Service failures in a cruise line context: Suggesting categorical schemes of service failures

This paper reports a study on service failures carried out on the prestigious Norwegian Coastal Voyage (NCV) or the Hurtigruten which is the brand name. Service failures are studied from the perspective of the service employees. Seven tour conductors participated in the study and in the research personal, face-to-face, in-depth interviews were carried out using an interview guide as a data collection tool. The respondents provided descriptive accounts of 51 service failure incidents which were used to develop a classification scheme of service failures. The service failure classification scheme consists of two main categories: 1. Service failures on excursions; 2. Service failures on the ships (while at sea or at harbour). The paper describes and discusses a set of sub categories within the main categories. The case study may advance the knowledge field of service failures particularly relevant for a high contact service like a cruise operation. © 2011 International University College. All rights reserved


Introduction
This paper reports a study on service failures from research carried out on the prestigious Norwegian Coastal Voyage (NCV) or Hurtigruten which is the brand name. The context of the study is interesting due to historical, cultural, geographical and tourism reasons. Historically, the Hurtigruten has been a main means of transportation along the Norwegian coast for more than a hundred years. Culturally, the Hurtigruten has provided pride and identity to the scattered coast line population in Norway, and geographically in the sense that the Hurtigruten has united the coastal settlements along the Norwegian coast from Bergen to Kirkenes. However, it's the tourism dimension which was given strategic priority by the introduction of "new" ships during the early 90s that brings this research into realisation, because the tourists' portion of the travellers on the Hurtigruten has gradually increased in numbers as well as in economic importance. The lack of substantial research on service failures in the cruise line industry and particularly on the Hurtigruten, provides the rationale for the research carried out. The choice of a cruise line context for the study is highly relevant as the cruise line industry is in "an early stage of development" (Biederman, 2008:197).
Service failures are an inevitable part of service processes (Hoffman and Bateson, 1997). Thus, the effective management of service failures is important for service quality improvements (Bejou and Palmer, 1998). From the tourists' point of view the most immediate evidence of the service quality occurs in the service encounters or in the "moments of truth" when the tourists interact with the service personnel (Czepiel, 1990, Grönroos, 2006. Onboard the Hurtigruten, as a high contact service (Lovelock and Wright, 1999), the front line personnel and especially the tour conductors, have vital service roles to play.They possess important high contact roles during phases of service delivery (Grove and Fisk, 1992). As the cruise line industry is featured by a high degree of interactions between the service employees and the travellers, there are many opportunities for service failures to occur.
Researching service failures is an important field within services management (e.g. Grönroos, 2000). Even though the issue of service failures has been examined for some time, academic research on service failures is "relatively recent and still in progress" (Lewis and McCann, 2004:6). Since 2004, several service failure studies have been conducted which evidence that this stream of research is still progressing. A primary reason for it is the increased role of the service sector as an important economic force, especially in Western societies as more than 2/3 of the entire "work force is employed in services" (Gummesson, 2000:7). This encompasses what Kim and Mauborgne (1999:44) have expressed as the "arrival of a knowledge economy". The knowledge economy is characterized by increased turbulence, uncertainty and ambiguity (Johannessen and Olsen, 2010). A growing academic literature is emerging on the knowledge economy (e.g David and Foray, 2003).
The cost of not delivering high class service is the "cost of quality", for example the risk of losing customers (Keaveney, 1995), negative "word of mouth" (e.g. Susskind, 2002;Swanson and Kelley, 2001), decreased employees' satisfaction and morale , reduced customer satisfaction (Smith et al., 1999;Zeithaml et al., 1988), and decrease in customer loyalty (e.g. Bejou and Palmer, 1998;Maxham, 2001 andMaxham andNetemeyer, 2002). Consequently, from the management's point of view it is essential to identify, describe and analyse the nature and the extent of service failures in order to better understand and judge customers' perceived level of service quality (Berry and Parasuraman, 1992;Zeithaml et al., 1988). Thus, according to Heskett et al., (1994), a shift to a quality focus is essential to the competitive survival of service businesses.
The study addresses two questions: 1. What service failures occurred on the Hurtigruten? 2. How may the service failures identified be categorised into a scheme of main and sub categories?
The first question is answered by providing a set of descriptive accounts of the service failure incidents reported. The second research question is answered by following a procedural approach to categorise the service failure incidents that were identified. The primacy of this paper is adequately to answer the second research question. In order to do so, a content analysis of a set of service failure incidents is required.
The paper is organised in seven parts. Following the introduction, the second part describes quite briefly the empirical context of the study. Part three constitutes a focused review of the empirical research on categorical schemes of services failures reported in the literature. In part four, the methodology details is outlined while part five entails the research findings. A discussion of the research findings follows in part six. Part seven ends the paper by discussing implications of the research findings and draws a series of conclusions. The limitations of the study are included in this part. This paper encompasses theoretical, methodological as well as practical features.

The empirical context
On the 2 nd July 1893, the Norwegian steamship D/S Vesteraalen embarked on a voyage from Trondheim to Hammerfest. About 70 hours later the ship arrived at Hammerfest. Three weeks earlier the Norwegian Parliament had agreed upon a proposal to establish a coastal route along the Norwegian coastline from Trondheim to Hammerfest. This was a new and quite astonishing idea, to sail days and nights along one of the most weather-beaten coasts in the world under all climatic conditions. Despite skepticism and some resistance from experts in sea transport, the Hurtigruten became a success in sea transport. In 1898, the route was extended to Bergen down south, and from 1908 to Kirkenes up north. Today, more than a 100 years later, the Hurtigruten has become a tourist attraction product of significant importance in Norway comprising 2500 nautical miles, 34 ports of call, 11 ships and an elevenday return voyage Bergen -Kirkenes. In appendix 1 the sailing route is presented as it is today (2010).
Historically, the Hurtigruten has been an important means of communication and transport to the scattered population along the Norwegian coast. However, due to improved infrastructure, such as better roads, a web of new airports and a network of fast running boats along the Norwegian coast; the role of the Hurtigruten has gradually changed from being an important means of infrastructure to a substantial tourist attraction product. In order to sustain the growing competition from the national and international cruise line industry, the Hurtigruten fleet has physically changed, especially during the 90s, by the investments in new and modern ships which appear like conventional smaller explorer cruise ships.
Two shipping lines were involved in the Hurtigruten trade at the time of the collection of the data. However, in 2006 the two shipping lines merged into one cruise line company, the Hurtigruten ASA. Management of the company has set the strategic goal of becoming the world's leader in explorer cruises within 2012. This is an ambiguous goal particularly in the light of the growing number of operators in the cruise line industry (Biederman, 2008).

The literature review
Research on service failures as a field within services marketing and management (Brown et al., 1994) was initiated by the classical work of Nyquist et al., (1985) on communication difficulties in service encounters. Since the time of that study, research has focused on several conceptual as well as theoretical and practical issues in a diverse set of service industries. However, according to Lewis & McCann (2004), academic research on service failures is relatively recent and still in progress. The situation is still the same today as new services are developed and established services like cruise operations are becoming more complex and competitive (Biederman, 2008). The increased turbulence and complexity in the business environment is by D'Aveni (1994) termed hyper-competition. This implies the need for better business performance by the suppliers in order to meet enhanced customer expectations to the quality of the services delivered (Grönroos, 1988, Grönroos, 2006. According to Lewis and Entwistle (1988), the relationships between a service provider and the customer are decisive in most service operations. Service failures arise when customers' experience dissatisfaction with the service provider's service delivery (e.g. Lewis & Spyrakopoulos, 2001;Bamford and Xystouri, 2005). A service failure is "a breakdown in the delivery of service, service that does not meet customer expectations" (Hoffman and Bateson, 1997:327). Obviously, the customer is the person who knows best when a breakdown in the service delivery occurs (Lovelock and Wright, 1999).This is a situation when the service provider does not deliver the expected level of service quality as perceived by the customer (Grönroos, 2006). However, the service employees are also considered to be important sources of information on service failure incidents (Bowen and Schneider, 1985;Tax and Brown, 1998).

The consequences of service failures
According to Lewis and Clacher (2001), service failures are inevitable especially in high contact service processes and to manage service failures well is an important task for service managers (Grönroos, 2006). During stages of service delivery mistakes will happen and things will go wrong. Consequently, when the service provider does not live up to customers' expectations to the service offerings, this may have negative consequences (Michel, 2001).
A review of the empirical research literature on the negative consequences of service failures illuminates a whole set of different consequences. In the Bitner et al., (1990) study, as an example, a decrease in employee morale and performance was emphasised.
Other studies point at consequences like dissatisfaction with the service provider (e.g. Kelley et al., 1993); a decline in customer confidence (e.g. Boshoff, 1997); negative wordof-mouth behaviour (e.g. Mattila, 1999; and customer defection (Miller et al., 2000). Thus, the occurrence of service failures may have a set of negative consequences for service providers. This implies the need for effective complaint management (e.g. Eccles and Durand, 1998).

The service failure classification process
A review of the empirical research on the service failure categorisation process shows the use of different research perspectives. The Nyquist et al., (1985) study for example, investigated communication difficulties from the perspective of the service firm employee contrasting for example the Bitner et al., (1990) study which was carried out from the perspective of the customer. The Edvardsson (1992) study which investigated service failures in an airline employed both perspectives.
In reviewing the empirical research on the service failure classification process, two opposing directions emerge: 1. The Bitner, Booms and Tetreault (1990) main classification system 2. The Edvardsson (1992)

main classification system
The Bitner et al., (1990) study suggested in a classical study three main categories of critical service incidents: 1) Employee responses to service delivery system failures. Service delivery system failures pertain to service failures in the core offering of a firm. An example in a cruise line operation can be a delay when a ship returns to the cruise port.
2) Employee responses to customer needs and requests. The second category of service failures pertains to employees' responses to customer needs and requests. In a cruise line operation, with a 24/7 service guarantee', a request for a service and not responded to, is an example within this category.
This category pertains to events and employee behaviours, good or bad, which are totally unexpected by the customer. In a cruise line operation, service employees may, quite unexpectedly, be rude and abusive. This is an example of deviating behaviour which needs to be dealt with by service management.
Several studies have employed the Bitner et al., (1990) classification system, for example Hoffman et al., (1995) and Hoffman and Chung (1999). However, the Bitner et al., (1994) study expanded it with a fourth main group: problematic customer behaviour. The root cause of customer dissatisfaction was the customer.
Consequently, reviewing the empirical research literature, the conceptual service failure classification scheme developed by Bitner et al., (1990), expanded with a fourth main group by Bitner et al., (1994), seems conceptually robust in order to classify critical service incidents across a range of service industries.
In a study of service failures in an airline, Edvardsson (1992) suggested three main categories: air transport, ground transport and other incidents. Edvardsson's advice was not to use a previously set of categories for classifying critical service incidents in unique service contexts, as the sequential phases of an airline travel (Gustafsson and Johnson, 2003). The rational for this argument is the key characteristics of an air line operation; the sequences of in-checking, the flight and the out-checking procedure.
The Bejou et al., (1996) study for example, tried to identify, describe and analyse service breakdowns in airlines in Sweden and US, and employed the same classification scheme as Edvardsson (1992). Thus, when examining service failure incidents in one organisation or in few organisations within the same industry, the Edvardsson's (1992) main classification scheme seems most promising due to the uniqueness of the research setting(s) while the Bitner et al., (1990) classification system seems more appropriate and useful when the research is carried out on a range of different organisations across service industries.
The service context in which a cruise line operator works is quite similar to that of an airline. The tourists travel to the cruise line port as air travellers to the airport, they check in, they participate on the cruise and when returning to the port, they check-out and they finally return back home. Obviously, service failures may take place during the whole travel process. However, it's on the cruise itself, on the ships and/or on excursions that the risk of service failures is most likely to occur.

The process of developing sub categories
The inductive process of developing varies across all studies. The Bitner et al., (1990) study developed 12 sub categories while for example the Chung and Hoffman (1998) study suggested 11 sub categories. The Edvardsson (1992) study suggested five sub categories in the main category air transport: delays, cancelled flights, delayed or damaged luggage, overbooking and other sources. The point seems to be that the sub categories have to be developed according to the information collected in each empirical setting. In this study, the analytical case approach suggested by Edvardsson (1992) is followed due to the service process characteristics of cruise operations which are comparable to those of airlines.

The methodology
The Critical Incident Technique (CIT) methodology guided the research procedure. The CIT, originally developed for research purposes by Flanagan (1954), has proved to be a robust and reliable research methodology to employ when the aim is to increase the knowledge about a phenomenon about which relatively little is known (Bitner et al., 1990). The CIT, as a qualitative research methodology, is essentially a classification technique analysing critical service incidents which are events that can be described in detail and that deviate significantly, either positively or negatively, from what the customers expect or consider normal in service encounters (Chung and Hoffman, 1998). In order to categories the service incidents, the methodology employs content analysis of "stories" or anecdotes" as data (Krippendorf, 1980). According to Michel et al., (2009), service failure information may be collected by means of three methods; total quality management (TQM), mystery shoppers and critical incidents. In previous empirical research on service failures, the last option is the most commonly used and is the choice in the research reported in this paper.
When using this methodology, data are collected by means of an interview guide and the results are content analysed Johnston, 1995). In this study, a case methodology was chosen (Merriam, 1998;Yin, 1994;2003). This approach is particularly useful when examining a phenomenon about which relatively little is known (e.g. Gummesson (2000), Mehmetoglu (2004), among others).
Information may be collected from both primary and secondary sources (Booth et al., 2003). The examination of secondary data such as annual and internal reports provided a thorough grounding into the Hurtigruten as a composite tourist attraction product. In a strategic report (1999), "Strategic Choices after 2001", as an example, a strategy of becoming an explorer cruise line company is discussed. In order to become more competitive internationally, a set of service quality issues had to be addressed, e.g. how to obtain customer satisfaction and loyalty. However, secondary data has often a "data-fit" problem (Kinnear and Taylor, 1991).Primary research information offers more specific and extensive data from informants who can provide information to the questions under scrutiny (Veal, 2006).
As evidenced in the literature section, the choice of empirical research perspective needs consideration. The research in this paper is carried out from the perspective of the service employees', a perspective which in previous empirical research has proved to be a reliable alternative to the customer's perspective. The service employee perspective values the decisive role of service employees for effective service delivery (Bettencourt and Gwinner, 1996).The primary reason for the choice was "practicality". Within the resources available, e.g. economy, it was considered more convenient to access service employees rather than the tourists traveling on the Hurtigruten.
A range of employee categories works on the Hurtigruten, and both back-stage and frontstage personnel are important for superior service delivery. However, the tour conductors belong to the service group that regularly have most face-to-face interactions with the tourists, and consequently know the tourists best. For many tourists, the tour conductor on the Hurtigruten is simply the service. The tour conductors are in favourable positions to both observe incidents of service failures and to be told "stories" or "anecdotes" from tourists about what caused them to be dissatisfied with the service offering. Thus, a decision was taken to include the tour conductors as subjects in the study.
The shipping personnel department released a list of thirteen tour conductors divided into two separate groups; regulars and substitutes. A decision was taken to include only regulars in the convenience sample due to their extensive experience and expertise in service operations. The regulars consisted of eight persons. One person did not want to participate in the study for personal reasons as the person did not approve the turn away from the "old" Hurtigruta towards the tourist "model" of the new ships.
In qualitative research, the personal interview is the most widely used method of research (Fielding, 1993). The personal, face-to-face indepth interview was judged to be an appropriate research tool due to the linkage to the CIT methodology, the limited number of informants and as the topic and questions under scrutiny required closeness to the respondents in order to collect "stories" or "anecdotes" of service failure incidents. According to Edvardsson (1992), data on critical service incidents can be collected in several ways and the personal in-depth interview is usually preferable due to "richness" and "depth" in data collected.
A pilot study was conducted before the main study. Two previously employed and highly qualified tour conductors participated in the pilot study. One issue in the pilot study was the consideration of collecting demographic information of the informants such a gender, age, education, home place etc. Another issue was the content and sequences of the questions in the interview guide. In light of the feedback, the demographic variables were assessed to be of minor importance in the study because the focus of the study was on the service failures reported to the tour conductors from the tourists. By not collecting demographic details, the thinking was that a good interview "atmosphere" would be created. The pilot study participators also insisted that collecting demographic information would perhaps hinder employee participation in the research. In the pilot study, the interview guide started with the open-ended questions "what happened and why it happened" (service failure "story"/"anecdote"). The pilot study participants suggested that the sequences of questions should be altered by starting in a "smoother" way before addressing the open-ended questions. Consequently, the order of the questions was adjusted.
An interview guide served as an information gathering tool. For the purpose of this paper, the questions "what happened and why did the service failure occur?" are linked to the two main questions raised in the research. The interviews with the tour conductors were conducted 2001. The long interviews (McCracken, 1988), each lasted for about two hours, were administered by the author. The interview process was carried out in three phases. First, the seven tour conductors were contacted in writing in order to explain the purpose of the research and to encourage them to participate in the research. Second, they were contacted by telephone in order to get their approval and to agree when to carry out the formal interviews. During the telephone conversation the tour conductors were asked to try to recall incidents of service failures, write down the details, and bring with them the notes to the formal interview meeting. Third, the long interviews with the tour conductors were carried out when the Hurtigruten was at sea between Harstad and Svolvaer. See informative map. During this part of the voyage, the tour conductors had the time available fully to concentrate on the interview situation as this part of the journey starts quite early in the morning and most of the tourists are occupied with the breakfast and some of them participate in an excursion between Harstad and Sortland. See map for geographical orientation (appendix 1). During the interviews, which were tape recorded, the tour conductors provided 7 to 8 descriptive accounts of service failures each. In total, information on 51 service failure incidents was collected. The "stories"/"anecdotes" then serve as data for the analytical purpose of suggesting categorical schemes of service failures in the case study context.

Findings
Research question 1: The first procedural step started with a repeated and careful reading of the 51 service failure incidents reported by the tour conductors. Appendix 2 provides a list of the service failure incidents reported by the service employees. An inspection of the list shows that 22 service failure incidents took place on excursions while 29 service failure incidents happened on the Hurtigruten while at sea or at harbour(s). The key words used in the appendix illustrate what happened when the service failures occurred. It is beyond the scope of this paper to provide full descriptive accounts of the total number of service failures that occurred. However, for the purpose of illustration, texts of a set of service failure incidents are provided under this heading. In the next part, answering question 2, additional information on a set of the incidents will be provided which is needed in order to develop the categorical schemes of service failures.
Service failures on the Hurtigruten while at sea or at a harbour(s): A total of 29 service failure incidents is allocated to the main category of service failures on the Hurtigruten while at sea or at a harbour. For the purpose of illustration, five service failure stories/anecdotes are provided.
Service failure story no. 6: The provision of TV channels A tourist group from Germany, according to the first interviewee, complained to the tour conductor about the lack of television channels from Germany. A female argued that "the Germans outnumber the English and US travellers on the NCV so why only the CNN?" According to the tour conductor, the group argued that they could not understand that the only international TV channel onboard was the CNN. What made the situation especially challenging, according to the tour conductor, was that the weather on the rest of voyage was rather bad, which reduced the total travel experience.
Service failure story no. 19: Cabin location A German couple in the 40-50s who travelled on the Hurtigruten for the first time was very dissatisfied with their cabin location because of noise from the propellers. The incident happened on one of the new ships which have several propellers in order to improve navigation. According to the tour conductor; the third interviewee, the man claimed that "although the ship was quite new, there had to be something wrong with the machinery". However, according to the tour conductor, the service staff could not offer a more favourable cabin location because the incident took place in the summer season and the ship was fully booked and all the cabins were occupied.
Service failure story no. 28: The menu in the ship's restaurant A married, middle-age couple from Germany travelled on the Hurtigruten for the first time. According to the tour conductor, the fourth interviewee, an important part of the core product is the breakfast, lunch and dinner meals served in the restaurant onboard. The gastronomic standards of these meals are high and the couple was tempted by the lavish food and simple ate too much. They incurred painful indigestion and could not fully enjoy the rest of the voyage due to sickness. However, according to the tour conductor, the female traveller said that "she herself was to blame for eating too much as she in the daily life, more or less, was on diet". Service failure story no. 40: "Poisoned" food in the ship's restaurant: Two Norwegian ladies in their 60s entered the Hurtigruten in Bergen to travel to Kirkenes up north and to Bergen down south. According to the tour conductor, they became very demanding customers from the moment they entered the ship. At the first dinner seating in the ship's restaurant they claimed that the fish served was poisoned and could not be eaten by human beings. According to the tour conductor, their enduring complaints and behaviour during the voyage created a nervous atmosphere among the waiters in the restaurant and embarrassed other tourists as well. The two ladies caused severe service problems and challenges for the service personnel during the voyage. The tour conductor said that "we, the service personnel, did what we could to please the ladies, but in vain".
Service failure story 45: Film episode A film team entered the Hurtigruten in the busy summer season. According to the tour conductor, the film crew behaved as if they "owned" the ship, for example moving the tourist away from their seating in order to get the "right" location for shooting the film. According to the tour conductor, several tourists complained, and a US lady claimed that "this would never happen on a cruise ship back home". The episode created a stressful atmosphere, and according to the interviewee, no information was given to her, or the tourists, from the ship's management about the film team entering the ship and the consequences of their presence.

Service failures on excursions:
A total of 22 service failure incidents are allocated to main category of service failures on excursions. For the purpose of illustration, three service failure stories are provided.
Service failure story no.1. Fog on the North Cape: According to the first tour conductor to be interviewed, a middle age tourist couple from Israel travelled on the Hurtigruen for the first time and they were especially looking forward to the North Cape excursion. Returning from the excursion, however, according to the tour conductor, the couple expressed disappointment. The cliff was covered in fog and they did not see anything at all except the outline of the North Cape building. "This was complete waste of money", they both uttered. According to the tour conductor, the couple did not raise a formal complaint but on several occasions they reminded her about the incident. According to the tour conductor, the couple argued that "they would not have participated if they had received weather forecast information prior to the excursion".
Service failure story no. 9. Excursion fully booked: An English tourist in his 40-50s experienced a service failure which seems to occur quite frequently. When buying his ticket in a travel agency in London, he also bought a package of seven excursions. What he did not know, because the travel agent failed to inform him, was that some of the excursions had to be bought onboard the ship. One of these excursions is the popular excursion to the Svartisen glacier which has a capacity limit of 90 tourists. When the Englishman discovered that this excursion was not included in his package and the excursion was fully booked, he flew into a rage. According to the tour conductor, the Englishman argued that "he had the right to participate". However, this was not possible due to the capacity limit of the landing boat. According to the tour conductor, the man was angry for the rest of the voyage, especially after hearing from fellow travelers that the excursion was "fantastic" and a once in a lifetime experience.
Service failure story no. 11. Unprofessional guide in Tromsø According to the second interviewee, a tourist group from Germany reported to the tour conductor that the local guide on a Tromsø guide lacked the required communication competency and skills. According to the tour conductor, a representative from the group claimed that the guide "had to read from the manuscript all the time" and "that the excursion was a total farce". The group was very disappointed with the total excursion experience. According to the second interviewee, the group was astonished by the low competency level of the guide as they thought the excursions "were of the same high quality as the rest of the Hurtigruten core product". However, according to the tour conductor, the group did not raise any formal complaint in order to get the excursion fee repaid.
Research question 2: The classification of the service failures: The challenge in answering the second question was to work out classification schemes which organised and systematised the service failure incidents in a convincing way. The work out of the service failure classification schemes consisted of two parts: First, the process of developing the main categories, and second, the inductive process of developing the sub categories. Following the approach of Edvardsson (1992), the analysis began with a careful scrutiny of the critical service incidents reported to the tour conductors and gradually a pattern emerged. As emphasised, the 51 service failure "stories" or "anecdotes" were identified and categorised within two main areas: 1. Service failure incidents on excursions 2. Service failures on the ship(s) while at sea or at harbour and The tourists on the Hurtigruten are offered eleven excursions during a round trip. Twenty two (43 %) of the service failures reported to the tour conductors occurred on excursions while twenty nine (57 %) happened onboard the ships while at sea or at harbour(s). Appendix 2 provides an overview of incidents which happened on excursions and on the Hurtigruten. The challenge in the inductive process was to develop a robust, reliable and comprehensive sub classification schemes within the two main categories. A content analysis of the service failure "anecdotes" or "stories" gradually revealed a "system" or "pattern" in the reported service failure incidents.

Service failures on excursions
The twenty two service failures which happened on excursions are categorised within five different sub groups. The groups will be described in turn.
Competency failures consist of service failure incidents caused by the service personnel due to shortage in core service competencies. An example is the service failure story no. 16. On several excursions, there is a limit on the total number of tourists to participate due to capacity restrictions. On an excursion to the Svartisen glacier, the tour conductor made a mistake by overbooking and two participants, the latest to register, could not participate. According to the tour conductor, there was nothing to do in the situation else than to apologise. The tourists got their money back but did not receive any compensation for the failure made.
Information failures consist of service failure incidents caused by incorrect information given to the travellers from the service provider. An example is the story no. 1. A tourist couple from Israel was especially looking forward to the North Cape excursion. However, they got very disappointed because the cliff was covered in fog and they did not see anything at all except the North Cape building. According to the tour conductor, the couple claimed that they would not have participated on the excursion if they had received appropriate weather forecast information.
Customer error failures consist of service failure incidents caused by the tourists' own shortcomings. An example is the story no. 10. A tourist couple from the US had booked an excursion from Molde to Kristiansund. See map for geographical orientation. When the bus was ready for departure, the couple did not show up. The couple became very angry because the bus did not wait for them and almost threatened the tour conductor in charge to reimburse the company.
Service failures in core delivery consist of service failures caused by a shortage in the core delivery. The service provider simply does not deliver the core product as promised. An example is the service story no. 2. A tourist couple participated on an excursion from one town to another by bus. The brochure informs that a dinner in a local restaurant is included. This did not happen on that occasion. The married couple was very disappointed and made a written complaint to management about the failure. The tour conductor in charge expressed the view that "such an incident should be avoided but even worse was that I was not involved in the decision made to leave out the local restaurant". Her view was that serving dinner in a local restaurant was an important part of the core product and was highly valued by the tourists.
The sub category other service failure incidents is a ragbag category and consists of service failure incidents on excursions that cannot be allocated to the other four groups. One example is the story no. 36. A tourist group from several countries participated on the Svartisen glacier excursion. They reported to the tour conductor that the excursion was overpriced and that the excitement of the excursion was not comparable to the price. According to the tour conductor, the price for this excursion had been raised by almost 40 % in two years.
The first main category: Service Failures on Excursions consists of five sub categories of which competency failures constitutes the largest one: about 40 % of the incidents. A summary of the research findings follows in Table 1 Service failures on the ships while at sea or at harbours. Twenty nine service failures happened on the Hurtigruten while at sea or at harbours. The failures are categorised into six different sub groups. They will be discussed in turn.
Customer needs and requests consist of service failure incidents where the tourists raise needs and requests to which the service employees have to respond. The "story" no. 7 illustrates this category. According to the tour conductor, the first interviewee, on a voyage with rather bad weather, some tourists felt that there was little to be occupied with on the ship. A small group of Germans approached her and asked for entertainment onboard, like folk music and local dance groups. However, according to the tour conductor, this was not planned for and the request could not be met.
Tourists' attitudes and behaviours consist of service failure incidents caused by the tourists' themselves. An example is the story no. 28, an incident which happened to a married couple from Germany. They were tempted by the gastronomic standards of the meals served in the ship's restaurant and simply ate too much. According to the tour conductor, the fourth interview, they incurred painful indigestion and could not fully enjoy the voyage due to sickness.
Two service failure incidents are categorised as what Lovelock and Wright (1999) call jay customers. The service failures are caused by the tourists themselves. The service failure story no. 40 previously described illustrates this category. The two "jay customers", two old ladies from Norway, caused severe service problems and challenges for the service personnel during the voyage. According to the tour conductor in charge, she was "completely exhausted when the two ladies left the ship in Bergen". The story no. 33 is similar to the story described except that the complainer was a man in his 50s from England who travelled alone. According to the tour conductor, the man argued that he had travelled on several international cruises before, and he claimed that the Hurigruten lacked luxury components like 24/7 personalised service. Due to the man's continuous complaining behaviour, he became a very demanding customer. According to the tour conductor, they could not satisfy the man as e.g. 24/7 personalised service is not a service standard on the Hurtigruten. Three service failures are service breakdowns caused by the service personnel onboard. An example is the story no. 24 about a German lady who usually ate dinner at the first seating in the restaurant and did not show up for dinner. However, according to the tour conductor, she came to the second seating and took a seat at a free table. The waiter in charge got angry because she did not come to the first seating according to agreement and that she just had "taken" a free table without "permission" from the waiter. The German lady was very embarrassed over the situation and, according to the tour conductor, later on sent a written complaint to the company about the waiter's rude behaviour.
Information failures consist of service breakdowns caused by the lack of or inappropriate information given by service personnel onboard the Hurtigruten. Two incidents are allocated to the category information failures. The "story" no. 45 illustrates this category. According to the tour conductor, a film team entered the Hurtigruten.
The film crew behaved as if they "owned" the ship; for example moving the tourists away from their seating in order to get the correct setting for filming. According to the tour conductor, no information was given to her and the tourists about the film team entering the ship and the consequences of their presence.
Service failures in core delivery consist of service failure incidents caused by a shortage in the core delivery. The service provider does not deliver the service product as promised. Four incidents are allocated to the core delivery category.

Discussion
The primacy of this part of the paper is to discuss how this case study on the Hurtigruten corresponds to previous research on service failures. As discussed in the literature part, two analytical approaches are used in categorising service failure incidents. The discussion below is coupled to the Bitner et al. (1994) classification scheme of four main categories as the analytical scheme is applicable also at the business level.
The first main category of service delivery system failures classifies service failures in the core delivery of the firm. An examination of the service failure list containing 51 service failure incidents reveals that 18 service incidents belong to this category. A couple of examples can be mentioned. When an excursion is overbooked this is due to a mistake made by a service attendant. The use of unprofessional guides on excursions is obviously a responsibility of the company, and when a bus collapses on an excursion this has to do with the core delivery. Changes and delays in the travel route also belong to this category. When An inspection of Appendix 2 exhibits the allocations made. These are:  service delivery system failures occur, the tourists do not receive value for money and the company fails to meet the customers' expectations.
The second main category, employee response to customer needs and requests, pertains to employee responses to customer needs and requests. An examination of the service failure list reveals that 16 failure incidents belong to incidents belong to the second main category. When the tourists request for the provision of TV channels, entertainment and library services, these incidents belong to the second main category. When the tourists miss the presence of higher ship officers on festive occasions and when ask for a change in menu, these incidents also belong to the second category.
Unprompted and unsolicited employee actions, the third main category, pertains to events and behaviour, both good and bad, that are totally unexpected by the customer. An assessment of the 51 service failure incidents shows that 8 incidents belong to this category. A couple of examples are: When a service attendant behaves in an impolite and rude manner, and a bus driver behaves in the same way on an excursion, they reveal unprofessional and unacceptable behaviour.
Problematic customer behaviour, the fourth main category, refers to customer behaviour and interaction that the attendants find difficult to handle. The service failure list returns 9 incidents that belong to the fourth category. A couple of examples are: When two ladies complain that the fish is "poisoned" this is evidently not true. In fact, the fish served onboard is excellent. And when locals drink too much alcohol embarrassing the tourists, they become a huge challenge to the service personnel. Table 3 below shows the correspondence between the research conducted on the Hurtigruten and the Bitner et al. (1994) study.
Consequently, the main classification scheme outlined by Bitner et al., (1990) and expanded with a fourth category by Bitner et al. (1994) suits the Hurtigruten 2001 study quite well. However, following the advice of Edvardsson (1992), the main classification scheme suggested in this research entails three main categories: Service failures on excursions; service failures on the ship(s) and other service failure incidents. The main categories in this investigation are comparable with the main groups suggested by Edvardsson's (1992) investigation of negative critical service incidents in an airline. The conclusion is that both approaches to categorise service failures are relevant as parameters for the research carried out on the Hurtigruten. However, the approach suggested by Edvardsson (1992) was favoured due to contextual reasons.

Implications and conclusions Theoretical implications
The present study examining service failures on the Hurtigruten has theoretical implications. First, the study evidences the robustness of the CIT methodology when examining a phenomenon about which relatively little is known (Gummesson, 2000). The procedural steps of the CIT have proved reliable when examining service failures in a cruise line context. Secondly, the use of a service employee perspective in the study supports previous empirical research which emphasises the important role of high contact service employees as important sources of information who can act as informants of "stories" of service failure incidents. Thus, the interactive role of front stage personnel is decisive for quality improvements and effective complaint management (e.g. Bettencourt and Gwinner, 1996). Thirdly, the complexity of many service processes, as on the Hurtigruten, supports the view that a classification scheme of service failures has to be contextualised, i.e. the scheme has to be developed according to how the services are delivered (Edvardsson, 1992). In the study reported, this implies the development of targeted service failure classification schemes. Finally, the study supports the view of customer information in general and service failure information in particular, as important sources of information for the purpose of services development and improvements (Gustavsen et al., 1999;Grönroos, 2006).

Implications for management
One lesson to be learned from the study is that shipping line management should continue to improve the quality of the service product. This is a challenging task when service processes are quite complex and involves a series of direct interactions between the tourists and the service personnel whose performances are likely to vary from one day to another. However, two areas seem to require special attention from management. First, the planning and execution of excursions, and second, the service operations in the ships' restaurants.
The planning and execution of an excursion is a rather complex task. Imagine the North Cape excursion which includes a bus trip to the plateau, eating a meal in an often overcrowded restaurant served by young inexperienced waiters, scrambling for a seat in the cinema to watch a video production, cueing in the souvenir shop and returning by bus to the ship after several hours. Obviously this highly complex service delivery process is demanding for any tour service provider. In order to avoid service failure incidents to occur on excursions two points seem important. First, the hiring and selection of service employees to front line positions, and second, the training of service attendants. The study reveals that in some cases service attendants are uncertain what to do when service failures occur. This implies that management has to address the issues of training and empowerment in to prepare the service employees for challenging service problems.
The ships' restaurants are places where, service failures are likely to occur. An inspection of the service failure list reveals that two distinct areas need inquiry. First is the hiring, selection and training service attendants. Second, is a thorough examination of the food preferences of the international travellers. If a service failure is caused by a rude and impolite service attendant this obviously has to do with weakness in the recruitment and training processes. And although the gastronomic trends move towards an international cuisine, management must remember that most of the tourists on the Hurtigruen are beyond their 50s and, according to the interviewees, are somewhat conservative in their food habits. As the food seems to be an important part of the core product, the menu should be given special attention by management in order to reveal changes in preferences.

Conclusions
This study has favoured a qualitative research methodology in order to identify, describe and analyse 51 service failure incidents on the Hurtigruten. The primacy of the research has been the work out of a comprehensive service failure classification scheme useful in a cruise line context by employing the Hurtigruten as an illustrative case. Two main categories and eleven sub categories of service failure incidents are suggested by approaching the data inductively. In the research literature induction is perceived at "the process of observing facts to generate a theory" (Ghauri and Grønhaug, 2002:14). However, this study does not suggest any theory on services failures applicable in industries across the service sector but makes a contribution to knowledge development on service failures in a specific context: within an explorer cruise line setting.
The research, however, may suffer from limitations and challenges. First, the use of a service employee perspective implies that the information reported on service breakdown is filtered. The tour conductors serve as intermediates of service failures incidents experienced by the tourists. In order to overcome this "filtering" problem, the tour conductors were contacted two weeks before the actual interviews and were encouraged to think through "stories" or "anecdotes" of service failure incidents which the tourists had reported to the. This cognitive approach was useful because the tour conductors were given sufficient time to recall and note down incidents before the actual conduct of the interviews. Second, the time of date collection may raise the question of the relevance of the data today. However, the time of the data collection is convenient for at least two purposes. First, the Hurtigruten had by the time expanded the fleet by new and "modern" ships which appeared like conventional smaller explorer cruise ships and had become a cruise line operator. Second, the incidents uncovered seem to be "universal" in the sense that they have relevance for the operations today. A round trip on the Hurtigruten at the time of the study is quite similar to a round trip as it is today. Third, the issue of quality in the research is an important one. There are several opinions on how to secure quality in quantitative studies (e.g. Johannessen et al., 2004). Veal (1997; discusses the two axioms of research quality, validity and reliability. Validity is perceived as "the extent to which information collected by the researcher truly reflects the phenomenon being studied" (Veal, 1997:35). The phenomenon is service failure incidents in a cruise line context and the validity of the research findings is closely linked to the recall and report of the incidents from the tour conductors. As they were given much time to recall and rethink incidents reported, the information collected provides a "pattern" or "system" of service failure incidents in that specific context. Reliability is by Veal (1997:35) perceived at "the extent to which research findings would be the same if the research was to be repeated at a later date with a different sample of subjects". However, as the intention of a single case study "is not to make generalisation but to investigate a "one off situation" (Clarke et al., 1998:103), the case study contributes to the knowledge field of service failure but is restricted to a specific context. This implies that more research on service failures is needed focusing on the cruise line industry as a whole.