THE EFFECTS OF FORCED CHANNEL MIGRATION IN
THE LIVE ENTERTAINMENT SECTOR
by
HENRIKE BRUNSTING
Rijksuniversiteit Groningen Faculty of Economics and Business
Msc Business Administration
Marketing Management & Marketing Research
May 2010
Van Brakelplein 30a 9726 HE Groningen 0031-(0)6 12244772
2 Title: The Effects of Forced Channel Migration in the Live Entertainment Sector
Student: Henrike Brunsting Student number: 1479520 Qualification: Master thesis
Study program: Msc Business Administration: Marketing Management & Marketing Research
Faculty: Economics and Business University: Rijksuniversiteit Groningen First supervisor: dr. J.E. Wieringa Second supervisor: dr. S. Gensler External supervisor: M. Manders
Organization: Mojo Concerts B.V. a Live Nation Company Completion date: May 2010
Summary: In this research a highly representative sample of the Dutch population is used to study the effects of forced channel migration in the live entertainment sector. The cancellation of ticket sales via outlet channels by Mojo Concerts in January 2009 is used as a case study. The results show that consumers which are confronted with a situation of forced channel migration show a higher reactance level and a lower satisfaction level compared with voluntary migration. Interestingly, this form of channel migration did not lead to a decrease in average demand over the year 2009 compared with the years before.
3
MANAGEMENT SUMMARY
This paper studies the effects of forced channel migration in the live entertainment sector. The elimination of the outlet channel by Mojo Concerts as from January 2009 is used as a case study. Until now, research in this field and especially in this sector was scarce. Forced channel migration is defined as: the elimination of a channel to compel customers to utilize a specific channel or channels in order to enhance the efficiency of the firm’s channel operations. Results show that when consumers are confronted with this type of migration that their reactance level increases and their satisfaction level decreases. Reactance can be defined as the negative feelings associated with being restricted in one’s freedom. A partial mediating effect of reactance on the relationship between forced channel migration and satisfaction is found. In spite of the influence of forced channel migration on the reactance and satisfaction level of the consumer, it did not significantly effect their average demand for live entertainment event tickets. The occurrence of switching behaviour, thus consumers purchasing tickets for an alternative live event upon unavailability of the desired tickets via an outlet, was marginal. For that situation, consumers indicate that they would purchase either no tickets at all or they would purchase them via another distribution channel.
4
PREFACE
Here before you lies the final product of my university education at the Rijksuniversiteit Groningen. At the moment that you are reading this, it is completed. But there was a time, back in august 2008, in which I just started to think about my graduation upon starting my Master in Marketing. Your master thesis research is something with which you can already define your interest, some kind of specialization. Moreover, it is a piece of work with which a lot of students struggle. Therefore, I thought, if I have to struggle it could better be on a to me very interesting research topic. One of my bigger interests, besides travelling, is (live) music. This led me to Mojo Concerts in Delft, one of the largest organisations in the live music business. I felt lucky to have landed a research topic there and on top of that an internship. It gave me three great opportunities (1) it gave me a platform to perform an interesting and relevant research; (2) I had the chance to gain valuable experience in organizing a large event, both from a marketing perspective as from a production perspective; (3) for a period of six months I got the great gift of being able to go to all concerts and festivals which Mojo organizes. A great thanks to Mojo, in particular to Marjanne Manders, Frederik van Alkemade and Joyce Mense. Then, after landing the most interesting research topic, I needed a supervisor who could help me get the most out of this opportunity. This is why I contacted professor Jaap Wieringa, who was my teacher in Advanced Market Research. I am glad to be able to say that some space just opened up in his pool of graduation students and that he saw something in my research proposal. Special thanks to dr. Wieringa for his feedback and constructive conversations. I have sincerely enjoyed working on this thesis and also hated it at times. A thank you to my family and friends for putting up with me during those latter times.
Enjoy!
5 TABLE OF CONTENTS MANAGEMENT SUMMARY ... 3 PREFACE... 4 TABLE OF CONTENTS ... 5 1 INTRODUCTION ... 6 1.1 Initial motive... 6 1.2 Problem statement ... 7
1.3 Structure of the report ... 8
2 MARKET BACKGROUND... 10 3 THEORETICAL FRAMEWORK... 16 3.1 Conceptual framework... 17 3.2 Conceptual model... 24 4 RESEARCH DESIGN ... 26 4.1 Data collection ... 26 4.1.1 Data design ... 26
4.1.2 Scales and measurement levels ... 27
4.1.3 Sample design ... 30
4.2 Data analysis... 32
4.3 Validity and reliability... 34
5 RESULTS... 35
5.1 Representativeness... 35
5.2 Analyses ... 36
6 IMPLICATIONS... 62
7 CONCLUSIONS AND RECOMMENDATIONS ... 66
7.1 Conclusions... 66
7.2 Limitations ... 66
7.3 Recommendations... 67
REFERENCES... 68
APPENDIX... 72
Appendix A: Interview questions... 72
6
1 INTRODUCTION
1.1 Initial motive
By now, extensive research has been performed concerning multichannel customer management (Neslin et al., 2008; Konuş et al., 2008; Neslin et al., 2006). Researchers were able to distinguish segments of consumers considering their information search and purchase behaviour in a multichannel environment. Categories included in this research stream are books, mortgages, electronics, holidays, clothing, computers and insurances. A relatively new field of study is forced channel migration. Many firms have been adding new channels in the last few years. Especially the opportunities of the Internet are attracting many companies. However, the (cost) efficiencies of this channel create the need for firms to eliminate less efficient channels. For instance, in the live entertainment sector where companies are reconsidering the outlets. Two noted companies in this sector have decided to cancel the ticket distribution through this channel, namely Joop van den Ende Theaterproducties BV and Mojo Concerts BV. Large scale research on this topic is nearly non-existent, especially in the live entertainment sector.
7 This cancellation has led to a major shift in the distribution shares of the different channels (see table 1).
TABLE 1
Shares distribution channels
Previous research of Mojo (CheckMarket, 2009) points out that 21% of the customers that bought their concert ticket through the Internet the last time they went to a concert indicate that they are influenced in the number of concerts they visit by the cancellation of the distribution of tickets through outlets. 15% of this group states that they buy concert tickets more often, 51% states that they buy tickets less often. An issue that rose from the communication management from Mojo is if they have actually lost customers all together and, if yes, what is the underlying reason for this loss. Due to the success of the ticket sales of this company, it cannot be concluded if customers are lost from the figures that are present in its database. On average sales have merely increased. However, existing figures do not show how many tickets were requested for a sold out show. Currently Mojo is having one of the best years in the company’s history. Therefore the issue for the communication management lies in the question if they have lost a share of customers of such substance that the company would notice this loss if sales would decrease in the future.
1.2 Problem statement
The shift in distribution shares has lead to the following question from the management of Mojo: Did Mojo lose customers due to the cancellation of the ticket distribution through outlets? And, if yes, what is the underlying reason for this loss?
These recent developments, the lack of existing research in this field in this sector and the question from Mojo are founded reasons for an empirical study.
Distribution channel 2007 2008 2009
Internet 49.20% 54.90% 93.40%
Telephone 7.60% 6.50% 6.60%
8 The research goal of this empirical study can be stated as:
What are the effects of forced channel migration in the live entertainment sector in The Netherlands?
In order to reach the research goal several questions need to be answered. The research questions are the following:
How can the multichannel environment in the live entertainment sector be characterized? What is the consequence of the cancellation of the outlet-channel for ticket sales?
Are the effects of forced channel migration equal for everyone?
The method of working in order to arrive to the answers on these questions is as follows. Konuş et al. were among the first to explore the effects of forced channel migration in their study in 2009. They have performed an experiment among a convenience sample of 117 undergraduate students. Participants in this experiments were assigned a channel via which they usually send the meter-reading of their electricity use (regular mail or e-mail) to a hypothetical company. The conceptual framework proposed in their study is used as a guideline and further developed according to relevant findings in the literature study. For example, the effect of forced channel migration on demand is added to the framework. Besides this, an important addition of this study to the one performed by Konuş et al. (2009) is the use of a highly representative sample of the Dutch population. Moreover, the data collected is based on an actual case of forced channel migration in the live entertainment industry. These additions improve the validity and reliability of the research results considerably.
1.3 Structure of the report
10
2 MARKET BACKGROUND
In this chapter relevant developments in channel usage in the live entertainment sector are discussed. The market background is mainly based on interviews with experts in the field (see appendix A), namely Aukina Buining (ticketing manager Mojo Concerts), Marjanne Manders (communication manager Mojo Concerts), Huib van der Ploeg (director of project management Live Nation Europe), Eugene Westerink (managing director CTS Eventim Nederland), Richard Heideman (marketing De Oosterpoort). Furthermore, figures of the Dutch Central Bureau of Statistics (CBS) are used.
The live entertainment sector in general offers five main channels for purchasing live event tickets, namely outlets, telephone, venue, brochure and the Internet. Some theatre and concert hall combinations use a presale system that is commonly utilized by theatres, that is presale via a brochure that appears at the beginning of each season.
11 costs of the Internet were relatively high because of the bank’s high charges for online banking and the printing and mailing costs of the tickets because at that time e-tickets were not developed yet.
When the contract with Ticket Service ended in 2008 and negotiations did not lead to a new agreement, Live Nation Europe introduced their new ticketing partner: the German company CTS Eventim AG. The costs of Internet tickets decreased because of lower fees for online banking and the development of e-tickets. In January 2009 the outlet channel was eliminated, forcing customers that previously used outlets as an information source (for example, a customer sees a poster in the post office of an upcoming Bob Dylan concert) or purchase channel to use telephone or Internet. This resulted in a better balance in service costs of Internet and telephone ordered tickets. For Mojo the cheaper and more efficient channel was now also cheapest and most efficient for the customer. E-tickets were only raised with ten percent service costs, while telephone ordered tickets were raised with and equal percentage of service costs and an additional handling fee of € 6.50 euro. Moreover the customer pays € 0.45 per minute plus the costs of their telephone provider. In 2009 the telephone was only used in 3.8% of the times a customer bought an event ticket.
Main distribution channels
The live entertainment sector distinguishes five main distribution channels, namely the Internet, telephone, venue, outlets and brochure. Relevant characteristics of each of the channels are discussed below.
Internet
12 7% 3% 9% 10% 15% 28% 32% 53% 0% 10% 20% 30% 40% 50% 60% Other
No computer, internet access Out of habit No experience Not available/does not trust/w orried Security/privacy Wants to see product themselves No interest
Internet. Research by CBS (2009b; figure 1) include figures from 2004 on reasons why consumers have never bought a product or service online.
FIGURE 1
Reasons why consumers have never bought online in 2004
(Source: CBS, 2009b)
These figures give an indication of the considerations of consumers concerning purchasing online, even though it is realistic that the shares of each reason have shifted since 2004 to some extent.
Eugene Westerink -managing director at CTS Eventim Nederland BV- can imagine that some consumers do not trust the Internet for ordering concert tickets of their favourite artist and blame ICT for not obtaining tickets for its concert (personal communication, 15 December 2009). Marjanne Manders –communication manager at Mojo Concerts- explains that in the past the TSN system was not always able to process the large number of simultaneous requests for tickets when a presale started for a popular concert. This also occurred for online orders. Many consumers still remember this, even though today online ordering via CTS is going very well. This could be a reason for consumers to still avoid using the Internet for purchasing tickets.
13 Telephone
Relatively little is known about the consumer that uses the telephone as a mean to order live event tickets. Aukina Buining -ticketing manager at Mojo Concerts- expects that consumers that order via telephone are mainly elderly people or youngsters that do not have access to iDeal or a credit card (personal communication, 14 December 2009). In contrast with general expectations, sales made by telephone did not increase due to the cancellation of outlets but in fact decreased.
Outlets
The outlet network of Ticket Service comprised mainly out of post offices, GWK’s, Free Record Shops, VVV’s and presale via the cash registers of the venues. Westerink argues that many consumers have a certain nostalgic feeling about the idea to sit in line in front of the post office in their sleeping bags hours before presale starts (personal communication, 15 December 2009). He can imagine it gives them a feeling of togetherness. Moreover, due to the failures of the Internet system in the past some consumers are still sceptic about this channel. If consumers are confronted with a sold out concert because they did not timely succeed in ordering tickets via the Internet, consumers may argue that this would not have occurred if the outlet channel had still been available to them.
Brochure
14 Venue
The distribution of live entertainment event tickets via the venue was previously done on basis of presale and sale on the day of the event. The cancellation of the outlet channel also included termination of presale via the venue. At the venue tickets for Mojo events can only be purchased on the day of the event since January 2009. As mentioned previously by Heideman, it is assumed that this affects sales at De Oosterpoort where consumers are still used to picking up their ticket in advance at the cash register. Tickets sold at the day the event takes place are mainly distributed to last minute attendants or to visitors that are not worried to miss out on tickets, for instance in the case of an event of which can be expected that ticket will still be available on the day of the show
The effect of forced channel migration
Experts are divided in their opinions on the effect of forced channel migration in the live entertainment sector in this particular case.
Westerink argues that the cancellation of ticket sales through outlets can lead to frustration by consumers that were not able to obtain a ticket for a well-desired event. “It can create a sense of powerlessness. Consumers believe that they did not have influence on the situation and feel that they would have obtained tickets if they would have laid in front of the post office the entire night”, Westerink explains. However, he does not believe that the possible frustration as a consequence of the cancellation has resulted in a loss of customers for Mojo.
Moreover, Heideman argues that frustration occurs among people that are not used to ordering online and now feel that they are forced to do so. Heideman believes that because of this, many of these people simply do not buy tickets at all anymore for Mojo shows in De Oosterpoort. Therefore, in the case of De Oosterpoort and other venues where customers are used to ordering via a brochure it is possible that customers retain from buying tickets through the remaining channels.
16
3 THEORETICAL FRAMEWORK
The marketplace is increasingly becoming a multichannel environment for consumers. Firms are adding new channels and this trend of proliferation has enabled firms to now also interact with their customers by means of the Internet, call centers, e-mail and weblogs. Simultaneously, firms are cancelling distribution channels they see as cannibalizing and/or cost inefficient (Ansari et al., 2008). Channel migration is only justified if it is beneficial to the company’s revenues. Channel migration is a newborn concept and it has recently started to receive more attention from a research perspective.
The study that is one of the first to explore the effects of forced channel migration is that of Konuş et al. (2009). The theoretical framework as constructed by Konuş et al. (2009) is considered and revised upon an extensive literature study. This chapter proposes the theoretical framework in the case of a cancellation of a distribution channel, which imposes forced channel migration upon the customer.
Forced channel migration
The cancellation of a channel confronts customers with forced channel migration. Konuş et al. (2009) describe forced channel migration as a strategy in which an organization obliges customers to use one channel over another for their information search, purchase or after sales. The organization uses coercive actions in order to enhance the efficiency of the firm’s channel operations. In general there are three situations in which a consumer is strongly directed in their channel choice; 1) the elimination of a channel; 2) high fees or rewards for utilizing a specific channel and 3) longer waiting times for using a specific channel. In this study situation one is subject and a definition of forced channel migration in its most literal sense is utilized: the elimination of a channel to compel customers to utilize a specific channel or channels in order to enhance the efficiency of the firm’s channel operations.
17 migration on overall demand in order to conclude on its influence on profitability. This effect is likely to be influenced by certain factors that can be modelled as a function of demand. The next paragraph considers possible factors and offers a conceptual framework. Paragraph 3.2 presents a depiction of the conceptual model.
3.1 Conceptual framework
In this paragraph the factors are discussed that are of influence on the relation between forced channel migration and demand. The factors considered are psychological reactance, customer satisfaction, use of a punishment or reward system, current channel use, customer value, age and gender. These factors are selected based on their likelihood to affect the mentioned relationship as suggested by literature. Moderating and mediating variables are distinguished. A moderating variable is defined by its indirect effect on the relationship between forced channel migration and demand. A mediating variable is characterized by its direct effect on demand.
Psychological reactance
18 Miyazaki et al. (2009) studied the moderating effect of psychological reactance on the consumption of products in several experiments with perceived consumption constraints. In the experiment in which a lack of channel access was used as a condition of constrained consumption, reactance was triggered and negatively affected the buying intention toward the product of subject. Furthermore, a study by Geyskens et al. (1999) shows that a coercive influence strategy such as channel migration results in stress and aggravation due to the loss of freedom. This indicates a direct effect of channel migration on reactance. This is supported by Konuş et al. (2009), their study indicates that reactance is stronger when channel migration is forced in comparison with voluntary channel migration. Therefore, the following hypothesis will be tested in this study:
H1: Forced channel migration leads to a higher level of reactance than does voluntary migration.
Customer satisfaction
19 line with the study by Geyskens et al. (19999) which shows that the coercive influence strategies associated with forced channel migration lead to aggravation and dissatisfaction. Therefore, the following is proposed:
H2: Forced channel migration leads to a lower level of customer satisfaction than does voluntary migration.
H3: Forced channel migration leads to a decrease in demand and this effect is mediated by satisfaction.
Moreover, research by Konuş et al. (2009) indicates that reactance has a significant mediating effect on the relationship between channel migration and satisfaction. Thus, it is suggested that the negative effect of forced channel migration on satisfaction proceeds from the positive effect of forced channel migration on reactance and the negative effect from reactance on satisfaction. This leads to the following hypothesis:
H4: The relationship between channel migration and satisfaction is mediated by customer reactance.
Punishment or reward system
20 that reactance primarily occurs if there is no reasonable explanation of the price difference between channels. For example, the Internet is cheaper for many products due to cost efficiencies from which the customer also profits and is generally more accepted as a reason for price difference (e.g. Brynjolfsson and Smith, 2000; Kaufmann et al., 2009). Furthermore, reactance can even be triggered among positively discriminated consumers in a perception of being deliberately influenced in their behaviour (Srivastava and Lurie, 2004). Nevertheless, rewards are more likely to enhance customer satisfaction and are associated with less reactance in a channel migration strategy than are punishments (Reynolds and Beatty, 1999). Considering the suggested effect of reactance and satisfaction on demand the following hypotheses are formulated:
H5: Customers confronted with a migration strategy reinforced with a reward system show an increase in demand via an other distribution channel.
H6: Customers confronted with a migration strategy reinforced with a punishment system show a decrease in demand via their usual distribution channel.
Current channel use
21
H7: Customers who use the conventional channel exhibit more reactance in response to a forced channel migration strategy than customers who already use the imposed channel.
H8: Customers who use the conventional channel exhibit lower satisfaction in response to a forced channel migration strategy than customers who already use the imposed channel.
Customer value
Wangenheim and Bayon (2007) found that high value customers of a firm significantly differ from low value customers in their response to negative and positive events as a consequence of overbooking capacity in service firms, for example airlines and hotels. They find that high value customers reduce the amount of transactions with the company more in case of a negative event, than low value customers do. Moreover, positive events (such as upgrading) only affect low value customers. This suggests a moderating effect of customer value on the relationship between channel migration and demand. Customer value is defined by the average number of tickets bought for live entertainment events by an individual. A customer of low value purchases up to two tickets per year on average, a customer of high value purchases three or more tickets per year on average.
In contrast, Konuş et al. (2009) find few differences between high and low value customers in relation to reactance. An argument that is opposed by the researchers from a channel perspective, is that the high value customers might be more experienced using alternative channels therefore making it easier for them to migrate. Furthermore, it could be that high value customers are more loyal to the organization and therefore are more accepting towards forced channel migration. Conversely, research shows that high value customers are generally more aware of their importance to firms and are expected to be treated with a higher level of service (Boland et al., 2002). The closer the relationship, the more a customer feels valuable and privileged. Considering the arguments that are proposed, the following hypothesis is stated:
H9: Customer value has a moderating effect on the relation between forced channel migration and reactance.
Age
22 results showed a decrease in level of reactance at an increase of age in years. This can be explained by the concept that older people are generally more emotionally stable and better capable of handling freedom-constraining situations and therefore showing less reactance than younger people.
Typically, other research on reactance did not include age as a variable in their study. Mostly due to the fact that in the majority of studies the sample consists of university students, thus the average age of participants is relatively young and shows little variance (Hong, 1990 in Hong et al. 1994; Joubert, 1990). Research by Ansari et al. (2008) indicates that elderly people are less likely to use the Internet, showing a significant age coefficient for channel selection. Therefore, the following is proposed:
H10: Age has a moderating effect on the relation between forced channel migration and reactance as young people and elderly people show more reactance than middle-aged people upon confrontation with forced channel migration.
Gender
The effect of gender on reactance is not unanimously recognized. Some studies showed a significant difference in psychological reactance between men and women, such as Joubert (1990) and Donnell et al. (2001). Joubert found that men scored higher than women on the reactance scale by Hong and Page (1989) for the small sample of 102 university students. Interestingly, Hong and Page did not find a significant result in their study. Joubert questions the reliability of the significant relationship due to the small sample size and its characteristics. He attributes the higher score on reactance by men to the local rules of fraternities which allows more latitude in the display of hostility and reactance. Moreover, Brehm and Brehm (1991, in Hong et al., 1994) argue that there is no reason to assume that differences in levels of reactance are explained by gender. Research by Hong (1989) provided evidence for this statement in their study. It is likely that gender is not of significant influence on the level of reactance as a consequence of forced channel migration.
Heterogeneity among consumers
23 found a positive effect of income on purchase volume. It is reasonable to assume that consumers who have more to spend, actually spend more. Educational level is also included as it serves as a valuable descriptor. Furthermore, Konuş et al. (2008) studied the existence of multichannel shopper segments and their covariates and found that consumers could be distinguished on several psychographic variables, namely innovativeness, loyalty, shopping enjoyment and price consciousness. It is valuable to include such psychographic variables in this study because it allows us to test if the effect of channel migration is equal for everyone and, if not, which segments can be distinguished based on among other these variables. The psychographic variables are discussed below.
Innovativeness
In this study, a consumer is considered innovative when it is among the first people to try a new product or service when it enters the market and is regularly seeking new experiences (Midgley and Dowling, 1978). Research suggests that consumers with a high level of innovativeness are more flexible and viable for marketing efforts (Ansari et al., 2008). Moreover, innovative people were early adopters of the Internet and pioneers in e-commerce. This implies that differences exist between consumers showing a high level of innovativeness and those showing a low level of innovativeness regarding their change in average demand upon forced channel migration.
Loyalty
24 (Wangenheim and Bayón, 2007). Either way, it is assumed that loyalty can be a distinguishing variable for segmenting consumers on change in average demand.
Shopping enjoyment
Shopping enjoyment can be defined as the hedonic value that consumers experience from searching for and purchasing products and services. Research indicates that consumers that enjoy shopping are more accepting towards purchasing online (Ansari et al., 2008). This suggests that variation in level of shopping enjoyment exists in consumer response to forced channel migration with respect to change in their average demand.
Price consciousness
Price consciousness is the extent to which consumers focus on paying low prices for a product or service (Lichtenstein et al., 1990). Therefore, price conscious consumers will also pay attention to the costs of a specific product that are associated with each distribution channel which one can utilize to purchase the item. Thus, in the case of forced channel migration these associated costs for the imposed channels are likely to have a stronger effect on highly price conscious consumers than on less price conscious consumers.
3.2 Conceptual model
This paragraph presents a depiction of the conceptual framework (figure 2) as outlined in the previous paragraph.
25 distinguishing quality of the psychographic variables is analysed and if segments can be identified, these will be described based on the demographic variables.
FIGURE 2 Conceptual model
Reactance Satisfaction
Age
26
4 RESEARCH DESIGN
In this chapter the research design is discussed. Paragraph 4.1 presents the method of data collection and consists of three subparagraphs which outline respectively the data design (4.1.1), the scales and measurement levels (4.1.2) and the sample design (4.1.3). The chapter proceeds with the explanation of the methods of data analysis in paragraph 4.2. Lastly, the validity and reliability of the data are discussed in paragraph 4.3.
4.1 Data collection
Further empirical research is performed by means of a survey. The conceptual model that is based on the literature study and interviews are the basis for the survey. The actual data collection is executed by market research company RMI by order of Mojo and the author of this report. Paragraph 4.1.1. discusses the data design, beginning with the survey method and preceding with the sample design. Paragraph 4.1.2 justifies the scales and measurement level used for the data collection. The sample design is presented in paragraph 4.1.3.
4.1.1 Data design
27 that it is fairly costly and time consuming. Nevertheless, Mojo finds that the benefits outweigh the costs of hiring a research company to administer the telephone survey.
Besides a telephone survey there is also an online survey performed. An important constraint of this method is that it rules out people that do not actively use the Internet. However, it is complimentary to the telephone survey and enables reaching fairly the entire target group. It is likely that the register for telephone survey rules out a certain type of person, for instance the assertive person that is innovative. Moreover, with the increase in cell phone use, many people are not even registered in the phonebook anymore. These groups of people can be targeted by the online survey. Combining these two survey means on can draw a highly representative sample of the population. Moreover, the respondents are selected by stratified sampling based on their geographical location in such a way that the sample offers a well representation of the target group on this variable. Furthermore, the segmentation variables age, gender, income and educational level are used as weighing factors resulting in a perfectly representative sample.
Personal interviews are not appropriate for collecting data for the reason that it is too expensive and difficult to realize a representative sample. However, the advantage of extra information that a personal interview can gain is realized in part by the interviews that were held with experts in the live music industry in order to determine relevant segmentation variables.
4.1.2 Scales and measurement levels
28 but this usually invites respondents to consider the other options less consciously. The justification of the choice for scales and measurement levels of variables that require this is outlined below.
Demand
As stated in the theoretical framework, firm specific demand is hard to measure in the live entertainment sector due to the fact that many visitors are not aware of which organization is responsible for the event. Therefore, demand is measured by a respondent’s average number of tickets purchased for live entertainment events per year. First the respondent is asked what the average amount of tickets bought per year is and subsequently one is asked whether the amount of tickets bought in 2009 is less, equal or more than the average amount. For respondents it might be hard to recall specifically the amount of tickets bought in 2008 and the amount in 2009, for that reason this approach is used and considered appropriate. The effect of channel migration on demand is measured by recoding the categories of the amount of tickets bought in 2009 compared to the average amount per year as follows. ‘Less tickets’ = -1; ‘Equal amount’ = 0 and ‘More tickets’ = 1.
Reactance
29 However, due to length constraints of the telephone survey the three items with the highest factor loadings are selected to measure the construct reactance. These are respectively “This leaves sufficient freedom to choose a distribution channel for buying my tickets”, “This makes me feel frustrated” and “This makes me feel forced to use the other distribution channels”.
Satisfaction
The measurement of satisfaction in relation with channel migration is performed by means of a three item bipolar scale, namely favourable-unfavourable, pleasant-unpleasant and negative-positive (Bearden and Netemeyer, 1999). One of the three items is reversely scaled to ensure that the respondent is still consciously considering the answer possibilities (Cooper and Schindler, 2003). To control for general satisfaction with live entertainment events, a simple one item scale for satisfaction is used ‘How satisfied are you overall with the organisation of live entertainment events’ with a 7-point Likert scale ranging from ‘highly dissatisfied’ to ‘highly satisfied’.
Psychographic variables
The scales of loyalty, innovativeness, shopping enjoyment and price consciousness are derived from the study by Konuş et al. (2008). Originally, they used multiple item scales in their study. However, in this study one item scales are used due to length constraints of the survey method. The item which showed the highest factor scores and/or applied to the product of subject is selected to measure the relative factor. Respectively, loyalty is measured by ‘I generally purchase the same brands’, innovativeness by ‘I am one of those people who tries a new product firstly just after the launch’, shopping enjoyment by ‘I like shopping’ and price consciousness by ‘I compare the prices of various products before I make a choice’. The scales are 7-point Likert scales, where one indicates ‘highly disagree’ and seven indicates ‘highly agree’.
Other variables
30 asked what the underlying reason is. Multiple options are given, moreover the respondent can answer ‘other’ and fill in their personal reason. Furthermore, respondents are questioned about the type of event they visit most (concert, festival or special event), the initiation of deciding to visit a concert, recognition of the Live Nation brand and their awareness of the cancellation of ticket sales via outlets.
TABLE 2
Operationalisation of variables in conceptual model
Variable Items Measurement level Scale
Demand Average amount of tickets bought per year ratio absolute number
Tickets bought in 2009 compared to average ordinal 3 categories
Reactance Extend to which one feels forced ordinal 7-point Likert
Level of frustration ordinal 7-point Likert
Level of freedom experienced ordinal 7-point Likert
Satisfaction Level of favourableness ordinal 7-point Likert
Level of pleasantness ordinal 7-point Likert
Level of positivity ordinal 7-point Likert
Customer value Average amount of tickets bought per year ratio absolute number
Current channel use Current most used distribution channel nominal 5 categories
Past most used distribution channel nominal 5 categories
Punishment system Influence sales if fee on regular channel ordinal 7-point Likert
Reward system Influence sales if discount on other channel ordinal 7-point Likert
Age Age in years ratio absolute number
Gender Sexe nominal 2 categories
Income Net monthly income in euro’s ordinal 13 categories
Level of education Highest level of education completed ordinal 8 categories
4.1.3 Sample design
31 match the population perfectly according to the Gouden Standaard 2009 of CBS. However, this sample requires a screening procedure to eliminate those who are not members of the group that we wish to study. This means that the members that are called are asked for their age first, may it be the case that they are younger than 15 or older than 65 they are kindly thanked and will not be questioned further. The Internet survey is administered to the online panel of RMI. The demographic characteristics of the database of panellists are registered, so that RMI is able to select a representative sample based on the same segmentation variables as used in the other sample.
The empirical data that is gathered is used to calculate the probabilities that one buys more tickets, an equal amount or more tickets based on the variables in the conceptual model. So, the probability that one purchases more tickets is P1, the probability one purchases an equal amount of tickets is P2 and P3 denotes the probability one purchases less tickets than they do on average per year. To determine the optimal sample size for these calculations one has to set the value of certain variables (Cooper and Schindler, 2003). First of all, the preferred precision of the probability estimation has to be determined. This research opts for a five percent precision, this percentage is widely used and acceptable. Then the confidence level needs to be set. A 95% confidence level is used to ensure that sound conclusions can be inferred from the analysis based on this data. The expected dispersion in the population is set on 50%. This allows the maximum sample size and corrects for possible skewness in the answers. Next, the optimal sample size can be determined by using the following formula:
n = (Z)² * (P * (100 - P) / S²)
in which:
n = required sample size
Z = factor for the calculation of the 95% confidence level = 1.96 P = expected dispersion percentage
S = precision of the probability estimation
32 The required sample size is 384. This is the number of respondents that is needed to complete the survey. However, the combination of survey methods requires a larger sample. It is decided that the sample size will be 500 of which 200 respondents will be selected for a telephone survey and 300 for the online panel.
4.2 Data analysis
Multinomial Response Models (MRM’s) are utilized because they allow probability estimation of a change in demand of live event tickets based on the predictive values of the variables included in the conceptual framework. The MRM’s are used as descriptive models to analyse the change in demand over the period before 2009 and 2009. The dependent variable change in demand is an ordinal variable consisting of three prescribed values. The values -1, 0 and 1 represent respectively a decrease in a consumers average demand of tickets, an equal average demand and an increase in a consumers average demand over the period before 2009 and 2009. MRM’s are appropriate in order to analyse categorical dependent variables with more than two response categories (Rodríguez, 2007). This type of model is especially appropriate for the analysis of nominal responses but also well suitable for the ordinal demand categories. However, this type of model does not make implicit use of the fact that the categories of the dependent are ordered. Therefore, the models are also estimated by Ordinary Least Squares (OLS) regression to assess the statistical validity of the MRM’s.
Two models are estimated, one providing the probability of an increase in average demand in the next year and one for the probability that a consumer will decrease its average demand. These two models are estimated by using equal average demand in the 2009 as the reference category. The following demand model explains consumer response in the form of a probability of consumer i to have either decreased or increased its average demand in year 2009 for live entertainment events in the Netherlands:
33 Pr (yi = 1)= α + β1channel + β2value + β3age + β4gender + β5satisfaction + β6event + β7trigger + β8income + β9innovative + β10loyalty + β11shopping + β12price + εt
yi = response category of change in demand for consumer i (-1 = decrease, 1 is increase) channel = channel use before 2009;
value = customer value (1 to 2 tickets per year is low; > 3 tickets is high); age = age of consumer in years;
gender = gender of consumer (0 = male, 1 = female);
satisfaction = general satisfaction with organisation of live events by consumer; event = type of event for which the consumer mostly purchases tickets;
trigger = usual mean by which information is received which triggers live event visit; income = net monthly income of consumer (≤ € 2500 = 1; > €2500 = 2; don’t want to say / don’t know = 3);
innovative = level of innovativeness of consumer; loyalty = level of loyalty towards brands of consumer;
shopping = general shopping enjoyment experienced by consumer; price = level price consciousness of consumer;
εt = disturbance term; α = constant.
34 corresponding demographic and psychographic characteristics. In contrast, if the results show that no valuable segments can be classified this is valuable marketing information as well as it can suggest that live entertainment event consumers can be treated the same with regard of communication and product offerings.
Mojo’s issue that underlies this research is addressed by the analysis of the two main questions: “Did Mojo lose customers due to the cancellation of the distribution of tickets through outlets?”, “And, if yes, what is the underlying reason for this loss?”. This is mainly done by the graphical and numerical presentation of the descriptive statistics, supported by Mann-Whitney U tests for significant differences between groups and Baron and Kenny’s (1986) four equation method for testing mediating effects.
4.3 Validity and reliability
Validity
The face validity of the data is covered with the justification of the scales and measurement levels (paragraph 4.1.2.). The contemplation of the number of items used for the different scales is concluded with a compromise between length of the questionnaire and completeness of the measurement of all domains of the construct. By selecting those items of scales that are relevant to the subject of study one can reduce the length of the questionnaire while simultaneously preserving the validity of the construct or scale.
Reliability
35
5 RESULTS
In this chapter the response on both the online as the telephone survey is discussed and the representativeness of the collected data. The analyses are outlined in paragraph 5.2.
5.1 Representativeness
Table 3 shows the demographical characteristics of the unweighted sample. Of the total of 800 respondents that were willing to participate in the survey, 259 respondents stated that they have never bought tickets to live entertainment events. Therefore, only the remaining 541 respondents were requested to fill in the complete questionnaire. The large share of respondents that never purchase any tickets for live entertainment events is not an issue for the analyses, because it is not a consequence of forced channel migration. As explained in the research design, a minimum sample of 500 relevant respondents drawn over the two survey methods was necessary to ensure a high level of representativeness. This minimum is amply met. A total number of 330 relevant responses are recorded via the online survey and 211 via the telephone survey.
TABLE 3
Demographical characteristics of the unweighted sample
Survey Response Relevant % male % female Av. age Mode income
level Mode education level Online 502 330 41.80% 58.20% 43 € 1501 - € 2001 HBO Telephone 298 211 43.60% 56.40% 37 € 2001 - € 2501 HBO Total 800 541 42.50% 57.50% 44 € 1501 - € 2001 HBO
Interestingly, the average age of the online panel is higher than that of the respondents reached via telephone. However, one cannot base any conclusions about online and telephone users on the demographics of this unweighted sample due to the process of partly directed calling in order to approach a representative sample of the Dutch population. Therefore, the telephone and Internet sample are combined for further analysis.
36 respondents is nearly equally divided in shares of males and females, with respectively 50.3% and 49.7%. The average age is 38. For analysis, respondent’s ages are divided over the following categories: 15 until 25, 26 until 35, 36 until 45, 46 until 55 and 56 until 65 years old. The distribution of the categories is fairly even (resp. 23.1%, 19.7%, 26.4%, 18.3% and 12.5%) with a slight underrepresentation of the category 56-55. The mode income level is € 1501 – € 2001 and has a relatively high standard deviation of 4.14 and is positively skewed (0.302) resembling the Dutch population. The mode education level is MBO, the distribution is negatively skewed (-0.303) which means that the more highly educated people outweigh the lesser educated people.
TABLE 4
Demographical characteristics of the weighted sample
Survey Response Relevant % male % female Av. age Mode income
level Mode education level Online 498 334 47.60% 52.40% 38 € 1501 - € 2001 MBO Telephone 302 211 54.50% 45.50% 38 € 2001 - € 2501 MBO Total 800 545 50.30% 49.70% 38 € 1501 - € 2001 MBO 5.2 Analyses
Frequency of live event ticket purchases
37
FIGURE 3
Frequency of average number of live event tickets bought per year in percentage of the total sample
The 545 respondents that purchase a minimum average of one live event ticket per year indicate to primarily purchase concert tickets, namely 65.5%, followed by special event tickets with 19.8% (table 5). 14.5% of the respondents mainly purchase festival tickets. There is no significant relationship between average number of live event tickets bought per year and the type of event where tickets are primarily purchased for (p = 0.446). However, there is a higher count for an average of one ticket per year for festival tickets compared to other numbers while other types of events count the largest frequency for an average number of two tickets bought per year.
TABLE 5
Percentages of type of live event mostly visited
Most people are usually triggered to attend a live event by friends and acquaintances that draw their attention to it (34%), followed by an advertisement in a newspaper or magazine (17%) or television commercial (15%) (figure 4). 1 Percent of the respondents generally uses advertisements in outlets as a source for the idea to attend a live event. In the specification of the ‘Other, namely…’ category respondents mention an array of different sources, such as
Type of live event Percentage
Concert 65.5% Festival 19.8% Special event 14.5% 31,9 20,8 21,2 8,6 6,0 3,8 2,7 0,5 1,5 0,1 1,4 0,4 0,5 0,2 0,1 0,0 0,2 0,0 5,0 10,0 15,0 20,0 25,0 30,0 35,0 0 1 2 3 4 5 6 7 8 9 10 12 15 20 24 25 30 How often per year on average do you buy a ticket for a live event
38 venue, artist or event calendars on the Internet, electronic newsletters, radio commercials or people visit returning events.
FIGURE 4
How are consumers triggered to attend a live event
Table 6 shows the reasons why people never purchase live event tickets. Multiple answers were allowed. More than half of the 255 respondents that never buy any live event tickets indicate that they do not have the desire to attend live events (54.9%). Moreover, 29% of the respondents indicated that they have little money to spend as a reason not to attend live events. The 19.6% of respondents that checked ‘other, namely…’ mostly answered that they do not have time to attend, they find the tickets too expensive or they are unable to attend due to health reasons. As discussed in paragraph 5.1, the respondents that never purchase any live event tickets were disclosed from further participation in the survey.
TABLE 6
Reasons why people never purchase any live event tickets
Reason why people never buy live event tickets Count Percentage
Someone else buys a ticket for me 6 2.40%
I receive a free ticket 4 1.60%
I do not have the desire to attend live events 140 54.90%
I have little money to spend 74 29.00%
I cannot order via the Internet 0 0.00%
I do not want to order via the Internet 1 0.40%
Other, namely… 50 19.60% Total 107.90% 15% 17% 3% 34% 2% 7% 1% 13% 8% B y a co mmercial o n t elevis io n
B y an ad vert is ement in a news p ap er o r mag azine B y a p o s t er in t he s treet s
B y friend s o r acq uint ances
B y an ad vert is ementl at a live event venue B y a p res ale b ro chure
39 Table 7 shows the average number of live event tickets bought per year over 2009 compared to the average per year in the period before 2009. The majority of consumers, namely 62.5%, has purchased an equal amount of tickets in 2009 on average. 16.6% has purchased more tickets in contrast with 20.7% that indicated to have bought less tickets compared with before 2009. The pre-categorized options to indicate a reason for this decrease ‘Offer less appealing’ and ‘Less money to spend’ were checked 27.9% and 29.5% of the times respectively. The category ‘Other, namely…’ was chosen by 42.6 percent of the respondents. Most of them mentioned time constraints related to work and family as the main reason why they bought less tickets in 2009. Some find tickets becoming too expensive, were unable to attend live events or they have lost interest. One respondent answered that it is unclear where to obtain live event tickets nowadays. Interestingly, none of the respondents indicated that they bought less tickets because the desired tickets were not for sale via outlets or brochure. Moreover, the data shows that 58.1% of the respondents are not even aware of the fact that some large organizers of live events do not sell tickets via outlets anymore since January 2009.
TABLE 7
Amount of tickets bought in 2009 compared to the average per year before 2009
Average amount of tickets in 2009 Percentage
Less 20.70%
Equal amount 62.50%
More 16.60%
Reasons why respondents bought less tickets
Offer less appealing 27.90%
Less money to spend 29.50%
Desired tickets not for sale via outlets or brochure 0.00%
Other, namely… 42.60%
Sales of live event tickets by distribution channel
40 channel is 12% larger, with an increase from 38% to 50% and is by far the largest distribution channel of live event tickets in 2009. In contrast, the outlet channel share decreased 5%, from 34% to 29% percent. Nevertheless, the share of sales via the outlet channel is substantial. The sales shares of the venue at the day of the event and telephone have decreased in favour of the Internet with 2% and 4% respectively. Hence, the Internet has gained share at the cost of the shares of the other distribution channels except for the presale brochure, which has remained stable at 5% of the total amount of live event tickets sold.
FIGURE 5 FIGURE 6
Share of distribution per channel before 2009 Share of distribution per channel in 2009
When one looks at the 106 respondents (20.3% of the sample) that have switched distribution channels between the period before 2009 and 2009, it is observable from which channel the respondents have emigrated from and to which channels respondents have migrated. Table 8 shows the emigration of consumers that used a specific channel before 2009 as a percentage of the channel’s distribution share in that period and as a percentage of the total market. The same is presented for the migration of consumers to a specific channel in 2009. Real change is represented as a percentage relative to the channel’s own size and to the market. The percentages of real change within the respective channels do not add up because they are relative to the channel’s own size. The real change in channel sizes within the market corresponds with the difference between figure 5 and 6. The outlet and telephone channel are subject to the greatest losses in their distribution share within the market. In terms of real change in perspective of the size of the distribution channel telephone and venue at the day of event have inclined the largest decrease, with a decline of respectively 39.0% and 16.4% in size. The presale brochure and online channel have increased over the period, with 16.0% and 28.7% respectively, compared to their previous share of 0.8% and
41 11.1% within the market. Interestingly, even though the tendency of migration to the Internet is facilitated greatly by the emigration from outlets and telephone, this tendency is compensated in part by the migration towards these channels.
TABLE 8
Emigration from and migration to distribution channels over the period before 2009 and 2009
Within channel Within market
Distribution channel Emigration Migration Real change Emigration Migration Real change
Outlet 24,9 % 11,0 % -16,0 % 8,6 % 3,2 % -5,5 %
Telephone 47,5 % 13,8 % -39,0 % 5,5 % 1,0 % -4,6 %
Presale brochure 12,0 % 25,2 % 16,0 % 0,7 % 1,4 % 0,8 %
Venue at day of event 32,7 % 18,7 % -16,4 % 3,4 % 1,6 % -1,7 %
Online 5,4 % 26,3 % 28,7 % 2,1 % 13,1 % 11,1 %
Total 20.30% 20.30% 0%
42 answer category received. The majority of the respondents with 51.4% indicates that it would purchase the desired tickets through a different channel, followed by 32.4% of the respondents claiming to purchase no live event tickets at all if such a situation would occur. 13.5% checked the ‘other, namely’ category and stated that they would then either turn to the black market, raise objections at the address of the operator of tickets for this live event or they would ask someone else to purchase a ticket for them. Only one respondent indicated that it would purchase tickets for a different live event in the case tickets would not be available via outlets, suggesting that switching behaviour is hardly an issue.
TABLE 9
Frequencies answer categories what people would do if the by them desired tickets were not available via outlets for two independent samples
Answer category Sample in train Sample at Carré Total frequency
I would purchase no tickets 25.00% 41.20% 32.40%
I would purchase tickets through a different channel 50.00% 52.90% 51.40%
I would purchase tickets for a different live event 5.00% 0.00% 2.70%
Other, namely 20.00% 5.90% 13.50%
There are no significant differences between the answer on the statement and age, type of event, gender, education level and city or village inhabitant. However, this is partly due to the small sample size. There are some notabilities in the dispersion of the answers on the statement. For instance, most of the festival and special event consumers state that they would purchase tickets through a different channel. Moreover, age has a wide range for both the answer ‘I would purchase no tickets’ as ‘I would purchase tickets through a different channel’, suggesting that age is not a distinguishing factor.
Reactance and satisfaction
43 standard deviations for the different levels of reactance and satisfaction for both migration strategies are presented in table 10. The mean level of reactance is 2.6 on a scale from one to seven in which one corresponds with a very low level of reactance and seven with a very high level. For satisfaction the mean level is 4.8 on a similar scale for satisfaction. Then, the respondents are confronted with the current situation, in which a number of large organizers of live events do not sell tickets anymore via outlets, and again the respondents are asked to indicate their level of reactance and satisfaction. The level of reactance in this situation of forced channel migration is significantly higher than when voluntary channel migration is hypothesized (M. = 3.7, p = 0.000), this result supports H1. Moreover, the mean level of satisfaction is significantly lower for the forced channel migration situation with a level of 3.5 compared with voluntary (p = 0.001), providing evidence to accept H2.
TABLE 10
Significance of differences in level of reactance and satisfaction between migration strategies
Level of reactance Level of satisfaction
Migration strategy M. S.D. M. S.D.
Voluntary 2.6 1.2 4.8 1.4
Forced 3.7 1.5 3.5 1.6
p-value test for difference
between strategies 0.000 0.001
The level of reactance and satisfaction in the situation of forced channel migration is controlled for the awareness of the respondent of the elimination of the outlet channel by a number of large live event organizers beforehand. The mean levels of both variables are similar for respondents who were aware of this fact and those who were unaware. The mean reactance level of aware respondents versus unaware respondents is respectively 3.6 (S.D. = 1.5) and 3.7 (S.D. = 1.5), for satisfaction these levels are respectively 3.4 (S.D. = 1.7) and 3.6 (S.D. = 1.6). Unsurprisingly, the Mann-Whitney U test shows no significant differences between these groups, neither for reactance (p = 0.518) nor for satisfaction (p = 0.251).
44 migration (resp. p = 0.012 and p = 0.036) in such a way that the satisfaction level in both scenarios is higher for people that indicate to be more satisfied with the organisation of live events in general.
A Mann-Whitney U test for differences in change in demand for respondents with low versus high general satisfaction with the organisation of live events does not produce a significant result (p = 0.148).
Mediation effects
The possible mediation effect of satisfaction on the relationship between migration strategy and change in demand cannot be tested for the reason that none of the respondents purchases less live event tickets due to forced channel migration. Therefore, it is impossible to measure the effect of satisfaction in this relationship because one cannot distinguish between groups classified by a particular change in demand that is explained by migration strategy and satisfaction level. Based on this fact one can conclude that migration strategy does not affect demand, providing support to reject H3.
The probable mediation effect of reactance on the relationship between migration strategy and customer satisfaction is tested by Baron and Kenny’s (1986) method of using four regression equations. However, before preceding with the method one condition has to be met. The independent, mediator and dependent variable should show intercorrelation (Baron and Kenny, 1986). Table 11 shows an overview of the intercorrelations between the variables, their mean and standard deviation. All variables are significantly correlated and proceeding with the method is allowed.
TABLE 11
Means, standard deviatons and intercorrelations between variables
Variable M. S.D. 1 2 3
1 Migration strategy 1.50 0.50 -
2 Reactance 3.16 1.45 0.369 -
3 Satisfaction 4.13 1.66 -0.407 -0.657 -
45
FIGURE 7
Paths of regression equations
TABLE 12 Results of regressions
Path Function of Regressed on R² Beta p
a reactance migration strategy 0.128 0.357 0.000
b satisfaction migration strategy 0.144 -0.380 0.000
c satisfaction reactance 0.429 -0.655 0.000
satisfaction migration strategy 0.463 -0.191 0.000
reactance -0.586 0.000
First, reactance is regressed on channel migration strategy resulting in a significant effect (p = 0.000). Thus the independent variable causes the mediator as required. Second, migration strategy significantly affects satisfaction in the regression equation (p = 0.000). Third, reactance negatively affects satisfaction significantly (p = 0.000). Lastly, satisfaction is regressed on both reactance and migration and also this relation is significant (p = 0.000). If the weight of the coefficient for migration strategy was non-significant then full mediation by reactance would be proven. Yet, the weight is significant suggesting that there is partial mediation or an indirect effect of reactance on the relationship between migration strategy and satisfaction. Because migration strategy causes reactance there is correlation between these variables which results in multicollinearity. Therefore, one needs to look not only at the significance of this equation but also at the size of the coefficients. Standardized coefficients (beta’s) are presented in table 12 which enables one to compare the strength of the coefficients. The weight of the coefficient of migration strategy is lower when satisfaction is regressed on both variables than when regressed on migration strategy alone, with beta’s of respectively -0.191 and -0.380. This indicates that reactance partially explains the relationship between migration strategy and satisfaction. Moreover, the R² is highest when both reactance and migration strategy are included (R² = 0.463) in the regression of satisfaction compared
Channel migration strategy Customer satisfaction
Reactance
a b
46 with the equation in which only migration strategy is included (R² = 0.144). Thus, it can be concluded that reactance is partially mediating the relationship between migration strategy and satisfaction. Therefore, H4 is only partially supported.
Punishment or reward system
The influence on demand for live event tickets is considered for two types of reinforced channel migration strategies, namely one with a punishment system and one with a reward system. A 7-point Likert scale measures the likelihood that respondents would buy more tickets via another distribution channel if they would receive a discount on these tickets, ranging from ‘definitely not’ to ‘definitely’. This represents the reward system. A mean level of 4.9 (S.D. = 1.8) is recorded for this system, suggesting an increase in ticket sales via another channel than the respondent’s usual channel if it is offered a discount. This result supports H5.
The effect of a punishment system was measured by hypothesizing a fee on tickets bought via the respondents regular channel. The scale measured the likelihood that a respondent would purchase less tickets via their usual channel if such a fee was imposed and ranged from ‘definitely not’ to ‘definitely’. A mean level of 4.8 (S.D. = 1.9) is recorded, indicating that in general people would buy less tickets via their usual distribution channel if confronted with a fee, supporting H6.
Mann-Whitney U tests indicate that respondents which would purchase more tickets via another channel in the presence of a reward system show a significantly higher satisfaction level for voluntary migration (p = 0.008). This group shows a mean satisfaction level of 4.8 for the voluntary migration situation compared to a mean level 4.4 of the group that would not purchase more tickets via another channel. This could suggest that the latter group is somewhat less flexible with regard of channel use.
47 regular channel. No other significant differences are found in reactance and satisfaction levels for both channel migration strategies.
Thus, results show relations between consumer reaction on reward and punishment systems and the reactance and satisfaction levels in case of a voluntary or forced migration strategy. However, the differences in the mean levels are marginal.
Current channel use
48
TABLE 13
Significance of differences between outlet channel users and Internet&telephone users before 2009
Voluntary migration Forced Migration
Average demand Change in demand Reactance Satisfaction Reactance Satisfaction
M. 2.9 -0.14 2.7 4.7 4.2 2.2 Outlet users S.D 3.5 0.63 1.2 1.4 1.3 1.4 M. 3.2 0.03 2.4 5.0 3.1 3.8 Internet&telephone users S.D. 3.1 0.61 1.2 1.4 1.4 1.7
p-value test for difference
between user groups 0.024 0.012 0.002 0.009 0.000 0.000
It is tested whether there is a difference in satisfaction and reactance level upon confrontment with forced channel migration for respondents that have switched from the outlet channel to another channel over the period before 2009 and 2009 and respondents that have not switched channels. A significant difference is found in the level of reactance for the group that has switched channels at the cost of the outlet share compared to the other respondents, with respective mean levels of 4.5 and 3.6 (p = 0.000) (table 14). Furthermore, a significant difference is found for the level of satisfaction (p = 0.000). The means levels of satisfaction are respectively 2.8 for outlet abandoning respondents and 3.5 for respondents that have not switched channels over the period before 2009 and 2009.
TABLE 14
Significance of differences in level of reactance and satisfaction in case of forced channel migration between switchers from outlet channel and non-switchers over the period before 2009 and 2009.
Level of reactance Level of satisfaction
Group M. S.D. M. S.D.
Switched from outlet channel 4.5 1.6 2.8 1.4
Did not switch channels 3.6 1.6 3.5 1.5
p-value test for difference
between groups 0.000 0.001
