videogame is related to lay-out instead of playability. By combining the argumentation for a team which has no changes in the technical operation team composition and the

(35)

argumentation for the technical operation team which has changes in the team composition, the following hypothesis has been made:

Hypothesis 3A: Changes in the technical operation team composition is positively related to the quality of the videogame sequel.

Hypothesis 3B: Changes in the technical operation team composition is mediated by the quality change of the videogame and is positively related to the sales of the videogame sequel.

Figure 3 shows an integration of the hypotheses mentioned in this section and the conceptual model as shown in figure 1.

============================================== As the figure shows, relationships between the change of the functional team composition and the team performance has been made, where the strength of this relationship will depend on the distance of knowledge search, as described by Miller, Fern and Cardinal (2007).

(36)

3. Research design/methods

In this chapter the dataset which has been gathered using internet, will be discussed. First the reason why choosing secondary data will be discussed, then the data gathering itself, which also includes the population and

sample size used during this research. Thereafter the variables will be explained, which includes the measurement scales used for testing the hypothesis. As last the procedure for using the multiple datasets and therefore calculations, comparisons and so on and so forth, will be

explained as the data gathered from websites couldn’t be used completely as some information was not useful for the purpose of this thesis.

Reason for choosing secondary data

For the purpose of this research and the limited time, the data gathered is secondary data. Secondary data has some advantages and some

disadvantages, as explained by Saunders and Lewis (2012). The main advantages, which have been the reasons for choosing the secondary data are; “the availability of the data, the format of the data and the lesser time needed to gather the data”. Some of the disadvantages are; “data only meets the needs partially, data are not always value-neutral, can be costly to obtain”, as mentioned in the book by Saunders and Lewis

(2012). The disadvantages mentioned are not so much a pitfall for the purpose of this research for a couple of reasons. First the disadvantage of “data doesn’t meet the needs completely”. For a lot of data already

(37)

useful variables (which will be explained in the next section). Beside the fact that using some data from internet was helpful in creating a new variable, it has also been the only way to execute this research. For

example, the developers mentioned at Mobygames.com has been counted to a number of ‘1305853’. When this data was not available at the website and was needed for the purpose of the research, it would have taken a much longer period of time to gather it by surveys, while the response rate would also be lower than 100%. This also explains the advantage of less time spent in gathering the data. As second disadvantage, the data could not be “value-neutral” as for example the scores an expert or

consumer gives to a videogame could be based on frustration of not being able to complete a videogame. However as the data is gathered from Metacritics.com, which by itself combines data from multiple websites and then creates one score on a scale, it is more likely that the scores are value-neutral. As last the data is obtained from websites, therefore the data gathering can be automated and this would cost time to create a tool to do this. For the advantage part, some data was already available as it has been gathered for other research purposes. This data has been gathered from three websites; Mobygames.com, Metacritic.com and Vgchartz.com. These websites have information about videogames, the team members participating in developing the videogame, review scores (both expert and consumer) and sales in numbers. As internet provides a lot of information, new data can be added to a former dataset which has been used for other purposes. The format of the data which was already

(38)

available is an access database file. However this data by itself was not sufficient to test the hypothesis or answering the research question. At least two sources were needed to make the data useful, the information which needs to be gathered is about videogame series and the firm a developer worked for.

Data gathering

The supervisor already had some data available which has been gathered for other researches. However as this data was not sufficient enough for the purpose of this research, new data was added which has been

gathered from Mobygames.com. The new data which has been gathered provides information about the GameSeries and its GameSequels (one GameSeries has multiple GameSequels). The other information which was needed for the purpose of this research is data about the developers and the firms they worked for. By having these two ‘extra’ datasets, the changes can be measured within each GameSeries, so between

GameSequels, and the distance of knowledge search can be determined, by checking at which firm a developer worked at, before participating in the current project team. This data was gathered by colleagues of Brunel IT, who used ASP.NET for gathering the information from

Mobygames.com. The data that was gathered has been gathered from the website using the search functionality with the term “Series”, which

(39)

Even the gathered data did not create a complete dataset without doing any adaptations to the gathered data. Therefore some calculations has been made to make sure that the data which has been gathered creates a new variable called “Distance Knowledge Search”. The steps taken will be described, using some figures to make sure that it becomes clear how the calculations has been made and why these calculations make sense.

Data Modification

As the data gathered by the Brunel IT colleagues was partly used, it has been described how it looked like and how it has been used for calculating the “Distance Knowledge Search” variable.

The first step was to get the data, which has been executed by Brunel IT employees. This data was placed in an Excel document containing 3 tabs.

The first tab (Series_Sequels) contains series and sequels data, which consists of the following columns A. SeriesGameID, B. SequelGameID, C. SeriesGameName, D. SequelGameName and E. Year.

(40)

The second and third tab (Developers_Firms) combined form the total dataset about developers and the firms they worked for at the time a videogame was created. These tabs consists of A. DeveloperID, B.

DeveloperName, C. GameName, D. Year, E. FirmName and F. “External?”.

The second step was calculation. The start has been made in comparing firm names, as this leads to “Extra-Organizational” which means that a check has been made to find out if the employee came from outside the firm or not. First a sorting was done at “DeveloperID” (ascending) + “Year” (descending) to find out whether a Developer changed firms. Then in column F, the following formula has been used:

=IF(AND(B3=B2 E3<>E2;);1;0)

(An example of the table used, is shown above)

This formula therefore checks if the DeveloperName is the same as the former row in the column and the FirmName is different. If this is true, then a ‘1’ has been placed in column “External?”, if not then a ‘0’ has been placed, where 1 means that the employee came from outside the

organization and therefore is part of the “Extra-Organizational” team members group.

(41)

Now that the Extra-Organizational employees are known, the next calculation was performed. This calculation is about finding out whether the developer was part of a former videogame development project team. As the developer tabs do not contain data about series, the second and third tabs had to be combined with the first tab, which has been done in Access. The connection from tab 1 to 2 and 3, was made using the column “GameName” and “SequelGameName”. As all tabs have different columns, the following columns were used: A. SeriesGameID, B. SequelGameID, C. SeriesGameName, D. SequelGameName, E. DeveloperName, F. FirmName and G. “External?” As the connection was made, the data file was

exported to Excel again. Then a new column was inserted as C, where the following formula was used:

=SUM((A2*100)+B2)

(An example of the table used, is shown in figure 4)

The reason for this is that the calculation helps to answer the question: “Was the team member part of the former videogame development project team or not?” The SeriesID was multiplied by 100, as the SequelGameID can be over 10. With the following example it can be shown that a multiplication of 10 wouldn’t be sufficient. For example, if the first series has 12 sequels, then we would have (1*10) + 1 until 12. Which would result in 11 until 22. As every series at least contains 2 sequels, this would lead to following numbers for the second series: (2*10) + 1 = 21 and (2*10) + 2, which is also 22. When the formula as

(42)

shown would have been used, then for the first series it would have been 101 until 112 and for the second series 201 and 202.

After this ‘dummy’ was created, a formula was used to compare if the videogame was a sequel of the former videogame or not. To compare one sequel videogame with another, the Cx has to be one higher than Cx-1. For example C3 has to be 1 higher than C2 to be the sequel of C2. In the case of the example given, the first series would be C2 until C13 (which is 12 sequels with number 101 until 112). C2 would be 101, where C3 would be 102, which is one higher. Another remark to be made is the movement of all other columns, so column C became column D, where D became E, E became F and so on and so forth. Before using the formula, a sorting was executed at “DevelopersName” and “column C” (both ascending).

The formula for Distance Knowledge Search (Column I) is:

=IF(H3=1;"Extra-Organizational";(IF(AND(F3=F2;C3=(C2+1));"Intra-Team";"Extra-Team")))

(An example of the table used, is shown in figure 5)

This formula checks if the ‘External?’ cell is 1, if so, then value = “Extra-Organizational”, else check if DevelopersName (F3) = DevelopersName (F2) and check if the ((SeriesGameID*100)+(SequelGameID+1) = Cx) = ((SeriesGameID*100)+SequelGameID = Cx-1). So this would be a

comparison of Cell X in Row r+1 to Cell X in Row r. If this is true, then “Intra-Team”, else “Extra-Team” is shown.

(43)

Combining Datasets

After the former steps has been executed, the new variable called “Distance Knowledge Search” was made. This variable had to be

connected to the other datasets, which contain data about Sales, Review scores and Positions of developers per videogame (to find out which functional team the developer worked for).

To create the connections, all 5 datasets were imported in Access.

Dataset Information

Mobygames_Series Created for Distance Knowledge Search Mobygames _Data Information about the Videogame

Mobygames _Credits Information about developers of the Videogame Metacritic _Data Information about the Videogame + Review

Score

VGChartz Information about the Videogame + Sales

First for every dataset the ‘GameNames’ were cleaned, to make sure that matching becomes more easily. To clean the names, a function of Access was used and a query which calls this function. The SQL query used is:

UPDATE <TableName>

SET <Column to be update> = Trim(CleanString(<Column to be cleaned>));

(The CleanString is a function, which includes a difficult check. This function is added in Appendix B.)

Then for all the datasets an export was made which contains the

SequelGameName and the SequelGameName_NEW (which is the update column). In Excel fuzzy lookup (an add-in) was used to make sure that videogames which do not match 1-on-1 can still be found. The confidence

(44)

has been placed at 90%. This matching led to one ‘overall matching table’ which contains all ‘cleaned’ SequelGameNames_NEW and matches to the SequelGameNames_NEW of the other datasets. This was then imported to Access again to create the table for connecting all dataset to each other.

As the series information was gathered for the purpose of this research, a link was made to all the other datasets to create new datasets needed for the purpose of this thesis. First a connection was made to the

Mobygames_Credits dataset. This dataset contains the positions a developer fulfilled during the videogame development cycle. These

positions are used to link the developers to the functional team that they worked at. As not all positions were needed for this research, a check has been made to get only some specific positions. The query to get only these specific positions is:

SELECT mobyvideogame_credit.name, mobyvideogame_credit.position, (Switch(position Like "*Design*",1)) AS priority_Design,

(Switch(position Like "*Art*",1)) AS priority_Art,

(Switch(position Like "*Producer*",1)) AS priority_Producer, (Switch(position Like "*Programm*",1)) AS priority_Programmer FROM mobyvideogame_credit

ORDER BY mobyvideogame_credit.position;

Thereafter the columns priority_Art and priority_Design has been merged using Excel, so that the three functional teams as mentioned in the

hypothesis section are separated. As the positions per developer per videogame has now been set to each videogame, it can be connected to

(45)

the Knowledge Distance Search variable. After this dataset was combined with the Knowledge Distance Search variable the table included all the GameNames, Developers, Positions and Knowledge Distance Search. Based upon this information a count could be executed for each team and each knowledge distance search. As three teams exists and three

knowledge distance search exists, nine queries were made to gather all the counts. An example of a query for the Designers team with “Extra-Organizational” as knowledge search distance is:

SELECT Knowledge_Teams.SequelGameName, Count(Knowledge_Teams.Design_Prio) AS COUNT_Design_Prio_Extra_Org

FROM Knowledge_Teams

GROUP BY Knowledge_Teams.SequelGameName, Knowledge_Teams.[Knowledge Distance Search], Knowledge_Teams.Design_Prio

HAVING ((Knowledge_Teams.[Knowledge Distance

Search])="Extra-Organizational")

AND (Knowledge_Teams.Design_Prio = "1");

To create the queries for the other functional teams and Knowledge Distance Search values, only bold text has to be changed. Every “Design_Prio” has to be changed to “Programmer_Prio” or

“Producer_Prio”, the Knowledge Distance Search has to be changed to “Extra-Team” or “Intra-Team” and the name of the column which is

written after AS “COUNT_Design_Prio_Extra_Org” had to be changed to a combination of functional team name and Knowledge Distance Search. After the nine tables were create, a merge was executed and at the same time a connection was created to the datasets of Vgchartz.com and

(46)

Metacritic.com, to get the connection towards the Sales, ExpertScore, Year, Genre Rating and Platform, respectively. An example of the data itself would be:

Measurement and scales

In the conceptual model and hypothesis some variables has already be mentioned, the abovementioned also shows some calculations already. In one list the following variables has been used for analysis:

• Quality of the videogame is the dependent variable and a mediator variable for sales of the videogame. This variable will only be seen as a mediator if changes in team composition of the artistic or technical operation team exist. This variable has been measured on the basis of 100 point scale. As has been explained at the website of Metacritic.com, the score was gathered on the basis of expert

reviews and when at least four experts has published their review, then it will be combined to become one overall average critics’ score. The 100 point scale is used in other articles as well, like Kawada and Yoshimura (2012)

(47)

• Sales of the videogame is the dependent variable which is measured by sales in millions of copies sold.

• Distance of knowledge search is the moderator variable which

therefore gives insight whether the relationship is influenced on the basis of distance used for new knowledge. Two questions will be used: 1. Was the team member already part of team which created former videogame? (intra-team if yes) And 2. Was the team

member already employed at the firm? (team if yes, extra-organizational if no)

• Year is a control variable which has values as years containing four digits.

• Genre is a control variable which has values of genres a videogame has, for example Football, Racing, Alternative sports and First

Person Shooter.

• Rating is a control variable which has values of E, E10+, T and M. These ratings has been based upon the ESRB Rating, which means that E = Everyone, E10+ = Everyone 10+, T = Teen and M = Mature. This therefore creates classes among ‘gamers’.

• Platform is a control variable which has values as Playstation 2, Playstation 3, XBOX, XBOX360 and Wii.

(48)

4. Results

In this chapter the results of this thesis research has been described, which is based upon outputs from SPSS. First the results of the

descriptives is shown, then a full correlation matrix was made to control the relatedness of the variables and last the Amos output is shown which has been used to calculate the relationships mentioned in the hypothesis. After these outputs has been shown, the results of the tests of the

hypothesis mentioned in this thesis has been shown.

Descriptives and Full Correlation Matrix

Based upon the descriptives, it can be argued that Designers (M=12,48 ; 12,21) have been changed most, than Programmers (M=9,66; 11,02), and last the Producers (M=7,43; 7,87). Another interesting result is the fact that in every team not even one change has been found outside the team but inside the firm (Extra-Team). Therefore this variable has been excluded in the table. After creating the descriptives, a correlation matrix was created to check if the variables are positive or negative related to each other. Table 1 shows the results of the integration of the descriptives and correlation matrix. Therefore it shows that the variables are related to each other.

(49)

Table 1

Descriptive Statistics and Correlations matrix

M SD 1 2 3 4 5 6 7 8 9 10 11 1.COUNT_Designer_Extra_Org 11,929 12,345 1 2.COUNT_Designer_Intra_Team 10,912 14,132 ,146* 1 3.COUNT_Producer_Prio_Extra_Org 7,1018 5,1909 ,555** -,004 1 4.COUNT_Producer_Prio_Intra_Team 7,0664 6,4574 -,015 ,484** ,027 1 5.COUNT_Programmer_Prio_Extra_Org 9,0575 8,4439 ,489** ,051 ,578** ,138* 1 6.COUNT_Programmer_Prio_Intra_Team 9,3584 13,255 -,008 ,703** -,070 ,600** ,110 1 7.Sales 1,2545 2,4661 ,057 ,110 -,042 ,185** ,120 ,296** 1 8.ExpertScore 7,5086 1,1714 ,055 ,134* -,090 ,181** ,071 ,201** ,272** 1 9.Sony 0,6814 0,467 -,169* -,072 -,082 -,008 ,008 ,026 ,043 -,034 1 10.M 0,2566 0,4377 -,006 -,078 ,114 ,047 ,109 -,108 ,007 -,021 ,076 1 11.FPS 0,0885 0,2846 -,046 -,052 ,135* ,081 ,029 -,073 -,037 -,014 -,121 ,423** 1 *. Correlation is significant at the 0.05 level (2-tailed).

(50)

Amos Output

The SPSS Add-in Amos has been used to test the impact from each

functional team to the sales of a videogame (Sales) and the critics’ score (ExpertScore) of the videogame. The result of Amos is presented in Table 3. Table 3 AMOS output Estimate P ExpertScore <--- COUNT_Design_Prio_Intra_Team 0.007 0.198 ExpertScore <--- COUNT_Design_Prio_Extra_Org 0.012 0.037 ExpertScore <--- COUNT_Producer_Prio_Intra_Team 0.015 0.187 ExpertScore <--- COUNT_Producer_Prio_Extra_Org -0.015 0.280 ExpertScore <--- COUNT_Programmer_Prio_Intra_Team 0.016 0.003 ExpertScore <--- COUNT_Programmer_Prio_Extra_Org 0.005 0.592 ExpertScore <--- SQUARE_Designer_Intra_Team -0.027 0.012 ExpertScore <--- SQUARE_Designer_Extra_Org -0.012 0.571 ExpertScore <--- SQUARE_Producer_Intra_Team -0.012 0.735 ExpertScore <--- SQUARE_Producer_Extra_Org -0.047 0.001 ExpertScore <--- SQUARE_Programmer_Intra_Team -0.003 0.854 ExpertScore <--- SQUARE_Programmer_Extra_Org 0.020 0.293 Sales <--- ExpertScore 0.423 0.002 Sales <--- COUNT_Design_Prio_Intra_Team -0.029 0.006 Sales <--- COUNT_Design_Prio_Extra_Org 0.039 0.002 Sales <--- COUNT_Producer_Prio_Intra_Team -0.008 0.718 Sales <--- COUNT_Producer_Prio_Extra_Org -0.066 0.023 Sales <--- COUNT_Programmer_Prio_Intra_Team 0.075 *** Sales <--- COUNT_Programmer_Prio_Extra_Org 0.031 0.083 Sales <--- SQUARE_Designer_Intra_Team -0.012 0.582 Sales <--- SQUARE_Designer_Extra_Org -0.070 0.098 Sales <--- SQUARE_Producer_Intra_Team 0.046 0.518 Sales <--- SQUARE_Producer_Extra_Org 0.024 0.799 Sales <--- SQUARE_Programmer_Intra_Team -0.027 0.462 Sales <--- SQUARE_Programmer_Extra_Org -0.036 0.352 Sales <--- Sony (PS2 and PS3) -0.098 0.852 Sales <--- M (Mature) 0.182 0.569 Sales <--- FPS (First Person Shooter) 0.205 0.549

(51)

Hypothesis testing

As has been argued, every relation from a change within a functional team composition towards the team performance (expressed by quality and sales of the videogame), has been hypothesized to be positive. Only one difference has been mentioned between the functional teams. This

difference is in the commercial teams (designers) which has been expected to only impact the Sales and not the ExpertScore, while the artistic (designers) and technical operation team (programmers) would impact the ExpertScore directly and Sales indirectly. For testing if every hypothesis is positively related, the output of Amos has been used. As can be seen, some relations between functional team recombinations are significant, while others are not.

Hypothesis 1A suggest a relationship between team recombination of

the commercial team composition (producers) to the Sales to be positive. The data of Amos shows that the relationship between changes in the commercial team to Sales is negatively related when changes are made (β = -0.066; P<0.05), therefore hypothesis 1A is not supported. At the same time the relationship of keeping producers inside the firm is not significant as do the results of the relation to ExpertScore.

Estimate P

Sales <--- COUNT_Producer_Prio_Intra_Team -0.008 0.718 Sales <--- COUNT_Producer_Prio_Extra_Org -0.066 0.023

(52)

the artistic team composition (designers) to the ExpertScore to be

positive. The data of Amos shows that the relationship between changes in the artistic team to ExpertScore is positively related when changes are made (β = 0.012; P<0.05), therefore hypothesis 2A is supported. At the same time the relationship of keeping designers inside the firm is not significant.

Estimate P

ExpertScore <--- COUNT_Design_Prio_Intra_Team 0.007 0.198 ExpertScore <--- COUNT_Design_Prio_Extra_Org 0.012 0.037

Hypothesis 2B suggest a relationship between team recombination of the

artistic team composition (designers) to Sales, which is mediated by ExpertScore, to be positive. The data of Amos shows that the relationship between changes in the artistic team to Sales is positively related when changes are made (β = 0.039; P<0.005), therefore hypothesis 2B is supported as well. At the same time the relationship of keeping designers inside the firm is also significant but as a negative impact on Sales (β = -0.029; P<0.01).

Estimate P

Sales <--- COUNT_Design_Prio_Intra_Team -0.029 0.006

Sales <--- COUNT_Design_Prio_Extra_Org 0.039 0.002

the technical operation team composition (programmers) to the

ExpertScore to be positive. The data of Amos shows that the relationship between changes in the technical operation team to ExpertScore is

Team recombination within the videogame industry : a good or bad decision to be made?

Thesis

Table of contents

Acknowledgement

Abstract

Introduction

1. Literature Review

Research gap

2. Hypothesis

Combined as a project team

Commercial team (Producers)

Artistic team (Designers and Artists)

Technical Operation team (Programmers)

3. Research design/methods

Reason for choosing secondary data

Data gathering

Data Modification

Combining Datasets

Measurement and scales

4. Results

Descriptives and Full Correlation Matrix

Amos Output

Hypothesis testing