Int. Journal of Business Science and Applied Management, Volume 3, Issue 1, 2008
Customer protest: Exit, voice or negative word of mouth
Bernt Krohn Solvang
Department of Work Life and Innovation, University of Agder
Service box 509, 4898 Grimstad, Norway
Tel: +47 37253147
Email: Bernt.K.Solvang@uia.no
Abstract
Of the three forms of protest the propensity of word of mouth (WOM) seems to be the most common,
and the most exclusive form of protest seems to be exit. The propensity for voice lies in between. The
costs linked to voice influence the propensity for WOM. The customers seem to do an evaluation
between the three forms of protest, yet the rational picture of the customers should be moderated.
Leaders should improve their treatment of the customers making complaints. The more they can treat
customer complaints in an orderly and nice way the less informal negative word of mouth activity they
will experience and they will reduce the exit propensity and lead the customers to the complain
organisation. They should also ensure that their customers feel they get equal treatment by the staff.
Keywords: voice, word of mouth, WOM, exit, satisfaction, loyalty
Bernt Krohn Solvang
15
1 INTRODUCTION
The customer’s potential to complain or make positive comments is hidden from the shop prior to
the purchase being made (Brief, 1998). This potential is of considerable significance for shops in a
market where there is competition and where keeping customers is of the greatest importance (Aaker,
1991; Fornell, 1992). Hirschman (1970) presents two main forms of protest: protest to the shop or to a
public complaints body (voice) or changing shop (exit). To complete the analysis of protest behaviour
we include the third form of protest, Word of Mouth (WOM), a complaint to friends and acquaintances .
Figure 1: Customer loyalty and freedom of choice. Developed based on the work of Hirschman
(1970).
The customers can choose among these forms of complaint, and according to Hirschman’s theory
(1970) it is the costs and potential gains of the two alternatives that decide which is chosen. The costs
of exit are connected to access to alternatives and to the degree of loyalty (Hirschman, 1970; Singh,
1991). Even though Hirschman’s theory looks at the relationship between the two forms of protest, an
empirical study of the three main forms is lacking. However, studies have been done on the relation
between voice and WOM (Bearden and Oliver, 1985; Richins, 1983; Singh, 1990b; Ping, 1997; Naylor
and Klaiser. 2000). Ping (1997) has considered the relation between satisfaction, exit costs and
complaint behaviour. We want to consider the relation between these three types of complaint:
• protest to the shop (voice) or a complaint to friends and acquaintances (WOM) or exit.
• What is the relation between these three types of protest?
• What influence the customer’s choice of protest method?
2 THE THEORETICAL PERSPECTIVE
Hirschman (1970) is focusing on the situation of choice when a customer is dissatisfied. A
dissatisfied customer could choose between various forms of protest methods as voice (complain to the
supplier, exit (leave the supplier for another one) or WOM (talk negatively to friends an acquaintance).
Hirschman (1970) did not treat WOM, but we include that form of protest here in order to obtain a
complete picture. This illustration of Hirschman’s (1970) theory shows three forms of protest. Exit,
voice and WOM are customer protests if they become dissatisfied with the delivery from the company
in relation to their own efforts to achieve what they want. According to Hirschman’s theory the loyalty
is a key variable. High degree of loyalty will increase the costs linked to exit. An alternative form of
protest (voice) is thus more likely. Accordingly the voice protesters are expected to be more loyal than
the average in the customer, as “the like hood of voice increases with the degree of loyalty”
(Hirschman, 1970: 77). Hirschman (1970: 35) sees the protest form of voice as a complement to exit
Loyalty
Exit
Voice
The
competitors
The company
The
customers
choice.
WOM
Efforts
Efforts
Int. Journal of Business Science and Applied Management / Business-and-Management.com
16
and not as a substitute for it. According to this theory exit is associated with costs and gains as the form
of voice is. The costs linked to exit might be of emotional character and travel distance and price and
quality variation. The voice costs are linked to bargaining power and by that to education.
We are focusing on the connection between these forms of protest since, to our knowledge these
connections have not been treated empirically in a study.
The form of protests
Dissatisfied customers may react in various ways that often manifest itself through frustration and
anger: for instance to go to the representative of the shop, respond in private by means of negative
comments to friends and acquaintances, or go to a formal complaints body (Richins 1983; Singh,
1990b; Singh, 1990; Singh, 1991; Strauss, Schmidt and Schoeler 2005). Singh (1990) divides
customers into four groups according to their pattern of response. The passive ones have a low score
for all types of reaction.
Only a few of those who are dissatisfied, make themselves heard, (Teknologisk Institutt 1993;
Grønhaug 1977). Andreasen and Best (1977) reported that more than half of those who were
dissatisfied did not complain at all, while according to Brief (1998) only 20% of Americans
complained in response to unsatisfactory service. The tendency to complain is, then, a function of
insufficient satisfaction and of frustration behaviour (Strauss, Schmidt and Schoeler 2005). Many of the
customers who are dissatisfied do not complain directly to the shop. Several authors, (Andreasen and
Best, 1977; Tarp, 1986; Grønhaug and Gilly, 1991; Hernandez et al., 1991; Bearden and Oliver, 1985;
Richins, 1983) see this in connection with the possibility of winning one’s case against the costs of
complaining in the best Hirschman tradition (1970). Berry and Parasuraman, (1991) state that
customers have a zone of tolerance where a performance that lies within the zone will be accepted.
Performances exceeding the zone create delight and loyalty, while performances falling short of the
zone create dissatisfaction.
Will loyalty be able to create an increased zone of tolerance or is the degree of loyalty not
significant for the zone? Will the zone of tolerance be able to reduce the tendency to protest?
Exit
The main behaviour in exit is to leave the shop and start being a customer in another shop. This
behaviour has costs and gains. The exit costs are related to access to alternatives and to the degree of
loyalty. Hirschman (1970) argues that the exit costs are higher in those cases where there is no
alternative shop. If the customer is convinced that complaining will be effective that could delay exit
(Hirschman, 1970: 37). Customer loyalty will work as a barrier against exit. The barrier may be
compared to a cost (‘protective tariffs’) (Hirschman, 1970:79). Exit propensity means the probability
of a customer choosing the exit option.
How is the exit propensity affected by the degree of loyalty? When loyalty increases, we expect
that the exit propensity will decrease since the costs of exit increase with increasing loyalty. How does
exit stay as a protest form in relation to voice and WOM? Do the customers have equal access to each
form of protest?
Voice
Voice works as a supplement to exit and not as a replacement for it (Hirschman, 1970: 35). A
decreasing degree of satisfaction and increasing exit costs may seem to play an equal role in
encouraging complaint behaviour (Ping, 1997). Those who are loyal are over-represented among the
complainers if Hirschman’s theory (1970) holds. The complainers score high on complaining and low
on the other forms of reaction. Those who are angry score high on comments to friends and
acquaintances (WOM), while the activists score high on complaining and high on third-party action
(consumer bodies) (Singh, 1990). Other factors which should be considered in an analysis of complaint
behaviour are understood risk, confidence in the success of the complaint as well as the advantages and
costs of complaining (Bearden and Teel, 1983; Tarp, 1986; Richins, 1985; Andreasen, 1997).
Grønhaug and Gilly (1991) suppose that the greater tendency to complain about the service
industries may be linked to the fact that the services are difficult to standardise. Grønhaug (1972) finds
that consumers with a high risk evaluation have a greater tendency to make use of consumer-related
sources of information, while those with a lower risk evaluation make greater use of market-dominated
sources.
When focusing on technology based service encounters Snellman and Vihtkari (2003) find that
customers who actually consider themselves guilty for the outcome were the most frequent
complainers, while the ones attributing the outcome to technology failures or service process failures
complain less often. Online customers are less likely to complain than offline customers and online
Bernt Krohn Solvang
17
customers are more sensitive to benefits/costs of complaining. The difference could be explained by a
difference in personal competence expressed where the offline customers have highest score (Cho,
Hiltz and Fjermestad 2002).
Comp lain, protest and avoidance have also been seen as negative effects of loyalty programs
(Strauss, Schmidt and Schoeler 2005).
Each fourth of the potential complainers do not complain. The main reasons for this are linked to
perceived costs of complaining as time and efforts (Voorhees, Brandy and Horowitz 2006). According
to Grønhaug and Zaltman (1981), making a complaint is positively linked to experience, education and
income, but negatively linked to age.
Voice handling
Poor handling of a complainer who chooses to complain instead of changing shop because he has
a certain loyalty will weaken the complainer’s faith in the supplier. This results in fewer satisfied
customers and reduced loyalty. The risk of exit and a reduction in repeat purchase increase, together
with the increased probability of negative private comments (WOM) (Bearden and Oliver, 1985;
Grønhaug, 1987; Richins, 1983). Griffin (1995: 191) points out those complainers who have obtained a
quick solution have a repeat-purchase tendency of 82%, in contrast to those who have experienced a
major problem without complaining and whose repeat-purchase tendency is 9%. Those who complain,
irrespective of the result, have a repeat purchase figure of 19%. Gilly and Hansen (1985) point out that
effective complaint handling results in customer satisfaction and loyalty.
We must suppose that many of the complainers are loyal customers. They choose to complain
instead of changing shop because their loyalty has increased the costs perceived in changing shop. On
the other hand a greater zone of tolerance among the loyal customers may keep them from
complaining. Good handling of this type of complainer will strengthen the complainer’s faith in the
supplier: ‘only moderate degrees of satisfaction with service recovery are needed to restore future
repurchase intention’ (Andreassen, 1997: 195; Singh, 1990b; Gilly and Hansen, 1985).
Good complaint handling results in satisfaction and increased loyalty, and reduces the probability
of negative private comments (WOM) (Bearden and Oliver, 1985; Grønhaug, 1987; Richins, 1983).
This in turn reduces the risk of exit and increases the probability of repeat purchase. Increased
probability of repeat purchase means a better financial result for the supplier. Calculations show that an
increase of 5% in the repeat purchase share from 60% to 65% increases receipts by 15%. ‘On the other
hand a fall in customer loyalty from, for example, 90 to 80 will result in future sales being halved.’
(Andreassen, 1997: 4) This is also shown by Oliver (1997, pp. 368-369). A better financial situation
helps the supplier to satisfy complainers. A weaker situation makes it more difficult for the supplier to
offer good complaint handling.
Negative WOM
Bearden and Oliver (1985) found that a higher potential loss stimulates various forms of
complaint, and that the extent of private complaint behaviour is inversely linked to satisfaction with the
response from the firm. They point out that if the organisation makes a mistake in its complaint
handling, this may lead to loss of goodwill and negative WOM. Grønhaug (1977) pointed out that the
complaints seem to build up round complex products which involve a high risk.
Richins (1983) found a connection between the consumers’ evaluation of the complaint handling
and comments about the shop. The more negative the complaint handling expected by the complainer,
the greater the probability of negative private comments (WOM). In another work Singh (1990b) points
out that exit and negative WOM are linked to an evaluation of the probability of the complaint being
successful. But Naylor and Kleiser (2000) do not find any effect of earlier complaint handling on
negative WOM. No complainers are less likely to engage in negative word of mouth than the
dissatisfied and recovery groups (Voorhees, Brandy and Horowitz 2006).
Some of the protest forms turn out the public against a firm that has wronged them. Protests
published at the Internet are rooted in injustice, identity and turn out as a personal grievance into a
“cause” worthy of public attention and support (Ward and Ostrom 2006).
Customer satisfaction
Customer satisfaction and dissatisfaction are associated with the expectations of the customer. If
high expectations are met, the customer will be satisfied, but if low expectations are not exceeded by
the delivery the customer will be dissatisfied (Oliver, 1997). The customer’s experiences could be
linked to various sources as service performance, product quality, transactions, product delivery and
other factors (Zeithaml, Parasuraman, Berry, 1990).
Int. Journal of Business Science and Applied Management / Business-and-Management.com
18
Churchill, Gilbert and Surprenant (1982) found possible effects of satisfaction dependent on
product characteristics (durable and non-durable). Whilst Snellman and Vihtkari (2003) do not find any
difference in complaining rate between customers in retail banking and traditional technology based
service encounter, while Oliver (1997) finds a greater tendency to complain about durables than about
non-durables, but the largest group is those who do not want to complain. This can also be linked to the
significance of design, which plays a central role for durables.
Grønhaug and Gilly (1991) point out that customer dissatisfaction can be connected with lack of
confidence concerning the transaction, and that much of the dissatisfaction could be linked to market-
institutional circumstances beyond the seller’s responsibility, such as no product delivered or a long
delivery time. Grønhaug and Zaltman (1981) find that it was the transaction frequency, and not the
qualities of the good, which best explains the variation in customer dissatisfaction. Ping (1997)
maintains that the tendency to complain is related to satisfaction and involvement in the relationship, in
the case of firms. Richins (1985) shows a positive connection between potential financial loss and the
tendency to complain.
However, there is no simple connection between satisfaction and loyalty. Even satisfied customers
can switch to another store because there is no one-to-one connection between satisfaction and loyalty.
The relation between satisfaction and loyalty is influenced by characteristics of the consumer such as
age and income (Homburg and Giering 2001). Bloemer and Kasper (1995) and Bloemer and de Ruyter
(1998), differentiate between two types of satisfaction. Manifest satisfaction conveys a customer who
has made a deliberate choice and has reached the conclusion that he/she is satisfied with the offer.
Latent satisfaction expresses an unconscious customer who has not compared the offer with other
suppliers. They find that an increase in the manifest satisfaction has a greater impact on customer
loyalty than an increase in the latent satisfaction.
Customer loyalty
In literature concerning consumer behaviour there are different approaches to view/define
customer loyalty. It is differentiated between consumer loyalty in the consumer goods market,
customer loyalty in the business-to-business market and the synthesis between consumer and customer
loyalty (Kotler 1987; Oliver, Rust and Varki 1997).
The loyalty phenomenon is characterized by diffuse and vaguely delimited contents of meaning
(Jacoby and Chestnut, 1978; Peter and Olson, 1993; Dekimpe and Steenkamp 1997). Hirschman’s
loyalty concept is equated with “non-exit” and hence it is too simple (Huefner and Hunt 1994). In
addition to being an unclear concept, several researchers have equated loyalty with repurchase (Carman
1970; Wind, 1978; Grønhaug and Gilly 1991). It is indicated that customer loyalty and repurchase can
be increased through establishing barriers that make it more difficult for the customer to go to another
store, and consequently repurchase increases (Aaker, 1991; Selnes and Reve 1994; Andreassen and
Bredal 1996).
Loyalty as a development pattern in phases: This concept in particular has given inspiration to our
approach. Oliver (1997) presents customer loyalty in the form of four Loyalty Phases, viewed as steps
of a loyalty ladder:
• Step 1 Cognitive loyalty – The customer has favourable knowledge of the supplier, but a
better offer will result in exit to the competitor. The loyalty is only based on cognition.
• Step 2 Affective loyalty – is an emotional attitude based on cognitive loyalty.
• Step 3 Conative loyalty – is intentional loyalty that includes a deeply felt obligation to buy.
• Step 4 Action loyalty – a determination to defy any obstacles in order to buy (Oliver, 1997:
392-393).
This seems to be a fruitful approach to this diffusing concept.
Research questions
1. How are the various forms of protest distributed?
2. How is the variation in satisfaction and loyalty distributed in each form of protest?
3. How could the variation in the propensity for each of the protest forms best is explained by
customer related variables?
4. How do external factors as competition and type of shop branch influence the factors
associated with each form of protest?
Bernt Krohn Solvang
19
3 METHODOLOGY
Sample of shops
We choose a quantitative design in order to be better able to answer our research questions. The
framework for the sample consists of four shops in the southern region of Norway, two in the grocery
trade and two in the furniture trade. For each shop 100 customers were selected, a sample of 400
customers altogether. In the case of the grocery shops the interviews were carried out outside the shops
on a Saturday and a Tuesday in October 1998. The sample of grocery customers was thus selected out
of convenience (those who came out of the shop).
In the case of the furniture shops the plan was to carry out the interviews in the shop. However,
because of a shortage of customers, a random sample of customers was selected from the shops’
customers list. The interviews were conducted by telephone. The Saturday customers were collected in
one group since customers on this specific day of the week can have a different shopping pattern with
several family members taking part.
The four shops differed on two criteria: type of trade to get variation in risks for the customers
(grocery and furniture) and competition situation and by that getting higher variation in the exit costs.
Consequently we included two grocery shops, one in a highly competitive area (low exit costs) and one
in a less competitive area (higher exit costs), and two furniture shops, one in a highly competitive area
and one in a less competitive area. All four shops are members of retail chains.
Definition of and Measurement of central variables
Loyalty
First we tried to establish an index variable based on loyalty as an attitude and a repurchase
indicator: the percentage share of the respondent’s own trade in that type of shop for the shop in
question. This index variable was not reliable since Cronbach Alfa came out under 0.7 (Hair,
Anderson, Latham and Black, 1998).
Then we established an index variable based on loyalty as an attitude and an indicator of an
emotional variable: To what extent the respondents would recommend the shop to others if they were
asked for advice. This indicator of affective loyalty come out with a significant Cronbach Alfa 0.70
(N=396). Consequently our indicator of Affective Loyalty is measured like this:
Measurement of loyalty
a) Self-evaluation of loyalty to the shop in question on a scale from 0 (extremely low) to 10
(extremely high).
b) Self-evaluation of to what extent the respondents would recommend the shop to others if they
were asked for advice on a scale from 0 to 10.
A reflective index (Troye, 1994) was worked out on the basis of these two indicators by the sum
(a+b). In a reliability analysis Cronbach Alfa came out with 0.70 (Hair, Anderson, Latham and Black,
1998). This indicates satisfactory reliability. The customer loyalty variable is then measures in values
from 0 to 20.
Satisfaction
Customer satisfaction comprises the opinion about the positive response in the exchange and the degree
of satisfied expectations (Andreassen, 1997).
Satisfaction was measured as follows:
a) Self-evaluation of satisfaction with the shop in question measured on a scale from 0 to 10.
b) Self-evaluation of the perceived balance between the costs related to being a customer in the
shop in connection with costs involving money and time, and the feeling of what one is left
with in return for these costs, measured on a balance scale from 0 to 10.
The sum of a) and b) make up our index variable for satisfaction, a reflective index measurement
(Troye, 1994). Coronach Alfa between these two indicators is 0.861, which indicates high reliability.
The satisfaction index is measured in values from 0 to 20.
Interaction between loyalty and satisfaction
By multiplying the two variables satisfaction and loyalty we got a new variable representing the
interaction between them.
Int. Journal of Business Science and Applied Management / Business-and-Management.com
20
Service quality
Zeithaml, Parasuraman, Berry (1990) presented five dimensions in their Service Quality Concept.
We have indicators to include three of these dimensions in our study. These are the following
dimensions:
a) Reliability (ability to perform the promised service)
b) Responsiveness (Willingness to help customers and provide prompt service)
c) Assurance (Knowledge and courtesy of employees and their ability to convey trust and
confidence)
As an indicator on Reliability we used respondent evaluation of the shop on how polite they found
the employees in the shop on a scale from 0 to 10. As an indicator on Responsiveness we used
respondent evaluation of the shop on willingness to serve you, they assessed the employees in the shop
on a scale from 0 to 10. As an indicator on Assurance we used respondent evaluation of the shop on the
level of relevant knowledge they assessed the employees in the shop on a scale from 0 to 10. We made
our index variable, service quality, by first running a factor analysis of theses three indicators (Principle
Component Analyses).
The component Matrix comes up with one component.
Table 1: Factor analysis of service indicators. Component Matrix
Polite staff (reliability) 0.900
Willingness to serve (Responsiveness) 0.901
Knowledge (Assurance) 0.800
N 396
The three indicators are all in compliance with a common factor we will call service quality. We
then performed a reliability analysis to see if these three variables could be joined together in an index
variable. Conbach’s ALPA=0.84. This indicates high reliability if we make an index variable consisting
of the sum of these three variable. Consequently this index variable is our service quality variable with
values from 0 to 30.
Exit costs: Self-evaluation of perceived costs in changing shop measured on a scale from 0 to 10.
Propensity to exit: Propensity to exit is a self-evaluation of the probability of the customer
continuing to use the shop in question. Those answering very likely or likely were given the value 0 for
the variable tendency to exit, while those answering fairly unlikely or unlikely and do not know were
given the value 1 for the variable propensity to exit. The group average is between 1 and 0 and is
interpreted as the propensity to exit for the group.
Voice costs: Self-evaluation of the costs related to complaining on a scale from 0 to 10.
Propensity to voice: Self-evaluation of the propensity to complain measured on a scale from 0
(have never complained to the shop) and 1 (have complained once or several times to the shop). The
group average lies between 1 and 0 and is interpreted as the propensity to complain for the group.
Negative Word of Mouth (WOM): Self-evaluation of to what extent one complains to friends
and acquaintances rather than to the shop measured on a scale from 0 to 10.
Propensity to WOM: Self-evaluation of the propensity to WOM measures on a scale from 0 to 1.
Those who found WOM actual or very actual we defined as high propensity (1) and those who found
WOM little or not actual as low propensity (0).
Experience with complaining: The method chosen was self-evaluation of how the complaint was
received and handled. 23% of the respondents had experience with complaints to the shop. They
answered according to these categories: bad (1), less good (2), satisfactory (3), good (4) and extremely
good (5).
Discriminated treatment: Self-evaluation of perceived discriminated treatment measured on a
scale from 0 to 1. “Some times we may feel that other customers are getting better treatment than
ourselves. To what extent is such discriminated treatment happing here?” Those who answered “it
Bernt Krohn Solvang
21
happens often” and those who answered “now and then” and those who answered “seldom” were all
given value 1 and those who answered “never “were given value 0.
Perceived risk linked to the shop: Self-evaluation of risk linked to the customers’ shop. To what
extent do you feel a risk by doing your shopping at this outlet? Scale from 0 to 10 where 10 are
measured as extremely high risk. This question was only presented to customers from the furniture
shops since the risk linked to grocery shopping is considered low.
Shopping frequency: How many times have you done your shopping in this outlet the last 4
weeks?
Age: The age of the respondents in years.
Education: The number of years of education after primary school.
4 ANALYSIS OF THE RESEARCH QUESTIONS
How is the various form of protest spread among the customers?
What are the portions of the various forms of protest? Our data from this research might give an
idea.
Table 2: the distribution of propensity and costs for each form of protest.
Exit Voice WOM
The propensity for 0.1 0.3 0.4
Costs linked to each protest form,
scale from 0-10.
2.3 3.0 42% answered actual
and very actual
N 380 396 396
According to these data, the propensity of Exit is lowest, of WOM protest highest and Voice in
between. Dissatisfied customers would choose to complain to friends and family four times more often
than make an exit and three times more often than to make “voice” to the shop. More than each three of
the customers in this sample have not been engaged in any form of protest. On the other hand only 2%
of the respondents have been engaged in all three forms. Of those with two protest forms, the ones with
Voice and WOM constitute the largest group (13%).
We do not have WOM costs measured in the same way as Exit and Voice, but the costs of WOM
seem to be low. The subjective costs linked to the other two forms are small and comparable.
How are the customers distributed on various combinations of protest forms?
Table 3: the distribution of combinations of the protest forms.
Protest form % N
No protest form 36 380
Exit 10 380
Voice 31 396
WOM 42 396
All three forms 2 380
Voice and WOM 13 396
Vo ice and exit 5 380
Exit and WOM 5 380
More than each three of the customers in this sample have not been engaged in any form of
protest. On the other hand only 2% of the respondents have been engaged in all three forms. Of those
with two protest forms, the ones with Voice and WOM constitute the largest group (13%). More than
each three of the sample does not make any protest at all.
Int. Journal of Business Science and Applied Management / Business-and-Management.com
22
Is there any association between the various forms of protest?
Are the resources favourable for each protest form accumulative or following a Matthew effect so
that those who have, shall have more and those who have less shall loose what they have (Merton,
1968)?
Table 4: Correlations between the propensities for protest forms
Propensities WOM Exit Voice
WOM - 0.04 0.00
Exit 0.3 - 0.12*
Voice 0.00 0.12* -
N 396 394 396
* Significant at 0.05
The only significant association between the propensities for protest is the association between
voice and exit. WOM has no significant correlation with the other two forms of protest.
WOM do not fit in with the pattern linked to voice and exit. Voice and Exit, however, are fitting in
well in the same dimension as we may call “formal active protest”, while the informal form of protest
of WOM do belong in another dimension which we may call “informal active protest”.
We do a small sociological analysis of each protest form in order to look for possible explanations
of the difference between the formal and informal form of protest.
Is there variation in sociological characteristics between these three forms of protest?
The similarities between the protest groups are more striking than the differences. The WOM
group and none protest have got the lowest degree of education, but the difference is not significant.
These results indicate that the active forms of exit and voice are linked with educational level.
Table 5: Age, education and sex within each form of protest.
Protest form Age Education after
primary school
Sex Sex Sample size
Mean years Mean years % Men % Women N
WOM 39 4.5 40 60 165
Voice 42 4.9 42 58 122
Exit 38 4.8 26 74 51
Non- protest 43 4.4 40 60 144
Table 5 shows no significant differences between the various forms of protest. The voice form of
protest has highest age, education and highest portion of men, but the differences are not significant.
How is the satisfaction and loyalty distributed in each form of protest?
Behind any form of protest there is some sort of dissatisfaction. The dissatisfaction and what
creates it could be linked to a lot of factors and unfulfilled expectations (Oliver, 1977).
Table 6: Satisfaction and loyalty in each protest propensity group.
Protest
propensity
Mean
satis-
faction
F value
on the
difference
between
(1) and (0)
Sig. Mean
loyalty
F value
on the
difference
between (1)
and (0)
Sig. N
Exit (1) 10.3 63.6 ** 6.1 39.4 ** 50
Exit (0) 13.8 11.3 326
Voice (1) 12.0 31.9 ** 9.4 10.2 * 116
Voice (0) 14.0 11.2 254
WOM (1) 12.6 12.8 ** 10.1 2.6 - 159
WOM (0) 13.9 11.0 211
None protest (1) 14.7 38.0 ** 12.0 16.1 ** 132
None protest (0) 12.5 9.8 224
** S<= 0.01
* S<=0.05
Bernt Krohn Solvang
23
Table 6 shows for all protest propensity groups, the customers without experience with the protest
form (with values 0) have highest score on satisfaction and loyalty. The differences are most profound
in the exit group. The differences in value on satisfaction and loyalty are all significant except for
degree of loyalty in the WOM group.
Exit seems to be the most potent form of protest with marked differences between those with exit
experience and those without. WOM experiences do not influence the degree of loyalty in any
significant way. The none protesters (with value 1) have both higher degree of satisfaction and loyalty
than the protesters (with value 0) on the none- protest variable.
What is the association between protest propensity and loyalty? According to Hirschman’s theory
(1970) voice propensity could be associated with high degree of loyalty and exit propensity would be
associated with low degree of loyalty since a high degree of loyalty would tend to prevent the customer
from exit. Table 6 seems to fit nicely to Hirschman’s theory (1970). The lowest degree of loyalty in the
group of high propensity for exit and the degree of loyalty in the two other groups are marked higher.
How coul d the variation in the propensity for each of the protest form best be explained?
We will perform a series of logistic regression with each protest form as dependent variable and
the theoretical based variables as independent. The results are listed up in Tables 7-12. We include an
interaction variable between loyalty and satisfaction with a view to survey interaction effects. We use
an exploratory approach since there are a lot of studies linked to each of the forms of protest.
Table 7: theoretical factors that might influence each form of protest
Theoretical factors Exit Voice WOM
Affective loyalty X X X
Satisfaction X X X
The interaction between satisfaction and loyalty X X X
Costs linked to the protest form X X
3
Service quality X X X
Transaction frequency X X X
Perceived risk linked to the shop
1
X X X
Voice experience in separate analysis since only 93 respondents had
experience
2
X X X
Discriminated treatment X X X
Age X X X
Education X X X
Sex X X X
1
This question was only asked to Furniture respondents and the variable is used in a special analysis.
2
Only a ¼ of the respondents had any complaining experience with the shop they left when interviewed so this
variable is studied in special analysis.
3
We have no variable describing how difficult the customers felt it was to talk to friends and acquaintances.
We will sum up the factors that could influence the forms of protest. Subsequently we will run
logistic regression and sum up with the significant variables for each form of protest.
Int. Journal of Business Science and Applied Management / Business-and-Management.com
24
The exit form of protest
We have seen the exit form of protest as the most exclusive one. How could we best explain the
variation in the propensity for exit? Our start model is based on variables in Table 7. The significant
model is presented in Table 8.
Table 8: Logistic regression with propensity to exit as dependent variable
Propensity to: Exit Significant test:
Wald statistics
ß
Satisfaction -.31**
12.7
Affective loyalty -.16** 7.6
Shopping frequency -.16** 7.1
Initial – 2 LOG likelihood 220.6
Model – 2 LOG likelihood 144.8
Difference 75.8
Significance for model P<.001
Nagelkerke R
2
.42
Prediction ability 91%
N 396
** S<= 0.01
Exit propensity could partly be explained by dissatisfaction, low degree of loyalty and low
frequency visit in the shop. Table 8 shows the association when satisfaction, loyalty and shopping
frequency increase the propensity for exit decrease. High shopping frequency seems to have a
preventive effect on exit propensity. This finding fits nicely to Hirschman’s (1970) theory. Loyalty and
satisfaction creates costs for the customers preventing them from making exit from the shop. The
model is significant and explains 42% of the variance leaving 58% for other factors and explanations.
In a special analysis of the customers with voice experience, we find a tendency showing “the
better the treatment of complaining customers, the fewer propensities for exit”.
The difference in evaluation the complain treatment between those without exit propensity (0) and
those with exit propensity (1) is 4.0 and 3.1 (N=92, F=7.0, Sig.=0.009).
In another special analysis of respondents linked to furniture shops, we could estimate the possible
effects of risk linked to shopping in the shop were the customers were interviewed. The risk evaluation
was done on a scale from 0 to 10. Average evaluation of risk was 2.1 (N=199). Those with low exit
propensity (0) had an evaluation score on 1.9, and those with high exit propensity (1) had an evaluation
on 2.8, N=190, F=4.0, Sig.=0.005. There seems to be a tendency that increasing risk evaluation is
linked to increasing exit propensity.
Bernt Krohn Solvang
25
How does external variables as competition and type of shop branch (external variables)
influence the factors associated with each form of protest?
Exit propensity influenced by external variables
We will trace possible effects of environment factors such as competition and of branch on the
factors explaining the variation in each form of protest.
Table 9: Effects of degree of competition and of branch on propensity to exit. Four analytical
models
Degree of
competition
Branch
Low High Grocery Furniture
Exit propensity 0.14 0.12 0.10 0.16
Factors explaining
variation in exit propensity
ß ß ß ß
Affective loyalty -.31** -.49** -.33** -.43**
Shopping frequency -.47** -.16** -.24**
Age -.10**
Initial -2 log likelihood 125.3 97.9 118.7 108.8
Model -2 log likelihood 93.0 52.1 94.4 54.1
Difference 32.3 45.8 24.3 54.7
Nagelkerke R
2
0.33 0.54 0.26 0.59
Percentage correct
predicted
88 94 90 94
N 179 200 197 199
** S<= 0.01
The effects of competition
When the competition increases the exit costs are reduced. Moreover, the quality of the offer from
the shops could be increased by the competition. When we compare the factors in Table 8, we find a
“better” model for explaining exit propensity when competition is high with some negative effect of
age reducing the propensity for exit. Shopping frequency seems to be more important in a competitive
environment and loyalty and satisfaction seem to reduce the propensity for exit both when the
competition is high and when it is low. When the competition is low there is an effect of loyalty, in
high competition the effect is linked to satisfaction. Does low degree of competition promote positive
attitudes towards the shops?
The effects of branch
Exit propensity seems to be higher in furniture shops then in grocery shops. The customers’
dependence of the shops might be higher for the grocery shops since they are more frequently visited
than furniture shops. The difference between grocery shops and furniture shops is linked to shopping
frequency which is a more important variable for grocery shops reducing propensity to exit. We tried to
include the risky variable in the furniture shop model, but it turned out to be not significant. The
satisfaction variable is a potent variable in both types of shops. In the furniture shops positive loyalty
attitudes seems to reduce the propensity for exit.
When competition is low, and for shops with lower visit frequency (furniture shops), the loyalty
seem to play an important role in preventing exit.
The four models in Table 9 are all significant. Models for furniture shops and shops in a
competitive environment have the strongest explanatory power.
Voice propensity
We start the study of variance in voice propensity with all the theoretical variables listed in Table
7. The final significant model for voice propensity is shown in Table 10. We do a separate analysis of
the customers with experience from previous complains.
Int. Journal of Business Science and Applied Management / Business-and-Management.com
26
Table 10: Logistic Regression with the propensity for voice as dependent variable
Propensity for voice Voice ß Significant test:
Wald statistics
Satisfaction -.19** 28.3
Age .02* 4.9
Initial – 2 LOG Likelihood 464.7
Model-2 LOG Likelihood 431.8
Difference 32.9
Model significance P<.001
Nagelkerke R
2
.12
Prediction ability 71%
N 396
** S<= 0.01
* S<=0.05
The propensity for voice is influenced by the satisfaction variable. The negative influence of the
satisfaction variable fits with Hirschman’s theory (1970), but the relatively weak effects could reflect
the effect of the theory of Zone of Tolerance (Berry and Parasuraman, 1991). Customers with high
loyalty refrain from making voice more often than customers with a lower degree of loyalty. An
increase in age increases the propensity for voice. Age is a resource for voice. The Logistic model is
significant and it exp lains only 12% of the variation in the dependent variable.
In a special analysis of the respondents with voice experience, we find the same tendency as we
found concerning exit propensity, but with opposite direction. The better the treatment of a
complaining customer, the more increased propensity for voice we have. Those who had not
complained to the shop had an average on treatment of 3.0, whilst those who had complained to the
shop had an average on 3.9 (N=93, F=2.2, Sig.=0.15). However, the difference is not significant.
In another special analysis of the respondents in the furniture shops, we studied the possible effect
of risk linked to do shopping in the actual shop on voice propensity. Those with low voice propensity
(0) had a risk evaluation on 1.8, whilst those with high voice propensity (1) had a risk evaluation on
2.7, (N=192, F=8.8, Sig.=0.003). The propensity for voice seems to proportional related to risk
evaluation, the higher risk evaluation the higher voice propensity.
Possible effects of external factors on voice propensity
We will see how these internal customer related factors are influenced when we differentiate
between high and low degree of competing environment for the shops and between grocery (with low
risk) and furniture (with higher risk) shops.
Table 11: Effects of branches and competition on the propensity to voice.
Degree of
competition
Branch
Low High Grocery Furniture
Voice propensity 0.30 0.34 0.34 0.29
Factors explaining
variation in voice
propensity:
Satisfaction -.18** -.21** -.20** -.19**
Age .03* ,03*
Initial – 2 LOG Likelihood 222.4 247.4 236.5 229.2
Model -2 Likelihood 204.9 228.3 217.6 215.1
Difference 17.5 19.1 18.9 14.1
Model significance P<0.001 P<0.001 P<0.001 P<0.001
Nagelkerke R
2
0.13 0.13 0.14 0.10
Percentage correct
predicted
73 73 67 74
N 179 194 184 186
** S<= 0.01
* S<=0.05
Bernt Krohn Solvang
27
Possible effects of competition on voice propensity
The effects of competition on voice propensity seem to be linked to one factor; satisfaction. Table
11 indicates that the higher the satisfaction the lesser the propensity for voice. When competition is
low, age could be a resource for voice propensity. The two models linked to competition are very week
and unable to explain much of the variation in voice propensity.
Possible effects of branch on propensity to voice
The level of voice propensity seems to be somewhat higher in grocery shops than what is the case
in a competitive environment.
Possible effects of branch on voice propensity are linked to age in the grocery shops. Age seems to
promote voice behaviour to a certain extent in the grocery shops. We tried to include the risk variable
in the model for Furniture shops, but it turned out as not significant.
The four models are all significant.
The voice propensity seems to be reduced by satisfaction in all the four models. Age seems to
promote voice to a certain extent when competition is low and in grocery shops.
WOM propensity
We noted that the propensity for WOM is the most common form of protest among the customer.
Again we start the study of variation in the WOM propensity with all the theoretical variables in
Table 7. In Table 12 we show the significant result. In addition we tried a model with those who had
complaint experience, but did not succeed in reaching a significant solution.
Table 12: Logistic Regression with the propensity to WOM as dependent variable
Propensity for WOM WOM
ß
Significant test :
Wald statistics
Satisfaction -.12* 9.7
Voice costs .17** 14.9
Discriminated treatment .79* 8.7
Initial-2 LOG Likelihood 390.5
Model -2 LOG Likelihood 352.1
Difference 38.4
Significance for model P<.001
Nagelkerke R
2
.17
Prediction ability 68%
N 288
** S<= 0.01
* S<=0.05
Factors influencing the WOM propensity are the satisfaction variable, voice costs, and
discriminating treatment. Increased satisfaction reduces the propensity for WOM. As voice costs
increase the propensity for WOM increase as well. This fit nicely in a rational model for customer’s
decisions. When the customers feel dissatisfied he/she normally evaluate either to voice or to WOM.
With high costs linked to voice the customer turn to negative WOM. Increased feeing of discriminating
treatment seems to increase the WOM propensity. The Logistic model is significant and it explains
17% of the variation in the dependent variable leaving room for other explaining factors.
A special analysis shows no significant difference in propensity for WOM between those who
have tried WOM and those who have not tried WOM with respect to treatment of complain.
In another special analysis of the furniture respondents we tried to trace effects of risk evaluation
to the actual shop. The risk evaluation was 1, 8 for those with low WOM propensity (0) and 2.6 for
those with high WOM propensity (1). N=192, F=6.2, Sig.=0.014. The difference in evaluation score is
significant. The higher the risk evaluation linked to a shop the higher the propensity for WOM.
We aim to investigate how these internal customers’ related variables are influenced by the
external variables as competition and shopping branch.
Int. Journal of Business Science and Applied Management / Business-and-Management.com
28
Table 13: Effects of branches and competition on the propensity to WOM
Degree of
competition
Branch
Low High Grocery Furniture
Voice propensity 0.50 0.39 0.45 0.39
ß ß ß ß
Satisfaction -.12* -.15*
Voice costs .13* .23** .19** .20**
Age -.03*
Discriminated treatment 1.0* 1,0* 0.8*
Initial – 2 LOG Likelihood 241.0 188.8 202.6 193.3
Model – 2 LOG Likelihood 224.8 166,2 178,7 177,4
Difference 16.2 22.6 23.9 15.9
Model significance P<0.001 P<0.001 P<0.001 P<0.001
Nagelkerke R
2
0.12 0.20 0.20 0.14
Percentage correct
predicted
60 74 64 68
N 174 151 147 147
** S<= 0.01
* S<=0.05
Effects of competition on WOM propensity
When competition increases the voice costs hold its important position, yet show an increase. The
discriminated treatment variable has a role in high competition, but is not present in low competition.
Experiencing discriminating treatment is probably a subject in WOM conversations!
The models are having poor explanatory power leaving most of the variance in WOM propensity
unexplained.
Effects of branch on WOM propensity
There is no clear difference between the two branches concerning factors for WOM propensity.
The voice costs are important in both branches and so is the equity treatment.
The level of WOM propensity seems to be high when competition is low and for customers in
grocery shops. The high level of WOM when competition is low could be explained by increasing costs
linked to an alternative form of protest, exit. WOM flourish more when alternative forms of protest are
more difficult.
Increased voice costs seem to increase the propensity for WOM. Feeling unequal treatment will
increase WOM propensity when there are low degree of competition and for customers linked to
furniture ships. The models are all significant, but their explanatory power is low.
5 CONCLUSION
Of the three forms of protest the propensity for WOM seems to be the most common with a
propensity factor of 0.4. The most exclusive form of protest seems to be exit with a factor score of 0.1.
The propensity for voice has a factor of 0.3.
Of the three forms of protest, we have the best model to explain the voice propensity. Nagelkerke
R
2
is 0.79. The model for exit propensity is second best having a Nagelkerke R
2
of 0.42. The model for
WOM propensity is not powerful, but we have identified some factors of importance to explain
variance in the WOM propensity. Nagelkerke R
2
is 0.14.
How do customers decide how to make a protest? A theoretical reflection
The effect of satisfaction on the propensity to perform a protest is strongest on the exit propensity
and weakest on the WOM propensity. The exit propensity seems to be the most serious form of protest.
An increase in the evaluation of risk linked to the shopping has positive influence on the propensity to
protest. An increase in the risk makes the deal more important for the customer.
The basis for any form of customer protest is low score on satisfaction. There are linkages
between the various forms of protest, exit and voice are positively correlated. The costs linked to voice
influence the propensity for WOM. The customers seem to do an evaluation between the three forms of
Bernt Krohn Solvang
29
protest. If the customers feel high voice costs, the WOM propensity increase. Voice is a more often a
chosen form of protest than exit, which seems to be more drastic and rare .
Customer protest seems to be a calculated behaviour governed by degree of loyalty, satisfaction
and of possible gains. If the costs linked to voice are high some customers prefer to go to friends and
acquaintances with negative WOM. The feeling of not being treated equally compared to other
customers is a strong motive for negative WOM. This fits into a calculated behaviour. The calculated
behaviour is seen as a sort of rational behaviour summing up feelings and factors linked to satisfaction
and calculating possible gains and losses, costs linked to exit and voice or WOM before the form of
protest is decided.
However, the calculated pattern is influenced and disturbed by a zone of tolerance created by
loyalty and by shopping frequency. The rational picture of the customers should also be moderated
since 1/3 of the customers (linked to grocery and furniture shops) do not use any form of protest. And
only a small number (2%) has experience in using all three forms of protest.
Treatment of customers complains is an important variable. A good treatment increase the
propensity for voice (instead of exit), while a good treatment reduce the propensity for WOM. A bad
treatment will increase the propensity for WOM, but reduce the propensity for voice.
Exit propensity
Exit propensity is influenced by satisfaction, loyalty, shopping frequency, risk evaluation and
treatment of complaints. Satisfaction, shopping frequency and treatment quality of complaining
behaviour will all reduce the propensity for exit if increased, and function as barriers for exit.
When the competition increases the exit costs are reduced. But the quality of the offer from the
shops could be increased by the competition. Shopping frequency seems to be more important in a
competitive environment and loyalty and satisfaction reducing the propensity for exit both when the
competition is high and when it is low. When the competition is low there is an effect of loyalty, in
high competition the effect is linked to satisfaction. Does low degree of competition promote positive
attitudes towards the shops?
Moreover, exit propensity seems to be higher in furniture shops than in grocery shops. The
customers’ dependence of the shops might be higher for the Grocery shops being more frequently
visited than a Furniture shop.
Voice propensity
Voice propensity is influenced by satisfaction and age, and a good complain treatment will
increase the propensity for voice (instead for exit). The negative influence of satisfaction on voice
propensity could have been weakened by a zone of tolerance since the effect variable is small, but
significant. An increase in risk evaluation does have the same effect. There are small effects if any of
external factors as competition and branch on the propensity to exit. Competition seems to make the
effects of satisfaction on voice propensity somewhat stronger
WOM propensity
Word of Mouth (WOM) is influenced by satisfaction/loyalty as the other two forms of protest. If
voice costs increases, the propensity for WOM also increases. A good treatment of complain behaviour
will reduce the propensity for WOM. Shopping frequency is also linked to WOM propensity, the
higher the risk evaluation, the higher the WOM propensity. The effects of competition seem to increase
the importance of voice costs and complain treatment. Complain treatment have a stronger effect in
Furniture shops than in Grocery shops.
For leaders
What measures should be made by leaders in shops in order to reduce formal and informal
protest? Firstly, they should make it more easy and comfortable for customers to make a complaint.
The more they can treat customer complaints in an orderly and nice way the less informal negative
word of mouth activity they will experience and they will reduce the exit propensity and lead the
customers to the complain organisation. Secondly, they should ensure that their customers feel they get
equal treatment.
Int. Journal of Business Science and Applied Management / Business-and-Management.com
30
REFERENCES
Aaker D. A. (1991) .Managing Brand Equity. New York: Free Press.
Andreassen T. W. and Bredal Dag. (1996). Kundepleie i praksis. Oslo: Ad Notam, Gyldendal.
Andreasen A.R. and A. Best 1977. Consumer complaint: Does business respond? Harvard Business
Review, July-August, 55-101
Andreassen T. W. (1997). Dissatisfaction with Services. Dissertation. Stockholm: Företaksøkonomiska
institusjonen, Stockholm University
Bearden W.O. and R.L. Oliver (1985). The Role of Public and Private Complaining in Satisfaction with
Problem Resolution. The Journal of Consumer Affairs. 19, 2, 222-240.
Bearden W.O. and J.E. Teel (1983). Selected Determinants of Customer Satisfaction and Complaint
Reports. Journal of Marketing Research, 20, 21-28.
Berry, L.L. and A. Parasuraman (1991). Marketing Service: Competing through Quality. New York:
The Free Press.
Bloemer, J.M., Kasper, H.D. (1995). The complex relationships between consumer satisfaction and
brand loyalty. Journal of Economic Psychology, 16,311-329.
Bloemer J.M., de Ruyter K. (1998). On the relationship between perceived service quality, service
loyalty and switching costs.- In: International journal of service industry management, 9p. 436-
454.
Brief P. A. (1998). Attitudes In and Around Organizations. Sage Publications Inc.
Cho Y, Im I, Hiltz R, Fjermestad J. (2002). The effects of post-purchase evaluation factors on online
vs. off line customer complaining behaviour: Implication for customer loyalty. Advances in
consumer research, Volume XXIX (29): 318-326.
Carman, J.M. 1970. Correlates of Brand Loyality: Some Positive Results. Journal of Marketing
Research nr.7: 67-76.
Churchill Jr., Gilbert A., Surprenant. (1982). An investigation into the determinants of customer
satisfaction. Journal of Marketing Research, Vol.19 Issue 4, pp 491- 504.
Dekimpe M.G., Steenkamp J.B. (1997). The Increasing Power of Store Brands: Building Loyalty and
Market Share. Long range planning.Vol.30.mr.6:917-930.
Griffin Jill. (1995). Customer Loyalty. How to Earn It. How to Keep It. New York: Lexington Books
Gilly M.C. and R.W. Hansen (1985). Consumer complaint handling as a strategic marketing tool. The
Journal of Consumer Marketing, 2, 5-16.
Grønhaug Kjell (1987). Exploring the Problem-Prone Consumers: Hypotheses and Empirical Findings.
European Journal of Marketing 21.1, 74-82.
Grønhaug Kjell. (1972). Risk indicators, Perceived risk and consumer’s choice of information sources.
The Swedish Journal of Economics. 7, 2, 246-262.
Grønhaug Kjell (1977). Kjøpers klageadferd: Noen undersøkelsesresultater (The Customer’s Complaint
Behaviour: Some Survey Results). Tidsskrift for Samfunnsforskning, 6, 6, 240-250.
Grønhaug Kjell and Mary C. Gilly. (1991). A transaction cost approach to consumer dissatisfaction and
complaint actions. Journal of Economic Psychology 12; 165-183.
Bernt Krohn Solvang
31
Grønhaug K. and G. Zaltman (1981). Complainers and no complainers revisited: Another look at the
data. Journal of Economic Psychology 1, 121—195.
Hair.J.F. R.E. Anderson, R.L. Tatham and W.C. Black. (1998). Multivariate Data Analysis. New
Jersey: Prentice Hall.
Hernandez S.A., W. Strahle, H.L. Garcia and R.C. Sorensen. (1991). A Cross-cultural Study of
Consumer Complaining Behavior: VCR owners in U.S. and Puerto Rico. Journal of Consumer
Policy 14, 35-62.
Hirschman, A. O. (1970). Exit, Voice and Loyalty. Harvard University Press, Cambridge Ma.
Homburg, C., Giering, A. (2001). Personal Characteristics as Moderators of the Relationship between
Customer Satisfaction and Loyalty. Psychology & Marketing. Vol. 18: 43-66.
Huefner, J.C., Hunt, K.H. (1994). Extending the Hirschman Model: When voice and exit don’t tell the
whole story. Journal of consumer satisfaction, dissatisfaction, and complaining Behavior. Vol 7.
P.267-270
Jacoby J., Chestnut, R. (1978). Brand Loyalty Measurement and Management Ronald Press
Publication, New York
Kotler Philip (1987). Marketing management New Jersey: Prentice-Hall.
Merton R.K. (1968). The Matthew Effect in Science. Science, 159 (3810): 56-63
Nayor G. and S.B. Kleiser. (2000). Negative versus Positive Word-of-Mouth. Journal of Consumer
Satisfaction, Dissatisfaction and Complaining Behavior. 13, 26-36.
Oliver R.L. (1997). Satisfaction. A Behavioural Perspective on the Consumer. New York: McGraw-
Hill Company.
Oliver R.L., Rust R.T. Varki S. (1997) Customer Delight: Foundations, Findings, and Managerial
Insight. Journal of Retailing. 73. NO.3:311-336.
Peter J.P. & Olson J.C. (1993) Consumer Behavior and Marketing Strategy Homewood Illinois 3rd ed.:
Irwin.
Ping R.A. (1997). Voice in Business-to-Business Relationships: Cost-of-Exit and Demographic
Antecedents. Journal of Retailing 73, 261-281.
Richins M.L. (1983). Negative Word-of-Mouth By Dissatisfied Consumers: A Pilot Study.
Journal of Marketing 47, 68-78.
Richins M.L. (1985). The role of product importance in complaining behavior. In H.K. Hunt and
R.L.Day (eds) Consumer Satisfaction, Dissatisfaction and Complaining Behavior. Bloomington,
Ind.: Department of Marketing, Indiana University. 50-53.
Selnes Fred and Reve Torgeir (1994) Relasjonsmarkedsføring-keiserens nye klær? Praktisk økonomi
og ledelse.2:61-70.
Singh J. 1990, B. Voice, Exit, and Negative Word-of-Mouth Behavior: An Investigation Across Three
Service Categories. Journal of the Academy of Marketing Sience.18,1, 1-15.
Singh J. (1990). A Typology of Consumer Dissatisfaction
Response Styles. Journal of
Retailing, 66, 57-99.
Singh J. (1991). Industry characteristics and consumer dissatisfaction. Journal of Consumers Affairs.
25, 19-57.
Int. Journal of Business Science and Applied Management / Business-and-Management.com
32
Snellman,K.,Vihtkari,T. (2003). Customer complaining behaviour in technology-based service
encounters. International journal of service industry Management,14 (2): 217-231)
Strauss B, Schmidt M,Schoeler A. (2005). Customer fustration in loyalty programs.
International Journal of Service Industry Management 16 (3-4): 229-252
Tarp (1986). Consumer complaint handling in America: An update study. Technical Assistance
research programs, Washington DC.
Teknologisk Institutt (1993). Kunden i fokus. Los-serien Consumer complaint handling in America:
Summary of findings and recommendations. Technical Assistance Research programs,
Washington. DC.
Troye S. V., (1994) Teori- og forskningsevaluering (Theory and Research Evaluation). Tano.
Voorhees, C.M., Brandy, M.K., Horowitz,D.M. (2006). A Voice From the Silent Masses: An
Exploratory and Comparative Analysis of Noncomplainers. Journal of the Academy of Marketing
Science.Volum 34,No. 4. pp 514- 527
Ward, J.C., Ostrom,A.L. (2006). Complaining to the Masses: The Role of Protest Framinig in
Customer- Created Complaint Web Sites. Journal of Consumer Research.Vol. 33 pp. 220-230.
Wind,Y. (1978) Issues and Advances in Segmentation Research. Journal of Marketing research 15,317-
337. John Wiley & Sons.
Zeithaml,V.A.,Parasuraman,A.,Berry,L.L. (1990). Delivering quality service. Balancing Customer
Perceptions and Expectations. New York: The Free Press.
