You are on page 1of 10


Journal of Retailing and Consumer Services 16 (2009) 271–280

Contents lists available at ScienceDirect

Journal of Retailing and Consumer Services

journal homepage:

Customer satisfaction and dissatisfaction in retail banking: Exploring the

asymmetric impact of attribute performances
Alessandro Arbore ,1, Bruno Busacca 2
Bocconi University, Via Bocconi, 8, 20136 Milan, Italy

a r t i c l e in fo abstract

The paper presents the results of an extensive study on the determinants of customer satisfaction for a
Keywords: retail bank. In doing so, it uses a revised version of the traditional analyses based on derived measures of
Retail banking attribute importance. The need for a revised methodology is prompted by the insights of the two- and
Customer satisfaction three-factor theories of customer satisfaction, such as Kano’s framework. Indeed, the evidence from
Three-factor theory 5000 customers of a prominent Italian bank confirms a non-linear and asymmetric relationship
between attribute performances and overall customer satisfaction. The results from both a traditional
and our revised approach are compared. While this approach can be applied across different industries,
it should not be assumed that the numerical results presented in the paper apply to contexts with
substantially different underlying characteristics. General trends and implications for banking services
are reported in the conclusions.
& 2009 Elsevier Ltd. All rights reserved.

1. Introduction 2. Background and literature review

The relevance of customer satisfaction in financial services has 2.1. Customer satisfaction in the banking industry
been widely investigated and assessed by the literature. Following
an increase in positive evidence, the focus on customer satisfaction A growing number of retail banks are directing their strategies
has become a primary goal, especially in the retail banking towards customer satisfaction. In fact, researchers have demon-
industry. A full understanding of its antecedents has therefore strated that customer satisfaction serves as a link to critical
become a critical issue for both researchers and practitioners. consumer behaviors, such as cross-buying of financial services,
This paper presents the results of an extensive study on the positive word-of-mouth, willingness to pay a premium-price, and
determinants of customer satisfaction in a prominent Italian bank. tendency to see one’s bank as a ‘‘relationship’’ bank (Winstanley,
It uses a revised version of the traditional analyses, based on 1997; Ehigie, 2006; Ndubisi, 2006).
derived measures of attribute importance (e.g. Dolinsky, 1991). The Ultimately, these behaviors proved to have a positive impact on
need for a revised methodology is prompted by the intuitions of key corporate outcomes, such as retention rates, average deposit
the two- and three-factor theories, which demonstrated a non- amounts, cost to the bank of providing services, and future
linear and asymmetric relationship between attribute perfor- earnings (Winstanley, 1997; Nagar and Rajan, 2005; Manrai and
mances and overall customer satisfaction (Anderson and Mittal, Manrai, 2007). These results are consistent with a broader
2000; Brandt, 1988; Gale, 1994; Johnston, 1995; Kano, 1984; literature assessing the positive relationship between customer
Matzler and Sauerwein, 2002; Oliver, 1997). satisfaction and firm profitability (cf. Anderson et al., 1994;
Fornell et al., 1996; Johnson et al., 1996; Ittner and Larcker, 1998;
Eklof et al., 1999; Zeithaml, 2000).
Following the evidence from these studies, the value of
identifying the most influential factors to trigger satisfaction and
dissatisfaction in retail banking becomes especially clear. This
 Corresponding author. Tel.: +39 02 5836 6529; fax: +39 02 5836 3790. information allows the firm to steer and optimize its marketing
E-mail addresses: (A. Arbore), efforts given its limited resources, keeping in mind that market- (B. Busacca). based assets, in turn, will increase shareholder value (Day and
Alessandro Arbore is Assistant Professor of Management at Bocconi Fahey, 1988; Srivastava et al., 1998).
University and Professor of Marketing at SDA Bocconi School of Management. He
Over time, researchers have identified many antecedents of
is senior researcher at the Customer and Service Science Lab, Bocconi University.
Bruno Busacca is Full Professor of Marketing at Bocconi University. He is customer satisfaction in different service industries (e.g. Oliver,
Director of the Masters Division at SDA Bocconi School of Management. 1993; Levesque and McDougall, 1996; Varki and Colgate, 2001).

0969-6989/$ - see front matter & 2009 Elsevier Ltd. All rights reserved.

272 A. Arbore, B. Busacca / Journal of Retailing and Consumer Services 16 (2009) 271–280

Table 1
Drivers of customer satisfaction in retail banking.

Dimension Attributes Authors Method

Functional Reliability, speed, accuracy, security, functionality Levesque and McDougall (1996), Winstanley (1997), Factor analysis, ordinary least squares
quality Jamal and Naser (2003), Jones (2004), and Ndubisi (OLS) regression, stepwise regression,
(2006) OLS regression
Relational Responsiveness, assurance, friendliness, courtesy, Levesque and McDougall (1996), Johnston (1997), Structural equation model (SEM),
quality commitment, communication Winstanley (1997), Lassar et al. (2000), Oppewal and critical incident technique (CIT), OLS
Vriens (2000), Jamal and Naser (2002), Jamal and regression, conjoint analysis, stepwise
Naser (2003), Jones (2004), and Ndubisi (2006) regression
Convenience Opening hours, travel distance, queuing time, parking Manrai and Manrai (2007), Levesque and McDougall Factor analysis
places, ATM availability (1996), Oppewal and Vriens (2000), Jones (2004),
Manrai and Manrai (2007)
Economics Charges/interest rates, price-quality ratio, price Levesque and McDougall (1996), Varki and Colgate Structural equation model (SEM),
fairness (2001), Nagar and Rajan (2005), Manrai and Manrai multivariate linear regression,
(2007), Matzler et al. (2006) multiple linear regression analysis
Tangibles Phisycal layout and furniture, phisycal facilities, dècor Positive impact on customer satisfaction: Levesque OLS regression
of branch environment, atmosphere of branch and McDougall (1996), Jamal and Naser (2002), Jamal
environment, cleanliness of branch environment, size and Naser (2003), no impact on customer satisfaction:
and furnishing of customer space, dress of the Wakefield and Blodgett (1999), Oppewal and Vriens
personnel (2000), and Jones (2004)
Problem handling and recovery Capability to avoid potential conflicts, efficient and Levesque and McDougall (1996),
effective handling of complaint, efficiency of solving Winstanley (1997), Ndubisi (2006),
problems, speed of finding and restoring errors and Manrai and Manrai (2007)

Several studies, for example, have emphasized the link between 2.2. The three-factor theory of customer satisfaction
satisfaction and the ‘‘service quality’’ construct (Taylor and
Baker, 1994; Levesque and McDougall, 1996; Johnston, 1997; Most of the studies reported on the pages above use indirect
Jamal and Naser, 2002; Ndubisi, 2006; Lassar et al., 2000; measures to assess the importance of attributes in retail banking.
Oppewal and Vriens, 2000). The traditional operationalization Typical indirect measures are the coefficients of a multiple
of service quality (Parasuraman et al., 1985, 1988) includes five regression where the dependent variable is an overall performance
dimensions: reliability, tangibles, responsiveness, assurance, rating (e.g. overall customer satisfaction), and the independent
and empathy. Building on the work of Parasuraman et al., other variables are specific performance ratings for each attribute
research in the banking literature distinguishes between two (Dolinsky, 1991; Wittink and Bayer, 1994; Anton, 1996; Taylor,
main dimensions affecting customer satisfaction: the quality 1997; Lowenstein, 1995).
of the core service provided by the bank and the quality of However, the implicit importance estimated this way would be
the relationship with the bank. Relevant drivers within the an average importance and would overlook any asymmetric
first dimension include ‘‘reliability’’, ‘‘security’’, ‘‘functionality’’, relationship between attribute performance and overall customer
‘‘accuracy’’, and ‘‘speed’’ (cf. Levesque and McDougall, 1996; satisfaction (Mittal et al., 1998; Matzler et al., 2004; Deng et al.,
Winstanley, 1997; Jamal and Naser, 2002, 2003; Jones, 2008). This is a relevant shortcoming of the previous literature,
2004; Ndubisi, 2006). Relationship drivers seem to be even more which conflicts with both evidence and theory.
important and include ‘‘responsiveness’’, ‘‘competences’’, ‘‘assur- Studies on customer satisfaction demonstrate that product
ance’’, ‘‘trust’’, ‘‘friendliness’’, ‘‘courtesy’’, ‘‘availability’’, ‘‘commit- attributes may affect overall satisfaction in a non-linear way.
ment’’, ‘‘flexibility’’, and ‘‘communication’’ (cf. Levesque and Preliminary studies from the 1970s and 1980s were successful in
McDougall, 1996; Johnston, 1997; Winstanley, 1997; Jamal testing a two-factor theory of customer satisfaction (Swan and
and Naser, 2002, 2003; Lassar et al., 2000; Jones, 2004; Combs, 1976; Maddox, 1981; Cadotte and Turgeon, 1988). They
Oppewal and Vriens, 2000; Ndubisi, 2006; Manrai and Manrai, demonstrated that consumers judge products on a limited set of
2007). attributes, some of which are relatively important in determining
Another recurrent key factor is the perception of costs, satisfaction, while others are not critical to consumer satisfaction
specifically ‘‘price fairness’’ and ‘‘price-quality ratio’’ (Levesque but unsatisfactory performance in these attributes may lead to
and McDougall, 1996; Winstanley, 1997; Varki and Colgate, dissatisfaction.
2001; Matzler et al., 2006). Further satisfaction drivers are Kano (1984) was the first to formalize these intuitions, adding
peripherals like the ‘‘location’’, both in terms of convenience and original contributions to the theory. Kano’s model was later
accessibility (Oppewal and Vriens, 2000; Manrai and Manrai, refined by other studies (Anderson and Mittal, 2000; Brandt, 1988;
2007). More controversial is the significant impact of other Gale, 1994; Johnston, 1995; Matzler and Sauerwein, 2002; Oliver,
‘‘tangibles’’, such as the ‘‘physical layout’’ of the bank (Johnston, 1997). These authors distinguish three categories of attributes:
1997; Oppewal and Vriens, 2000; Jones, 2004; Manrai and Manrai, basic factors, excitement factors, and performance factors (Fig. 1).
2007). Basic factors (or Kano’s dissatisfiers) are minimum requirements
Finally, researchers emphasize the positive role played by that do not have a positive impact on satisfaction if expectations are
problem-handling skills, like ‘‘capability to avoid potential con- exceeded, but generate dissatisfaction if they are not fulfilled.
flicts’’, ‘‘efficient and effective handling of complaint’’, ‘‘efficiency Excitement factors (or Kano’s satisfiers) are delighting attributes
of solving problems’’, and ‘‘speed of finding and restoring errors’’ that have no impact on dissatisfaction, but can foster satisfaction if
(Levesque and McDougall, 1996; Ndubisi, 2006; Manrai and delivered. Finally, performance factors (or Kano’s one-dimensional
Manrai, 2007). A summary of this literature review is offered in factors) have a symmetric impact on both satisfaction and
Table 1. dissatisfaction, in proportion to their level of fulfillment.

A. Arbore, B. Busacca / Journal of Retailing and Consumer Services 16 (2009) 271–280 273

Exciting attributes Overall satisfaction and attribute performance were measured

using a Likert-like scale of 1 (extremely low) to 10 (extremely high).
A complete set of attributes was first identified, based on previous
Performance literature and suggestions from company management. Then, a
subset of relevant satisfaction drivers was identified through both
Expectations Expectations qualitative and quantitative exploratory analysis (e.g. Cooley and
not fulfilled exceeded Lohnes, 1971; Darlington, 1990). Having a very large dataset, a
cross-validation approach was used and a stepwise regression
procedure was run (e.g. Darlington, 1990; Lindeman et al., 1980;
Neter et al., 1985; Stevens, 1986). Finally, using variance inflation
Basic attributes factors (VIF), we verified that multicollinearity among the
Dissatisfaction variables was not severe.
The idea behind the technique used in this paper is to assess
Source: Adapted from Matzler et al., 1996
the importance of these attributes by distinguishing – ex post – the
Fig. 1. The three-factor theory of customer satisfaction. Source: Adapted from following cases:
Matzler et al. (1996).

 the impact of negative performance of an attribute (i.e. negative

After Kano, different methodologies have been proposed to ratings) on unsatisfied bank customers;
identify the three factors. The most common methods are Kano’s  the impact of excellent performance of an attribute (i.e. top
questionnaire, the critical incident technique, the analysis of rating scores) on delighted customers.
complaints and compliments, the importance grid, and the
penalty-reward-contrast analysis. Appendix 1 sums up these To estimate these effects, two dummy variables and two
contributions and their possible limitations. For detailed discus- interaction terms for each relevant attribute are introduced into
sion, refer to the article by Busacca and Padula (2005). The a classic multivariate regression model (see the formula below).
methodology refined for this study is a development of the Overall satisfaction is the dependent variable.
penalty-reward-contrast analysis (Brandt, 1988). The first dummy is set to ‘‘1’’ when the following conditions are
simultaneously met: the customer is unsatisfied with the company
(e.g. overall customer satisfaction below 5 in a Likert-like scale
3. Hypotheses
from 1 to 10) and s/he gave negative feedback on the attribute in
question. Symmetrically, the second dummy associated with each
Zielke (2008) recently recalled that only a few authors have
attribute is set to ‘‘1’’ when: the customer is delighted with the
developed hypotheses on the asymmetric effects of performance
company (e.g. overall customer satisfaction above 8 on a scale of
dimensions on satisfaction. Even more surprisingly, almost every
1–10) and s/he provided positive feedback on the attribute in
study on customer satisfaction for financial services (see the
previous section) implicitly assumes no asymmetric effects at all.
Finally, as we said, two further terms for each attribute must be
We claim that this is a major shortcoming in the retail banking
added. They are two interaction terms multiplying the actual
literature, conflicting with the theoretical background introduced
performance of the attribute by the dummy variables identified
above (Kano, 1984).
above. By doing this, the two terms will provide two independent
Hypothesis 1. Significant non-linear effects of performance adjustments to the traditional coefficient of the attribute. Specifi-
dimensions on overall satisfaction exist in retail banking. cally, their parameters will provide the following:

Hypothesis 2. When the traditional coefficients of a multiple  a corrective adjustment to estimate the extent to which top
regression are used as indirect measures of attribute importance, performance in this attribute contributed to high levels of
the estimated results may disregard significant non-linear effects. customer satisfaction;
 a corrective adjustment to estimate the extent to which poor
To estimate asymmetric effects using indirect measures of
performance in this attribute contributed to low levels of
attribute importance, we propose a new approach, presented
customer satisfaction.
below. More specifically, regarding the list of relevant attributes in
retail banking, this study will follow an exploratory approach (see
Sections 4 and 7). We acknowledge the limits of this step, while In formula,
recalling that our main focus is on the evaluation of asymmetric
Y^ i ¼ b0 þ b1j X ji þ b2j D1ji þ b3j X ji D1ji þ b4j D2ji þ b5j X ji D2ji
where Yi is the overall satisfaction (from 1 to 10) expressed by the
ith customer and Xji the performance evaluation (from 1 to 10) of
4. Data collection and methodology the jth attribute, expressed by the ith customer. D1ji ¼ 1 if the ith
customer expressed a negative evaluation of the jth attribute
For the purposes of this study, customer satisfaction of a (Xjio5) and overall dissatisfaction with the bank (Yio5); 0
prominent Italian bank was analyzed. The research is based on a otherwise. D2ji ¼ 1 if the ith customer expressed a positive
survey of a sample of 5209 customers. Data were collected by a evaluation of the jth attribute (Xji47) and overall satisfaction
certified research company using computer-assisted telephone with the bank (Yi47); 0 otherwise.
interview methodology. The sampling frame was the set of the It is important to note that the specification of the dummy
Bank’s customers with either a checking or a savings account at variables may lead to a spurious correlation problem. For this
the time of the survey (March–April 2006). Stratified random reason, Table 2 details their Pearson correlation coefficients with
sampling was used, based on geographic area, number of months the dependent variable, confirming that no strong relationship
within the customer base, average number of transactions per exists due to our mathematical process (Studenmund, 2001). In
month, and average monthly balance. fact, all the coefficients range from 0.57 to 0.26 (in absolute value),

274 A. Arbore, B. Busacca / Journal of Retailing and Consumer Services 16 (2009) 271–280

that is, a moderate to weak relationship. Nonetheless, caution is The estimated model provides very good measures both in
warranted when interpreting the t-scores and overall fit of the terms of statistical significance and overall fit, explaining almost
regression, since they could be overstated in the presence of strong 60% of the variance in customer satisfaction. Nevertheless, as
spurious relationships. already said, the assumptions of linearity are inconsistent with
The empirical results are presented in Section 5. While we are both theory and empirical evidence. The relationship between
confident that this general approach is broadly applicable across attribute-level performance and overall satisfaction is more
different industries, it should not be assumed that our specific complex and requires specific adjustments.
findings apply to contexts with substantially different underlying The proposed adjustments are computed through the metho-
characteristics from the context considered in this paper. dology described in the previous section. The new results are
reported in Table 5. This also provides company management with
a visual representation of the results, making them easier to
interpret. Notably, the estimated impacts of low performance on
5. Empirical findings
dissatisfaction are significantly different from the impacts of high
performance on high satisfaction (adjusted beta in Table 5), thus
The case of an Italian retail bank is used to show how the confirming Hypothesis 1. In addition, the non-significant levels of
analysis of customer satisfaction must take into consideration the the reference cases (i.e. the cases with both dummies set to zero,
asymmetric impact that attribute performances may have on it. as well as those receiving scores of 5, 6, or 7) seem to corroborate
The results from both a traditional and our revised approach are the disconfirmation of expectations paradigm for customer satisfac-
compared. tion (Oliver, 1980, 1997), since neutral ratings would have no
The typical procedure used in the banking literature on statistical impact on overall satisfaction.
customer satisfaction estimates the importance of the attributes All this information can be used for a more accurate
through either standardized or unstandardized regression coeffi- consideration on what priorities to use when allocating scarce
cients. We started out with the same procedure. The dependent resources to optimize customer satisfaction. In fact, to better
variable, again, is ‘‘overall customer satisfaction with the bank’’. understand the role of each specific attribute in our value
The unstandardized results are shown in Table 3, while Table 4 proposition, it is important to
lists our set of critical value drivers in decreasing order of
importance, at least according to this approach. Appendix 2  look at the impact of both low and high performance on both
provides a definition of the variables, as they were described to the dissatisfaction and satisfaction,
respondents. Nonetheless, we would like to understate the  check current performance as reported by customers, both on
relevance of our list of attributes (we only claim internal validity), average, and with dissatisfied customers, and
while keeping attention on the issue of asymmetric effects.  compare the information in the two points above across all the
relevant attributes.
Table 2
The moderate correlation between our dummy variables (D1j and D2j) and the
Accordingly, two different charts can be drawn up (examples
dependent variable (Y).
are given in Figs. 2 and 3): one to help foster customer satisfaction
Pearson correlations between the dummy variables and overall customer and one to minimize dissatisfaction. These are a variation of

Table 4
D1_satisfaction with the family banker 0.568a
Derived measures of importance according to a traditional regression analysis.
D1_solid company 0.556a
D1_innovative company 0.554a
Value drivers Estimated impact on Current
D1_easiness to contact the company 0.577a
customer satisfaction performance
D1_technical quality of the contents received 0.495a
D1_provision of suitable investments 0.531a
(1) Financial results from subscribed 0.30 7.44
D1_financial results from subscribed products 0.515a
D2_satisfaction with the family banker 0.405a
(2) Overall satisfaction with the family 0.19 8.54
D2_innovative company 0.477a
D2_solid company 0.480a
(3) Provision of suitable investments 0.15 7.73
D2_easiness to contact the company 0.354a
(4) Solid company 0.14 8.02
D2_technical quality of the contents received 0.263a
(5) Technical quality of the contents 0.10 8.02
D2_provision of suitable investments 0.513a
D2_financial results from subscribed products 0.556a
(6) Innovative company 0.09 8.01
(7) Easiness to contact the company 0.06 8.40
Correlation is significant at the 0.01 level (2-tailed).

Table 3
Traditional analysis to elicit attributes’ importance.

Parameter Estimated beta Standard error t Value Pr4jtj

Intercept 0.1774 0.1162 1.53 0.1271

Easiness to contact the company 0.0592 0.0122 4.85 o.0001
Provision of suitable investments 0.1492 0.0137 10.91 o.0001
Technical quality of the contents received 0.0961 0.0136 7.07 o.0001
Financial results from subscribed products 0.2957 0.0107 27.59 o.0001
Innovative company 0.0888 0.0163 5.44 o.0001
Solid company 0.1391 0.0162 8.59 o.0001
Overall satisfaction with the family banker 0.1885 0.0112 16.81 o.0001

R2 ¼ 0.58.
Adjusted R2 ¼ 0.57.

A. Arbore, B. Busacca / Journal of Retailing and Consumer Services 16 (2009) 271–280 275

Table 5
Derived measures of attribute importance: the new insights.

Impact of low performances on dissatisfaction Reference cases Impact of high performances on high satisfaction

Adjusted Pr4jtj Intercept Pr4jtj beta Pr4t (one Adjusted Pr4 jtj Intercept Pr4jtj
beta correction tail) beta correction

Intercept 7.9817 o.0001

Easiness to contact the company 0.0382 0.6283 0.2082 0.5548 0.0611 n.s. 0.099** 0.0021 0.4686 0.048
Provision of suitable investments 0.0538 0.4688 0.7383 0.0332 0.00315 n.s. 0.166*** o.0001 1.1491 o.0001
Technical quality of the contents received 0.0946 0.2796 0.7075 0.0847 0.0586 n.s. 0.2111*** o.0001 1.4929 o.0001
Financial results from subscribed 0.2408*** o.0001 1.8663 o.0001 0.0596 0.0017 0.1383*** o.0001 0.798 0.0004
Innovative company 0.0351 0.7325 0.7537 0.124 0.0458 n.s. 0.1353** 0.0014 0.8948 0.0041
Solid company 0.4289*** o.0001 2.2971 o.0001 0.0672 n.s. 0.1772*** o.0001 1.0677 0.0007
Overall satisfaction w/ the family banker 0.3623*** o.0001 1.9451 o.0001 0.1044 n.s. 0.3259*** o.0001 1.8992 o.0001

R2 ¼ 0.6829.
Adj. R2 ¼ 0.6808.

0.35 simplification for company management. We propose, instead,

iority a heuristic interpretation of the two graphs, where the
0.30 gh Satisfaction w/family banker attributes in the upper left corner of our space represent our top
Impact on Satisfaction

priorities (either to increase customer satisfaction or to

0.25 Financial results from Technical quality of the contents sent reduce dissatisfaction, depending on the graph) and those in the
subscribed products
0.20 bottom-right corner should be the last investment areas to
Solid company consider. It is important to remember that the insights from
0.15 Availability of the graphs will be integrated and interpreted with the
suitable of investments
Innovative company sensitive sight of company management, which will consider,
Easiness to contact among other things, the marginal costs of the improvements and
the company
0.05 the different lifetime value of different customers (e.g. Baesens
w et al., 2004).
0.00 Above all, the new insights confirm the asymmetric impact of
7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 attribute performances on customer satisfaction. Further, the new
Current Performance results help to explain the inaccuracy of previous estimates,
obtained using the traditional method presented in the retail
Fig. 2. An example of impact-performance chart to improve customer satisfaction.
banking literature (Hypothesis 2).
To provide an example of the potential misleading conclusions
of the traditional approach, we start by considering the ‘‘technical
quality of the contents received from the bank’’. With a beta
ity coefficient equal to 0.096, the traditional analysis would classify
2.5 ior
r this value driver as one of the least important in our set (Table 4).

Solid company
Hi Indeed, by averaging its implicit importance, the exciting nature of
2 this feature would not emerge in traditional analysis. Table 5 and
Satisfaction w/family banker
Financial results from Fig. 4 now tell us a different story. Fig. 4 represents the
1.5 subscribed products relationship between this variable and overall customer satisfac-
tion. The solid line is the new estimate, while the dotted line is the
1 Technical quality of estimate based on the old linear parameter. The lines are drawn
Innovative company the contents sent assuming that the remaining variables perform at their average.
Provision of
ity This implies a baseline level of customer satisfaction equal to 8.
0.5 suitable of investments ior
Easiness to contact Pr The new approach reveals substantial contribution from the
the company Lo attribute of ‘‘technical quality of the contents received from this
0 bank’’ in explaining top levels of customer satisfaction. Table 5
4 4.5 5 5.5 6 6.5 7
reveals the exciting nature of this element, since the adjusted beta
Current Performance as perceived by coefficient estimating the impact of high performance on high
DISSATISFIED customers satisfaction is almost twice the adjusted beta coefficient capturing
the impact of low performance on dissatisfaction (0.21 vs. 0.09).
Fig. 3. An example of impact-performance chart to reduce customer dissatisfac-
tion. The opposite is true for basic factors, such as ‘‘being a solid
company’’ (0.18 vs. 0.43). Finally, for the so-called performance
attributes, the two coefficients are very similar, as in the case of
diagonal impact-performance models already used in other ‘‘overall satisfaction with the family banker’’ (0.33 and 0.36).
studies (e.g. Hawes and Rao, 1985; Slack, 1994), but now include The technical quality of the contents sent by this bank to its
the three-factor theory intuitions. For impact-performance appli- clients was actually the second most important factor in explain-
cations, diagonal line models were more reliable than quadrant ing high levels of customer satisfaction. This information can
methods (Bacon, 2003). therefore be usefully combined with the current performances of
Purposely, we do not provide any algebraic formula to rank our the attributes. In our case, for example, this value driver showed
priorities with mathematical precision. Those kinds of approaches interesting margins of improvement, suggesting new opportu-
have some theoretical shortcomings and represent a risky over- nities for company management.

276 A. Arbore, B. Busacca / Journal of Retailing and Consumer Services 16 (2009) 271–280

Impact* of high performances (8 to 10)

9 on satisfaction with the bank:
Impact* of low performances (1 to 5) +0,20 + 0,21 x
on dissatisfaction with the bank:
-0,71 + 0,09 x Traditional beta: 0,10

Current performance,
as perceived on average: 8,0
Overall Satisfaction with the bank

Current performance,
as perceived by
dissatisfied customers: 6,8

* Impact on average satisfaction

1 5 8 10

Customers’evaluation about the technical quality

of the contents received from the bank

Fig. 4. A non-linear relationship between attribute performance and overall satisfaction.

On the contrary, ‘‘easiness to contact the bank’’ was one doing so, however, we must keep in mind that the level of attribute
of the top performing dimensions of the organization we studied. satisfaction can cause a change in attribute importance. A specific
Still, by the same token its estimated contribution to customer methodology to deal with this issue has been suggested in this
satisfaction was not very high. Given both the marginal contribu- paper and its application has been illustrated through an extensive
tion to customer satisfaction and the current performance, such study on an Italian retail bank. The asymmetric and non-linear
organization may currently be overinvesting in this dimension. relationship between attribute performances and customer satis-
Yet, before considering a reallocation of resources, it is also faction was confirmed by our findings.
important to take a closer look at the impact on customer While this approach can be applied across different industries,
dissatisfaction. In our specific case, this information also con- it should not be assumed that the numerical results presented in
firmed the possibility of freeing up resources due to the out- the sections above apply to contexts with substantially different
standing performance underlying characteristics. Nonetheless, the study helps to explain
This conclusion would have been different in the case of a basic a number of interesting insights for the retail banking industry.
attribute, such as being considered a reliable, solid company. The Specifically,
traditional analysis classified this attribute as being of average
importance. Conversely, while being a solid company will not
delight any customer, it has the potential to have a tremendous  reputation is an increasingly critical, basic attribute and this is
impact on dissatisfaction (Table 5 and Fig. 5). For no reason, then, probably emphasized by the recent subprime mortgage crisis.
would it be appropriate to disinvest in ‘‘basic’’ attributes. In the As in the study reported in this paper, being a solid bank will
study, this attribute actually represented the major dissatisfaction not delight any customer, but a poor perception of this attribute
trigger, with a tremendous risk in terms of negative word-of- would immediately foster high dissatisfaction,
mouth. In these cases, it would be undoubtedly worthy a prompt  individual consumers are becoming more and more sophisti-
recovery of the most critical clients – critical both in terms of cated, confirming the social trend of customer evolution and
economic value and visibility. empowerment. In our case, two important antecedents of
satisfaction were the availability of a wide range of suitable
investments, and the technical quality of the informative
6. Conclusions and implications material sent to the clients,
 interpersonal relationships are confirmed to be a tremendous
Identifying the antecedents of customer satisfaction has driver of both customer satisfaction and dissatisfaction.
become a major concern for both researchers and marketing By providing social and confidence/trust benefits (Adelman
practitioners, especially in retailing and service industries. In et al., 1994; Goodwin and Gremler, 1996; Hennig-Thurau

A. Arbore, B. Busacca / Journal of Retailing and Consumer Services 16 (2009) 271–280 277

Impact* of high performances (8 to 10)

9 Impact* of low performances (1 to 5) on satisfaction:
on dissatisfaction: +0,35 + 0,18x
-2,30 + 0,43x
Traditional beta: 0,14

Current performance,
as perceived on average: 8, 2

Current performance,
as perceived by
Overall Satisfaction with the bank

dissatisfied customers : 5,6

* Impact on average satisfaction

1 5 8 10

Customers’ evaluation about the solidity of the bank

Fig. 5. A non-linear relationship between attribute performance and overall satisfaction.

et al., 2002), this relationship contributes to the creation selection, we do not claim external validity for this outcome,
of value and, in turn, to customer loyalty (Reynolds and assuming a conservative position in the presence of multiple
Beatty, 1999; Butcher et al., 2001). A pertinent suggestion statistical comparisons. From this point of view, we confirm the
for retail banking can be adapted from recent studies on exploratory nature of our research, whose main goal was to
salespersons (Palmatier et al., 2007a, 2007b). The more a illustrate a different approach for an improved analysis of
customer perceives that all the employees with whom s/he satisfaction and dissatisfaction drivers. Additional research might
deals behave consistently, the more likely s/he will infer test our framework both in different cultures and different
that this behavior is due to underlying causes related to the businesses. For future studies, we also suggest multi-item
bank, such as employee selection or training. So, in this case measures of performance: some attributes of a value proposition
too, a well-established reputation and a positive image would are broad in scope and simply cannot be assessed with a single
be both the result and a signal of interpersonal consistency, question. In addition, a composite score resulting from a multi-
leading to a much stronger and positive attitude toward the item measure tends to be more consistent over time and, in
bank. general, more reliable, because random errors can average out
(Spector, 1992).
From the point of view of the procedure, we showed that no
7. Limitations and directions for future research strong correlations are generated by our methodology. None-
theless, VIF analysis is always warranted, as well as a correlation
The overall approach that we propose has been illustrated test between our dummy variables (D1j and D2j) and the
using a large customer base from the Italian banking context. dependent variable. In the presence of strong correlations, caution
Nonetheless, while we are confident that our procedure is widely is needed in interpreting overstated t-scores and overall fit of the
applicable across other retail and service businesses, it should not regression.
be assumed that our specific findings apply to contexts with A final remark for any future implementation of this method
different underlying characteristics from the one considered here. relates to the heterogeneity of the sample (Allenby et al., 1998).
Caution should also be used in considering the list of attributes This kind of analysis shall always be applied to a relatively
that we selected for this study, since a quantitative exploratory homogenous subset of clients. Accordingly, either a preliminary
procedure was adopted. Although stepwise regression with cross- cluster analysis or a latent class analysis is strongly suggested (e.g.
validation is the least controversial methodology for variable Kamakura and Russell, 1989; Mizuno et al., 2008).

278 A. Arbore, B. Busacca / Journal of Retailing and Consumer Services 16 (2009) 271–280

Appendix 1

The different methods to identify the three factors of customer satisfaction are listed below.

Method Authors Results Comments

Kano’s Kano (1984), Berger et al. (1993), Three-factor  Especially used in the QFD literature
questionnaire Matzler et al. (1996), and Yang theory supported  Reliability and validity not thoroughly tested.
(2005)  Weak in practice, since the allocation of attributes to categories is based on a
frequency distribution of the responses
 Time-consuming and costly method

Critical incident Swan and Combs (1976), Two-factor theory  Widely used in service quality research
technique Maddox (1981), Silvestro and supported  External validity questionable, since based on anecdotes
Johnston (1990), Stauss and Three-factor  Complex method: difficulties in processing and analyzing anecdotal materials
Hentschel (1992), Johnston theory supported
(1995), and Backhaus and Bauer
Analysis of Cadotte and Turgeon (1988) Three-factor  Validity and reliability questionable (see CIT)
complaints theory supported  Bias entailed in the contents
Importance grid Vavra (1997); Matzler and Three-factor  User-friendly approach
Sauerwein (2002), and Busacca theory supported  Validity questionable
and Padula (2005)  No theory behind the method
 The use of explicitly derived importance is a source of misinterpretation

The penalty- Brandt (1988), Mittal et al. Three-factor  Uses data from ‘‘typical’’ satisfaction surveys, therefore, no additional data
reward- (1998), Anderson and Mittal theory supported collection is required
contrast (2000), Matzler and Sauerwein  User-friendly approach.
analysis (2002), and Matzler et al. (2004,  Typically, small amount of overall satisfaction explained
The shapely value Conklin et al. (2004) Two-factor theory  Computational difficulties (sampling of attribute combinations is necessary)
supported  Identifies only the most relevant combination of dissatisfies (or key-enhancers)

Appendix 2

List of attributes: the original questions in the survey are listed below:

Attribute/variable label Question in the survey

Financial results from ‘‘On a scale of 1–9 (with 1 meaning completely dissatisfied and 9 meaning very highly satisfied), how satisfied are you with the overall
subscribed products results of the financial products that you subscribed with our bank? You can consider both the current value of your investment and
the monetary pay back that you have received till now.’’
Overall satisfaction with ‘‘Please think of the overall experience with your family banker. On a scale of 1–9 (with 1 meaning completely dissatisfied and 9
the family banker meaning very highly satisfied), overall, how satisfied are you with your family banker?’’
Provision of suitable ‘‘Please think of the range of investments and financial products that our bank currently offers to its customers. How well do you think
investments our offer suits your financial needs? (on a scale of 1–9 (with 1 meaning completely dissatisfied and 9 meaning very highly satisfied).’’
Solid company ‘‘Overall, solvency, liquidity and the quality of assets make a solid bank. In other words, we think of a solid bank as a bank where your
money will always be safe.
On a scale of 1–9, with 1 meaning that you totally disagree and 9 meaning that you fully agree, would you refer to our bank as a solid
Technical quality of the ‘‘We send you a monthly newsletter together with your financial statement. On a scale of 1–9, how much do you appreciate the
contents sent technical quality of the contents and the analysis that we send you? 1 means that you do not appreciate them at all and 9 means that
you appreciate them highly.’’
Innovative company ‘‘We think of an innovative bank as a company that keeps offering new services to its customers as well as old services in new,
innovative ways.
On a scale of 1–9, with 1 meaning that you totally disagree and 9 meaning that you fully agree, would you refer to our bank as an
innovative company?’’
Easiness to contact the ‘‘Think of all the possible ways that you can contact our bank when you need to. Based on this information and on your own
company preferences, how easy do you think it is to contact us? Please express your evaluation on a scale of 1–9, with 1 meaning very hard to
contact and 9 meaning extremely easy to contact.’’

Anderson, E.W., Mittal, V., 2000. Strengthening the satisfaction-profit chain. Journal
of Service Research 3 (2), 107–120.
Adelman, M.B., Ahuvia, A., Goodwin, C., 1994. Beyond smiling – social support and
Anton, J., 1996. Customer Relationship Management: Making Hard Decisions with
service quality. In: Rust, R., Oliver, R. (Eds.), Service Quality – New Directions in Soft Numbers. Prentice-Hall, Upper Saddle River, NJ.
Theory and Practice. Sage, Thousands Oaks, CA, pp. 139–171. Bacon, D.R., 2003. A comparison of approaches to importance–performance
Allenby, G.M., Arora, N., Ginter, J.L., 1998. On the heterogeneity of demand. Journal analysis. International Journal of Market Research 45 (1), 55–71.
of Marketing Research 35 (3), 384–389. Backhaus, K., Bauerm, M., 2000. The impact of critical incidents on customer
Anderson, E.W., Fornell, C., Lehmann, D.R., 1994. Customer satisfaction, market share satisfaction in business-to-business relationship. Journal of Business to
and profitability: findings from Sweden. Journal of Marketing 58 (3), 53–66. Business Marketing 8 (1), 25–44.

A. Arbore, B. Busacca / Journal of Retailing and Consumer Services 16 (2009) 271–280 279

Baesens, B., Verstraeten, G., Van den Poel, D., Egmont-Petersen, M., Van Kenhove, P., Matzler, K., Bailom, F., Hinterhuber, H.H., Renzl, B., Pichler, J., 2004. The asymmetric
Vanthienen, J., 2004. Bayesian network classifiers for identifying the slope of relationship between attribute-level performance and overall customer
the customer lifecycle of long-life customers. European Journal of Operational satisfaction: a reconsideration of the importance–performance analysis.
Research 156 (2), 508–523. Industrial Marketing Management 33 (4), 271–277.
Berger, C., Blauth, R., Boger, D. et al., 1993. Kano’s methods for understanding Matzler, K., Hinterhuber, H., Sauerwein, B., 1996. How to delight your customer.
customer defined quality. Journal of the Japanese Society for Quality Control Journal of Product and Brand Management 5 (2), 6–18.
(Fall), 3–35. Matzler, K., Sauerwein, E., 2002. The factor structure of customer satisfaction:
Brandt, R.D., 1988. How service marketers can identify value-enhancing service an empirical test of the importance grid and the penalty-reward-
elements. The Journal of Services Marketing 2 (3), 35–41. contrast analysis. International Journal of Service Industry Management 13
Busacca, B., Padula, G., 2005. Understanding the relationship between attribute (4), 314–332.
performance and overall satisfaction: theory, measurement and implications. Matzler, K., Würtele, A., Renzl, B., 2006. Dimensions of price satisfaction: a study in
Marketing Intelligence and Planning 23 (6), 543–561. the retail banking industry. International Journal of Bank Marketing 24 (4),
Butcher, K., Sparks, B., O’Callaghan, F., 2001. Evaluative and relational influences on 216–231.
service loyalty. International Journal of Service Industry Management 12 (4), Mittal, V., Ross, W.T., Baldasare, P.M., 1998. The asymmetric impact of negative and
310–327. positive attribute-level performance on overall satisfaction and repurchase
Cadotte, E.R., Turgeon, N., 1988. Key factors in guest satisfaction. The Cornell H.R.A. intentions. Journal of Marketing 62 (1), 33–47.
Quarterly 28 (4), 45–51. Mizuno, M., Saji, A., Sumita, U., Suzuki, H., 2008. Optimal threshold analysis of
Conklin, M., Powaga, K., Lipovetsky, S., 2004. Customer satisfaction analysis: identifi- segmentation methods for identifying target customers. European Journal of
cation of key drivers. European Journal of Operational Research 154 (3), Operational Research 186 (1), 358–379.
819–827. Nagar, V., Rajan, M.V., 2005. Measuring customer relationships: the case of the
Cooley, W.W., Lohnes, P.R., 1971. Multivariate Data Analysis. Wiley, New York, NY. retail banking industry. Management Science 51 (June), 904–919.
Darlington, R.B., 1990. Regression and Linear Models. McGraw-Hill, New York, NY. Ndubisi, N.O., 2006. A structural equation modelling of the antecedents of
Day, G.S., Fahey, L., 1988. Valuing marketing strategies. Journal of Marketing 52 (3), relationship quality in the Malaysia banking sector. Journal of Financial
45–57. Services Marketing 11 (2), 131–141.
Deng, W., Kuo, Y., Chen, W., 2008. Revised importance–performance analysis: three Neter, J., Wasserman, W., Kutner, M.H., 1985. Applied Linear Statistical Models:
factor theory and benchmarking. The Service Industries Journal 28 (1), 37–51. Regression, Analysis of Variance, and Experimental Designs. Irwin, Homewood,
Dolinsky, A.L., 1991. Considering the competition in strategy development: an IL.
extension of importance–performance analysis. Journal of Health Care Market- Oliver, R.L., 1980. A cognitive model of the antecedents and consequences
ing 11 (1), 31–36. of satisfaction decisions. Journal of Marketing Research 17 (November),
Ehigie, B.O., 2006. Correlates of customer loyalty to their bank: a case study in 495–507.
Nigeria. International Journal of Bank Marketing 24 (7), 494–508. Oliver, R.L., 1993. Cognitive, affective and attribute base of satisfaction response.
Eklof, J.A., Hackl, P., Westlund, A., 1999. On measuring interactions between Journal of Consumer Research 20 (December), 418–430.
customer satisfaction and financial results. Total Quality Management 10 (4 Oliver, R.L., 1997. Customer Satisfaction. A Behavioral Perspective on the Consumer.
and 5), 514–522. McGraw-Hill, New York, NY.
Fornell, C., Johnson, M.D., Anderson, E.W., Jaesung, C., Bryant, B.E., 1996. The Oppewal, H., Vriens, M., 2000. Measuring perceived service quality using inte-
American customer satisfaction index: nature, purpose, and findings. Journal of grated conjoint experiments. International Journal of Bank Marketing 18 (4),
Marketing 60 (4), 7–18. 154–169.
Gale, B.T., 1994. Managing Customer Value. The Free Press, New York, NY. Palmatier, R.W., Scheer, L.K., Steenkamp, J.-B.E., 2007a. Customer loyalty to whom?
Goodwin, C., Gremler, D.D., 1996. Friendship over the counter: how social aspects of Managing the benefits and risks of salesperson-owned loyalty. Journal of
service encounters influence consumer service loyalty. In: Brown, S.B., Bowen, Marketing Research 44 (2), 185–199.
D., Swartz, T. (Eds.), Advances in Services Marketing and Management, vol. V. Palmatier, R.W., Scheer, L.K., Houston, M.B., Evans, K.R., Gopalakrishna, S., 2007b.
JAI Press, London, pp. 247–282. Use of relationship marketing programs in building customer–salesperson and
Hawes, J.M., Rao, C.P., 1985. Using importance–performance analysis to customer–firm relationships: differential influence on financial outcomes.
develop health care marketing strategies. Journal of Health Care Marketing 5 International Journal of Research in Marketing 24 (3), 210–223.
(4), 19–25. Parasuraman, A., Zeithaml, V.A., Berry, L.L., 1985. A conceptual model of service
Hennig-Thurau, T., Gwinner, K.P., Gremler, D.D., 2002. Understanding relationship quality and implications for future research. Journal of Marketing 49 (Autumn),
marketing outcomes: an integration of relational benefits and relationship 41–50.
quality. Journal of Service Research 4 (3), 230–247. Parasuraman, A., Zeithaml, V.A., Berry, L.L., 1988. SERVQUAL: a multi-item scale for
Ittner, C.D., Larcker, D.F., 1998. Are nonfinancial measures leading indicators of measuring consumer perceptions of service quality. Journal of Retailing 64 (1),
financial performance? An analysis of customer satisfaction. Journal of 12–40.
Accounting Research 36 (Suppl.), 1–35. Reynolds, K.E., Beatty, S.E., 1999. Customer benefits and company consequences of
Jamal, A., Naser, K., 2002. Customer satisfaction and retail banking: an assessment customer–salesperson relationships in retailing. Journal of Retailing 75 (1),
of some of the key antecedents of customer satisfaction in retail banking. 11–32.
International Journal of Bank Marketing 20 (4), 146–160. Silvestro, R., Johnston, R., 1990. The determinants of Service Quality: Hygiene and
Jamal, A., Naser, K., 2003. Factors influencing customer satisfaction in the retail Enhancing factors. Quality Services II, Selected Papers, Warwick Business
banking sector in Pakistan. International Journal of Commerce and Manage- School, England, pp. 193–210.
ment 13 (2), 29–53. Slack, N., 1994. The importance–performance matrix as a determinant of
Johnston, R., 1995. The determinants of service quality: satisfiers and dissatisfiers. improvement priority. International Journal of Operations and Production
International Journal of Service Industry Management 6 (5), 53–71. Management 14, 59–75.
Johnston, R., 1997. Identifying the critical determinants of service quality in retail Spector, P., 1992. Summated Rating Scale Construction: An Introduction. Sage,
banking: importance and effect. International Journal of Bank Marketing 15 (4), Newbury Park, CA.
111–116. Srivastava, R.K., Shervani, T.A., Fahey, L., 1998. Market-based assets and shareholder
Johnson, M.D., Nader, G., Fornell, C., 1996. Expectations, perceived performance, and value: a framework for analysis. Journal of Marketing 62 (1), 2–18.
customer satisfaction for a complex service: the case of bank loans. Journal of Stauss, B., Hentschel, B., 1992. Attribute-based versus incident-based measurement
Economic Psychology 17 (2), 163–182. of service quality: results of an empirical study in the German car service
Jones, C., 2004. Developing a scorecard for service quality. Management Services 48 industry. In: Kunst, P., Lemmink, J. (Eds.), Quality Management in Services. Van
(April), 8–13. Gorcum, Assen, pp. 59–78.
Kamakura, W.A., Russell, G.J., 1989. A probabilistic choice model for market Stevens, J., 1986. Applied Multivariate Statistics for Social Sciences. LEA, Hillsdale,
segmentation and elasticity structure. Journal of Marketing Research 26 NJ.
(November), 379–390. Studenmund, A.H., 2001. Using Econometrics, forth ed. Addison Wesley Longman,
Kano, N., 1984. Attractive quality and must-be quality. Hinshitsu: The Journal of the Boston, MA.
Japanese Society for Quality Control 14 (2), 39–48. Swan, J.E., Combs, L.J., 1976. Product performance and consumer satisfaction: a new
Lassar, W.M., Manolis, C., Winsor, R.D., 2000. Service quality perspectives concept. Journal of Marketing 40 (April), 25–33.
and satisfaction in private banking. Journal of Services Marketing 14 (3), Taylor, S.A., 1997. Assessing regression-based importance weights for quality
244–271. perceptions and satisfaction judgments in the presence of higher order and/or
Levesque, T., McDougall, G.H., 1996. Determinants of customer satisfaction in retail interaction effects. Journal of Retailing 73 (1), 135–159.
banking. International Journal of Bank Marketing 14 (7), 12–20. Taylor, S.A., Baker, T.L., 1994. An assessment of the relationship between service
Lindeman, R.H., Merenda, P.F., Gold, R.Z., 1980. Introduction to Bivariate and quality and customer satisfaction in the formation of consumers’ purchase
Multivariate Analysis. Scott Foresman, Glenview, IL. intentions. Journal of Retailing 70 (2), 163–178.
Lowenstein, M.W., 1995. Customer Retention: An Integrated Process for Keeping Varki, S., Colgate, M., 2001. The role of price perceptions in an integrated model of
Your Best Customers. ASQC Quality Press, Milwaukee, WI. behavioural intentions. Journal of Service Research 3 (February), 232–240.
Maddox, R.N., 1981. Two-factor theory and consumer satisfaction: replication and Vavra, T.G., 1997. Improving Your Measurement of Customer Satisfaction: A Guide
extension. Journal of Consumer Research 8 (1), 97–102. to Creating, Conducting, Analyzing And Reporting Customer Satisfaction
Manrai, L.A., Manrai, A.K., 2007. A field study of customers’ switching behaviour Measurement Program, Milwaukee, WI.
for bank services. Journal of Retailing and Consumer Service 14 (7), Wakefield, K.L., Blodgett, J.G., 1999. Customer response to intangible and tangible
208–215. service factors. Psychology and Marketing 16 (1), 51–68.

280 A. Arbore, B. Busacca / Journal of Retailing and Consumer Services 16 (2009) 271–280

Winstanley, M., 1997. What drives customer satisfaction in commercial banking. Zeithaml, Z.A., 2000. Service quality, profitability, and the economic worth of
Commercial Lending Review 12 (3), 36–42. customers: what we know and what we need to learn. Journal of the Academy
Wittink, D.R., Bayer, L.R., 1994. The measurement imperative. Marketing Research 6 of Marketing Science 28 (1), 67–85.
(4), 14–22. Zielke, S., 2008. Exploring asymmetric effects in the formation of retail
Yang, C., 2005. The refined Kano’s model and its application. Total Quality price satisfaction. Journal of Retailing and Consumer Services 15 (5),
Management 16 (10), 1127–1137. 335–347.