ROLE OF PERSONNEL-RELATED SERVICE QUALITY …
Segmenting Members of a Retailer Loyalty Programme
Using Personnel-Related Service Quality Dimensions
Irena Ograjenšek
Vesna Žabkar
University of Ljubljana
Faculty of Economics
Kardeljeva pl. 17
1000 Ljubljana
Slovenia
irena.ograjensek@ef.uni-lj.si
vesna.zabkar@ef.uni-lj.si
Abstract
Ever since the advent of the smart loyalty cards, loyalty programmes (LP) have been transcending their traditional role as creators of exit barriers by transforming themselves into facilitators of customer data collection. Apart from demographic and socio-economic data, behavioural (transaction) as well as psychographic (survey) data are being collected for known LP members. Integral analysis of these data can be of immense value for retailers striving to improve service quality for different customer segments. Using the SERVQUAL model as a starting point, our study focuses on the issue of LP members’ segmentation on the basis of three distinct personnel-related perceived service quality dimensions (appearance, empathy and assurance), which serve as inputs into clustering process. For cluster profiling, selected demographic, socio-economic, and transaction variables are used. Apart from methodological issues, managerial implications of findings are discussed.
Keywords
loyalty programmes, service quality, customer data analysis, customer segmentation
Segmentiranje članov programa zvestobe na osnovi vpliva prodajnega osebja na zaznano kakovost storitev v trgovini na drobno
Povzetek
Programi zvestobe so tradicionalno sicer zasnovani kot sredstvo preprečevanja prebega kupcev h konkurenčnim podjetjem. Z razvojem informacijske in telekomunikacijske tehnologije pa postaja njihova glavna privlačnost vloga, ki jo igrajo v procesih zbiranja in analize podatkov kupcev – članov programa zvestobe. Integralna analiza demografskih, socioekonomskih, transakcijskih in anketnih podatkov kupcev omogoča podjetjem učinkovito usmerjanje procesov celovitega obvladovanja kakovosti storitev na ravni posameznih segmentov kupcev. Izhodišče pričujoče študije so na osnovi anketnih podatkov kupcev identificirane dimenzije zaznane kakovosti storitev, ki se nanašajo na prodajno osebje (videz, vzbujanje zaupanja in odzivnost) in služijo kot vhodni elementi v proces segmentacije kupcev. Profili segmentov so pripravljeni s pomočjo izbranih demografskih, socioekonomskih in transakcijskih podatkov. V prispevek vključujemo iz rezultatov analize izhajajoča priporočila vodstvu podjetja.
Ključne besede
programi zvestobe, kakovost storitev, analiza podatkov kupcev, segmentacija kupcev
1. INTRODUCTION
Companies in the most developed economies have been aware of the importance of product and process quality for many decades due to theoretical and practical contributions of quality experts such as Deming, Juran, Crosby, Feigenbaum, Ishikawa and Taguchi [Peace, 1993; Drummond, 1994; Hagan, 1994; Cole and Mogab, 1995; Swift, 1995; Bisgaard, 1998; Easton and Jarrell, 2000]. Service industries embraced the basic quality improvement ideas simultaneously with the manufacturing sector, yet neglected the use of statistical methods in quality improvement processes even more than their manufacturing counterparts. One of the reasons for such a state of affairs is given by differences in nature of services and manufactured goods. They have always been emphasised in the literature, especially with regard to measurability of service quality attributes and, consequently, characteristics of the measurement process. Therefore, the use of traditional statistical quality control toolbox in quality improvement of service processes has usually a priori been limited to the most basic of tools (e.g. control charts). This does not mean that complex statistical methods have not been used in the service sector. The toolbox developed by social sciences (containing e.g. exploratory and confirmatory factor analysis) has been proposed as the one to be of use in quality improvement of service processes by authors such as Parasuraman et al. [1985, 1988, 1994], Teas [1993, 1993a, 1994], Zeithaml et al. [1993], Cronin and Taylor [1992, 1994], Lytle et al. [1998], etc. For the better part, the toolbox consists of a series of instruments for measurement of perceived (subjective) service quality which, among other things, differ importantly with regard to applicable measurement scores. SERVQUAL as the most frequently applied model (proposed by Parasuraman et al.) is based on comparisons of customers’ quality expectations and perceptions.
Although our study does not classify as another attempt to prove the SERVQUAL generalisability, it does use this model as a starting point. However, our discussion focuses on the issue of retailer loyalty programme members’ segmentation from the viewpoint of those service quality dimensions, which pertain to retailer’s sales personnel. Data provided by a large Slovenian retailer are used in the project. Our paramount research goals are to find out how many loyalty programme member segments can be identified with regard to sales personnel-related perceived service quality dimensions, what are their characteristics, and what kind of managerial measures (if any) are necessary to improve the level of sales personnel-dependent perceived service quality.
2. RETAILER LOYALTY PROGRAMMES AND THEIR CHARACTERISTICS
Before going into a detailed discussion of our empirical research project, it is necessary to devote some attention to the issue of retailer loyalty programmes and their characteristics. A brief historical overview shows that loyalty programmes as we know them today have existed for a relatively short period of time: a little over twenty years. Large airlines were the first companies to start playing with the idea of creating special customer programmes with two distinctive features:
• they would include a large number of customers;
• activities conducted in the programme framework would stimulate members to increase their purchase frequency.
In 1981 American Airlines launched the programme called AAdvantage [Kasper et al., 1999]. It quickly turned into a source of competitive advantage and other airlines were forced to respond in kind. Further development is well known: loyalty programmes soon outgrew the confines of the airline industry. The expansion that started in closely linked industries (e.g. car rental and hospitality industry) later advanced into retail trade, banking and insurance, etc.
Hopf and Ograjenšek [1999] define a loyalty programme as a company’s organised and structured form of loyalty efforts. Rayner [1996] is more technical and speaks of a loyalty programme as a mechanism for identifying and rewarding loyal customers - those staying with the company for longest, purchasing most frequently and spending on average most per purchase.
A loyalty programme could be viewed either as a functional (marketing) or strategic (management) instrument, which includes various different measures to improve customer loyalty in the framework of a special customer club. Measures such as for example organisation of exclusive events, regular mailings (including monthly or quarterly information bulletins, birthday cards, invitations to exclusive events, etc.), have a sole purpose of emotionally involving and binding club members, and, consequently, increasing purchase frequency.
Practical value of loyalty programmes in continuous quality improvement of service processes can be determined from two points of view, namely business and methodological [Ograjenšek, 2002]:
• Business aspect historically precedes the methodological and could therefore also be called the traditional one. It is focused on creation and maintenance of loyalty as part of the defensive business strategy. Defensive strategy aims at reducing the number of exiting and switching customers. In other words, it tends to minimise customer turnover and maximise customer retention. One way of accomplishing these goals is to introduce switching costs [Bharadwaj et al., 1993]. The other is to “produce” highly satisfied customers. If positioned correctly, loyalty programmes create a long-term customer loyalty (manifested in repeat purchase), and do not turn into a relentless short-term chase after rewards. Members do not perceive programmes as devices, which limit their freedom of choice [Marshall, 1999]. Additionally, according to Bolton et al. [2000] although members are increasingly exposed to a complete spectrum of service experiences (including service failures) with the same company, they seem to be more tolerant of failures and less ready to switch to competition.
• Methodological aspect, which could also be called the modern one, has gained in importance with development of modern information technology and use of the so-called smart loyalty card for customer data collection. From this point of view, loyalty programmes could be regarded as an important data source for analyses necessary in continuous quality improvement of service processes.
The framework which brings both aspects together is Fornell’s [1992] classification of business strategies, which can be supplemented as shown in Figure 1.
Figure 1: Bringing the Business and Methodological Aspects of Loyalty Programmes Together in the Framework of Fornell’s Classification of Business Strategies
[pic]
While maintenance of loyalty (creation of switching barriers) is the focus of the business aspect of a loyalty programme, methodological aspect is accounted for by data generation and analysis. Figure 1 also shows that methodological aspect of loyalty programmes is of vital importance not only for creation and implementation of a defensive, but also to form an offensive business strategy. It could therefore be argued that due to recent advances in information technology, loyalty programmes are transcending their traditional role as creators of exit barriers for the customer. More and more, they are becoming facilitators of data collection and analysis. In this role, they help companies segmenting the customer database and reshaping its profile, focusing on those segments that can be served well for a profit. As Reichheld and Sasser [1990: 109] put it: “Achieving service quality doesn’t mean slavishly keeping all customers at any cost. There are some customers the company should not try to serve.”
It goes without saying that in retailing, the role sales personnel play is crucial for shaping customer perceptions of service quality and determining the level of customer satisfaction. The effect is probably even larger in the framework of a loyalty programme aiming at creation of long-term relationships, since liberally made and often repeated promises which are not kept create more room for development of negative feelings towards the company in question (especially if the customers feel trapped). On a more basic level, however, our research proposition is the following:
Proposition: Tangible and intangible characteristics of sales personnel (such as physical appearance, willingness to listen, assure and oblige, etc.) are the key to high levels of perceived service quality.
Our empirical research strives to verify this research proposition using a subset of items from the SERVQUAL scale.
3. SERVQUAL: AN OVERVIEW
SERVQUAL is a result of a systematic on-going study of service quality that begun in 1983. The model defines quality as the difference between customers’ expectations and perceptions with regard to the service delivered in the past. The respondents are asked to answer two sets of questions dealing with the same subject. One set of questions is general (e.g. quality of service in financial institutions), the other pertaining to a company in question (e.g. quality of service in bank X).
Respondents choose from a seven-point modified Likert scale to indicate the degree of their agreement with each of the given statements. For each of the items (service attributes), a quality judgement can be computed according to the following formula:
Perception ( Pi ) - Expectation ( Ei ) = Quality ( Qi ) (1)
The SERVQUAL score (perceived service quality) is obtained by the following equation:
[pic] (2)
The Pi ( Ei gap scores can be subjected to an iterative sequence of item-to-item correlation analyses, followed by a series of factor analyses to examine the dimensionality of the scale.
Using the oblique rotation that identifies the extent to which the extracted factors are correlated, Parasuraman et al. [1988] discovered five quality dimensions:
• tangibles: physical facilities, equipment, and appearance of personnel;
• reliability: ability to perform the promised service dependably and accurately;
• responsiveness: willingness to help customers and provide prompt service;
• assurance: knowledge and courtesy of employees and their ability to convey trust and confidence;
• empathy: caring and individualised attention that company provides its customers with.
As shown in Llosa et al. [1998], the SERVQUAL scale has been adopted as a standard by other researchers in the field of service quality both in business-to-business and mass markets. This, however, does not mean that the scale is not subject to constant re-examination and criticism concerning issues such as:
• object of measurement: Kasper et al. [1999] state that it is not clear whether the scale measures service quality or customer satisfaction;
• questionnaire wording: the same group of authors [Kasper et al., 1999] points out that it might be better not to use negatively worded questions in order to avoid interpretation problems;
• questionnaire length: Johns and Tyas [1996] deem the questionnaire in its original form too long;
• timing of questionnaire administration: the issue of whether to distribute the questionnaire before or after the service experience is discussed by Carman [1990], Smith [1995] as well as Johns and Tyas [1996];
• problems of Likert scale application: a comprehensive overview given in Krosnick and Fabrigar [1997] includes number and labeling of points, inclusion of the middle alternative, equality of distances between points, etc.;
• use of (Pi-Ei) difference scores: Teas [1993, 1993a, 1994] shows that increasing P-E scores may not always correspond to increasing levels of perceived quality;
• generalisation of service quality dimensions: Buttle [1995] lists a number of replication studies to illustrate that the number of distinct perceived quality dimensions found in replication studies conducted in different service industries varies from one to nine, thus the generalisation is low, or, as stated by Babakus and Boller [1992: 253], “the dimensionality of service quality may depend on the type of services under study”;
• static nature of the model: Haller [1998] points out that for a number of long-term service processes (such as e.g. education), both perceptions and expectations - and consequently quality evaluations - change in time, therefore a dynamic model of service quality should be developed.
Given the fact that ours was not to be a SERVQUAL replication study, we applied a shorter version of the questionnaire, concentrating on tangible and intangible characteristics of sales personnel in relation to perceived service quality. We included only the items of performance and decided to adopt a five-point Likert scale with the goal of increasing the response rate and response quality. With regard to the timing of questionnaire administration, we resolved to collect data away from the actual point of service delivery (unrelated to a single encounter).
4. EMPIRICAL RESEARCH
4.1 A Brief Introduction of the Retailer
The retailer whose loyalty programme members participated in our research project belongs to the group of those large Slovenian companies whose almost monopolistic status (market share above 70 per cent over the last several years) will probably be seriously eroded in the decade following the EU accession. The company has a widespread network of retail outlets where, as early as in 1996, more than 5,000 different products were available for sale. It is renowned for its environmental concerns. Apart from them, the company has also been involved into a number of cultural, educational, sport, and humanitarian projects.
4.2 A Brief Introduction of the Retailer Loyalty Programme
Rayner [1996] argues that there is considerable communality between the systems required to run payment card programmes, and those required to run customer loyalty programmes. The former can be easily amended to add customer loyalty benefits. She also points out that retailer-issued payment cards are typical differentiating services aimed at developing long-term loyalty. They might have a narrower appeal than loyalty programmes, but generate a stronger customer commitment.
To benefit from advantages of both loyalty programmes and payment card, retailers may pursue one of the two paths:
• they introduce the payment card first and either add loyalty benefits later, or run the card in conjunction with a separate customer loyalty programme;
• they capitalise on the infrastructure developed for a loyalty programme by expanding it into payment options.
In the year 1992, our retailer trial-issued a payment card to its own employees and employees of affiliated companies. The card was only accepted at retailer’s points of sale and was not promoted in any way. With this act, foundations for a long-term development of a loyalty programme were laid. After very good results of the trial card issue, the public nation-wide card launch followed a year later, and an independent unit (the Card Centre) was established as an IT-support and transaction authorisation centre. During the following years, in co-operation with several external partners, a number of loyalty initiatives were undertaken in the card promotion framework, bringing it closer and closer to the final transformation into a formal loyalty programme. Implementation of the final step (gradual replacement of bar-code cards with smart cards and transformation of the Card Centre into the Customer Care Centre) started in 2002.
4.3 Sample Selection Procedure and Sample Characteristics
Before taking a closer look at the sample characteristics, it is necessary to describe the sample selection procedure, which consisted of three steps:
• In the first step, a stratified random sample of 600 units was selected from the retailer database of 45,958 members (the ZIP code areas serving as strata).
• Cross-verification of each unit’s address and phone number was carried out in the second step. 36 units (or 6 per cent) had to be excluded from the sample because their phone numbers were either non-existent or could not be positively identified for various reasons (e.g. phone number registered under a different name).
• The remaining 564 units were contacted by phone in the three-day period from June 26th to June 28th, 2001. 201 or almost 40 per cent (39.7 per cent of the applicable sample) agreed to participate in the survey.
Although the final result is not a perfect stratified sample due to elimination of several selected units without replacement (because of temporal and financial project constraints), it comes close enough for the goals of the proposed empirical project.
Sample characteristics closely resemble characteristics of the population of loyalty programme members. As shown in Table 1, there are 62.7 per cent males and 37.3 per cent females in the sample, with the average age of 42.9 years. The youngest cardholder is 24 and the eldest 79 years of age.
Table 1: Gender Structure and Age Characteristics of the Cardholders’ Sample
|Gender |Number |% of Units |Average Age |Minimum Age |Maximum Age |
| |of Units | | | | |
|Male |126 | 62.7 |43.3 |24 |69 |
|Female | 75 | 37.3 |42.2 |24 |79 |
|Total |201 |100.0 |42.9 |24 |79 |
2.5 per cent of the respondents finished only the primary and 62.2 per cent the secondary school (in the population of cardholders, these percentages amount to 7.3 per cent and 65.9 per cent). The remaining 35.3 per cent held at least a college if not a university degree (26.9 per cent in the population of cardholders).
As in the cardholder population, a little more than 50 per cent of respondents came from the two largest Slovenian metropolitan areas. The average length of membership, however, was slightly longer (3.5 as opposed to 2.9 years).
More than 80 per cent of respondents were owners or co-owners of a house or an apartment; roughly the same percentage of them were married or living together with a partner, and almost 70 per cent had to provide for one or two dependent persons.
4.4 A List of Variables
Apart from demographic and socio-economic variables used to describe population and sample characteristics and to profile clusters later on, the following transaction and survey variables were used in the analysis:
• survey variables (perceived service quality items pertaining to retailer’s sales personnel measured on the five-point Likert scale with options 1 – strongly disagree, 2 – disagree, 3 – neither disagree nor agree, 4 – agree, 5 – strongly agree):
– SQ_1: Retailer’s sales personnel are properly dressed.
– SQ_2: Retailer’s sales personnel are neat.
– SQ_3: Retailer’s sales personnel are always friendly.
– SQ_4: Retailer’s sales personnel can be trusted.
– SQ_5: Retailer’s sales personnel are very busy, therefore it is understandable that they cannot help me immediately.
– SQ_6: Retailer’s sales personnel are very busy, therefore it is understandable that they cannot spend a lot of time dealing with my requests.
• transaction (behavioural) variables (for the period January – June 2001):
– total amount spent in the six-month period using the payment card issued by the retailer;
– number of retailer’s outlets visited in the six-month period;
– maximum number of cardholder’s visits to one outlet in the six-month period;
– total number of visits to retailer’s outlets in the six-month period.
Following is a brief description of methodology used in the empirical project.
4.5 Methodology
To define the underlying structure in the data matrix of service quality perceptions we used factor analysis which enabled us to identify the separate dimensions of the structure and determine the extent to which each variable was explained by each dimension. Oblimin with Kaiser normalisation rotation method was used due to expected correlation among factors. Significant loadings were interpreted. Factor analysis was further used for data reduction by calculating scores for each underlying dimension and substituting them for the original variables [Hair et al., 1998].
Factor scores were then applied to group respondents into clusters, which should exhibit high within-cluster homogeneity and high between-cluster heterogeneity. Distance measures of similarity (Euclidean distance) were applied. A combination of hierarchical and non-hierarchical clustering algorithms was employed (hierarchical method was used to specify cluster seeds for a non-hierarchical method). Cluster analysis respecification showed that one of the observations had to be deleted as an outlier and clustering algorithm repeated. Finally, the clusters were interpreted and named, followed by validation and profiling of the clusters.
5. RESULTS
Perceptions of service quality pertaining to retailer’s sales personnel (6 survey items) were factor-analysed to determine the underlying factors related to the SERVQUAL instrument. Correlation matrix of these variables showed that over half of the correlations were significant at the 0.01 level (see Table 2).
Table 2: Pearson Correlation Coefficient Matrix (with p-Values in Parentheses)
|Item |SQ_1 |SQ_2 |SQ_3 |SQ_4 |SQ_5 |SQ_6 |
|SQ_1 |1.000 |0.481 |0.237 |0.286 |0.094 |0.205 |
| | |(0.000) |(0.001) |(0.000) |(0.189) |(0.004) |
|SQ_2 |0.481 |1.000 |0.261 |0.401 |0.108 |0.227 |
| |(0.000) | |(0.000) |(0.000) |(0.133) |(0.001) |
|SQ_3 |0.237 |0.261 |1.000 |0.432 |0.266 |0.209 |
| |(0.001) |(0.000) | |(0.000) |(0.000) |(0.003) |
|SQ_4 |0.286 |0.401 |0.432 |1.000 |0.259 |0.287 |
| |(0.000) |(0.000) |(0.000) | |(0.000) |(0.000) |
|SQ_5 |0.094 |0.108 |0.266 |0.259 |1.000 |0.516 |
| |(0.189) |(0.133) |(0.000) |(0.000) | |(0.000) |
|SQ_6 |0.205 |0.227 |0.209 |0.287 |0.516 |1.000 |
| |(0.004) |(0.001) |(0.003) |(0.000) |(0.000) | |
The Kaiser-Meyer-Olkin Measure of Sampling Adequacy with value of 0.700 was in the acceptable range. Bartlett's Test of Sphericity (216.4, df. 15, Sig.0.00) showed that non-zero correlations existed at the significance level of 0.000. This provided an adequate basis for proceeding with the factor analysis.
The first step in the factor analysis procedure was to select the number of components to be retained for further analysis. The importance of each component as well as their relative explanatory power as expressed by their eigenvalues were analysed. The scree test indicated that three factors might be appropriate. Although the eigenvalue for the third factor was low (0.755) relative to the latent root criterion value of 1.0, we considered inclusion of this factor as well. The three factors represented 53 per cent of the total variance of the six variables (two factors accounted for 48 per cent of the variance).
The size of communalities (see Table 3) shows variance in a particular variable accounted for by the three-factor solution. Extraction method used was the Principal Axis Factoring, followed by the Oblimin rotation method with Kaiser Normalisation.
Table 3: Communalities
|Variable Code |Variable Description |Communalities |
|SQ_1 |Retailer’s sales personnel are properly dressed. |0.436 |
|SQ_2 |Retailer’s sales personnel are neat. |0.666 |
|SQ_3 |Retailer’s sales personnel are always friendly. |0.478 |
|SQ_4 |Retailer’s sales personnel can be trusted. |0.520 |
|SQ_5 |Retailer’s sales personnel are very busy, therefore it is understandable that |0.540 |
| |they cannot help me immediately. | |
|SQ_6 |Retailer’s sales personnel are very busy, therefore it is understandable that |0.554 |
| |they cannot spend a lot of time dealing with my requests. | |
As shown in Table 4, each factor is composed of variables with loadings of 0.50 or higher. Variable SQ_1 and SQ_2 loaded significantly on Factor 1, variables SQ_3 and SQ_4 on Factor 3 and variables SQ_5 and SQ_6 on Factor 2. All three pairs of variables vary together (for all three pairs, both variables are of the same sign, suggesting that these perceptions are quite similar among respondents). Factor 1 seemed to capture personnel appearance, Factor 2 tapped into empathy and Factor 3 revealed assurance.
Table 4: Pattern Matrix
|Variable Code |Variable Description |Personnel Appearance|Empathy |Assurance |
|SQ_2 |Retailer’s sales personnel are neat. |0.79 |- 0.00 |- 0.05 |
|SQ_1 |Retailer’s sales personnel are properly |0.62 |0.03 |- 0.05 |
| |dressed. | | | |
|SQ_6 |Retailer’s sales personnel are very busy, |0.15 |0.73 |0.09 |
| |therefore it is understandable that they | | | |
| |cannot spend a lot of time dealing with my | | | |
| |requests. | | | |
|SQ_5 |Retailer’s sales personnel are very busy, |- 0.14 |0.70 |- 0.14 |
| |therefore it is understandable that they | | | |
| |cannot help me immediately. | | | |
|SQ_3 |Retailer’s sales personnel are always |- 0.01 |0.00 |- 0.69 |
| |friendly. | | | |
|SQ_4 |Retailer’s sales personnel can be trusted. |0.20 |0.06 |- 0.55 |
In our case it was reasonable to expect that perceptual dimensions would be correlated (see Table 5). The application of an oblique rotation was thus justified. Validation of factor analysis was performed by splitting the sample into two sub samples and re-estimating the factor model to test for comparability [Hair et al., 1998]. The results proved to be stable within our sample.
Table 5: Factor Correlation Matrix
|Factor |Factor Label |1 |2 |3 |
|1 |Personnel appearance |1.00 |0.30 |- 0.59 |
|2 |Empathy |0.30 |1.00 |- 0.51 |
|3 |Assurance |- 0.59 |- 0.51 |1.00 |
Factor scores for each of the three factors were saved. Each of the respondents was therefore assigned three new variables (factor scores for Factors 1–3) that replaced the original six variables in the cluster analysis. Our next objective was to segment customers into groups with similar perceptions of identified service quality dimensions (personal appearance, empathy and assurance).
The sample had been examined for the outliers and one strong candidate for deletion was found. Given that the set of three factor scores was metric, squared Euclidean distances were chosen as the similarity measure. The standardisation of the variables was not undertaken because all variables were measured on the same (five-point) scale. The within-case standardisation was not appropriate, as the magnitude of the perceptions was an important element of the segmentation objectives.
In the clustering process, we decided to combine the hierarchical and non-hierarchical cluster methods. First we used the hierarchical procedure to identify the appropriate number of clusters. Then we used the non-hierarchical procedure to fine-tune the results.
In hierarchical cluster analysis, Ward’s algorithm was chosen to minimise the within-cluster differences and to avoid problems associated with linkage methods. The within-cluster sum of squares coefficient in agglomeration schedule revealed large increases in going from four to three clusters (223.35 - 180.71 = 42.64), three to two clusters (310.08 - 223.35 = 86.73) and two to one cluster (579.00 - 310.08 = 268.92). Because the largest increases were observed when going from three to two clusters and from two to one cluster, the three-cluster solution was selected (from a manageable number of clusters in the range of two to five). Furthermore, a three-cluster solution was suggested based on the visual representation by a dendrogram. The absolute and relative sizes of the three clusters are given in Table 6.
Table 6: Number and Percentage of Cases in Each Cluster
|Cluster |Number of Cases |% of Cases |
|1 |115 | 59.6 |
|2 |37 | 19.2 |
|3 |41 | 21.2 |
|Total |193 | 100.0 |
The non-hierarchical technique was then applied, using cluster centroids as seed points. Identified groups were of the same size as the groups resulting from the hierarchical clustering procedure. A three-cluster solution was thus confirmed, with Clusters 2 and 3 being of the similar size (containing about 20 per cent of respondents each).
Interpretation and profiling of clusters was provided through the mean values (centroids) on each of the three rating variables (see Table 7). It could be noted that members of Cluster 1, which is the largest cluster, focus their attention, relative to members of Cluster 2 and Cluster 3, on personnel appearance and empathy. Quite the contrary can be stated for members of Cluster 2 who are focusing on assurance and not on personnel appearance or empathy. For members of Cluster 3, the focus is more on personnel appearance and less on assurance than Cluster 2, however also less on empathy than Cluster 1.
Table 7: Final Cluster Centres
|Factor |Cluster 1: |Cluster 2: |Cluster 3: |
| |High Perceived Service |Low Perceived Service |Low Perceived Empathy |
| |Quality |Quality | |
|Personnel appearance | 0.38 |- 1.29 | 0.22 |
|Empathy | 0.56 |- 0.61 |- 0.97 |
|Assurance |- 0.50 | 1.09 | 0.31 |
For the profiling stage, we focused on demographic and behavioural (transaction) variables not included in the cluster solution (see Tables 8 and 9).
Table 8: Comparisons for Nominal and Ordinal Variables,
Description of Clusters (n = 193)
|Variable |Variable Value |Number of Cases |Total |
| | |Cluster 1 |Cluster 2 |Cluster 3 | |
|Gender |Male |73 |28 |22 |123 |
| |Female |42 |9 |19 |70 |
| |Total |115 |37 |41 |193 |
| |Maribor |18 |10 |7 |35 |
| |Other |60 |13 |20 |93 |
| |Total |115 |37 |41 |193 |
| |Secondary school |81 |15 |24 |120 |
| |College or university |31 |21 |16 |68 |
| |Total |115 |37 |41 |193 |
| |Other |14 |5 |9 |28 |
| |Total |113 |36 |40 |189 |
| |Tenant |19 |6 |6 |31 |
| |Total |115 |37 |41 |193 |
|Age |1 |115 | 43.03 | 10.02 | 0.93 |
| |2 |37 | 41.00 | 9.98 | 1.64 |
| |3 |41 | 42.85 | 8.85 | 1.38 |
| |Total |193 | 42.60 | 9.76 | 0.70 |
| |2 |37 | 4.05 | 1.94 | 0.32 |
| |3 |41 | 3.56 | 1.95 | 0.30 |
| |Total |193 | 3.51 | 2.15 | 0.15 |
| |2 |36 | 1.19 | 0.92 | 0.15 |
| |3 |40 | 1.35 | 0.86 | 0.14 |
| |Total |191 | 1.35 | 1.01 | 0.07 |
| |2 |37 | 86,755.27 | 86,579.13 | 14,233.52 |
| |3 |41 | 89,863.99 | 73,467.65 | 11,473.72 |
| |Total |193 | 103,553.60 | 131,603.24 | 9,473.01 |
| |2 |37 | 5.19 | 4.25 | 0.70 |
| |3 |41 | 6.78 | 9.20 | 1.44 |
| |Total |193 | 5.65 | 6.11 | 0.44 |
| |2 |37 | 7.68 | 7.51 | 1.24 |
| |3 |41 | 11.17 | 15.91 | 2.49 |
| |Total |193 | 12.75 | 15.08 | 1.09 |
| |2 |37 | 14.89 | 12.81 | 2.11 |
| |3 |41 | 24.83 | 44.24 | 6.91 |
|Total |193 | 22.81 | 28.29 | 2.04 | |
The F-ratios showed differences in the group means for several behavioural (transaction) variables:
• Length of membership in the loyalty programme: F-ratio = 1.69, Sig. 0.19;
• Maximum number of cardholder’s visits to one retailer’s outlet (the most frequently used outlet by a given cardholder) in the six-month period: F-ratio =3.64, Sig. 0.03;
• Total number of visits to retailer’s outlets in the six-month period: F-ratio = 1.81, Sig. 0.17.
Here, the profiling process indicated that members of Cluster 1, which rated higher on perceived personnel appearance and empathy, tend to have a shorter length of membership in the company’s loyalty program, visit one (“the preferred”) of the company’s outlets significantly more often than members of Cluster 2, and are characterised by an overall higher total number of visits to retailer’s outlets than members of Cluster 2.
As far as Cluster 2 (which contains loyalty programme members focusing primarily on assurance) is concerned, it could be ascertained that of all three clusters, this one is characterised by the longest average length of membership on one, and the lowest average number of visited outlets as well as the lowest average number of visits to retailer’s outlets in the six-month period on the other hand. Total amount spent is also the lowest in this group of loyalty programme members, although differences among clusters are not statistically significant. Interestingly, Cluster 2 is also marked by the lowest percentage of females (24 per cent).
Cluster 3 with low perceived empathy (caring and individual attention provided to customers) is - controversially - characterised both by larger than average number of visited outlets and larger than average number of visits to retailer outlets in the six-month period. Given the largest percentage of females in this cluster (46 per cent) this comes as no surprise since research indicates that women are prepared to visit an outlet several times to check on alternatives [Barletta, 2003; Quinlan, 2003]. Also, female customers might appreciate more help from retailer’s sales personnel and therefore show less understanding for them being busy, not being able to help immediately, or not being able to spend some time dealing with their requests.
6. MANAGERIAL IMPLICATIONS
As we pointed out earlier, we believe that in retailing in general, and in the framework of a loyalty programme in particular, the role sales personnel play is crucial for shaping customer perceptions of service quality and determining the level of customer satisfaction. In our empirical project we investigated tangible and intangible characteristics of sales personnel (such as physical appearance, willingness to listen, assure and oblige, etc.) in relationship to perceived service quality. In order to exclude possibility of sales personnel bias, we collected the customer data on perceived service quality away from the point of sale. All results of the empirical project discussed in this paper are in line with our research proposition.
Managerial implications of our findings can be summarised as follows:
• Cluster 1 represents the solid membership base with high perceived quality and high number of visits to retailer outlets on one, yet with relatively shorter length of membership in the loyalty programme on the other hand. The main managerial challenge in this group would be keeping the members happy and in spending mood. Giving more emphasis on friendliness of sales personnel as well as their ability to convey trust could be the key.
• Cluster 2 are disgruntled loyalty programme members (predominantly male with high education), who “had been there and seen it all”. They cannot be bought with friendliness and trust without delivery; they need to be regularly invited back to retailer outlets and shown that their individual needs can be taken care of immediately.
• Cluster 3 members (females form the majority in this group) do not need any further push to visit several retail outlets of this particular company a number of times. They should, however, be well attended to once in the shop. Sales personnel should be prepared to dedicate enough time to answer their questions and help them find solutions to their problems.
A loyalty programme is defined as a company’s organised and structured form of loyalty efforts. Can these efforts produce negative effects? Not directly, perhaps, but sometimes indirect influences turn out to be the ones that are most problematic. Our empirical findings indicate that both the retailer in question and retailers in general should keep an eye on links between the length of membership and behavioural (transaction) variables to avoid the negative “loyalty programme effect”. The retailer ( loyalty programme member relationships might go stale due to predictability of the shopping experience. Also, interactions with the same sales personnel might become a put-off for long-term members. Carefully timed promotions and events, thorough employee training and rotation among outlets, as well as continuous analysis of customer data could help prolong the active stage in the loyalty programme member lifecycle.
7. THE FUTURE OUTLOOK
The results of our empirical analysis are a snapshot at a given point in time for a known retailer and should be interpreted with regard to this limitation. The variables used in the analysis were cut out of the broader model of SERVQUAL and give us only a limited view of the service quality dimensions. Several presented findings, however, might be of general interest to retailers involved in loyalty programmes.
The following issue would certainly need further investigation: in what ways could customer data analysis of current loyalty programme members be applied to identify specific segment behaviour and needs in order for management to act upon them with the goal of improving perceived service quality level, and, ultimately, the bottom line. In the future, we hope to broaden the scope of our analysis in this regard.
Further analytical challenges include introduction of the time perspective (to account for effects of implemented managerial measures against the negative “loyalty programme effect”) as well as predictions of profiles and behaviour of potential new members based on customer data analysis of current loyalty programme members (to improve the efficiency of direct marketing activities). Additionally, we would like to analyse the influence of demographic shifts in the population (longer life expectancy, higher percentage of females in higher age classes, etc.) both on effectiveness of loyalty programmes and on personnel-related service quality dimensions. Finally, the general trend towards multi-company and multi-industry loyalty programmes (European examples include Airmiles in the Netherlands, Payback in Germany as well as Nectar in the UK) should present us with interesting personnel-related quality problems to tackle, starting with the most pressing issue of employees’ dilemma whether to be loyal to their respective companies or the loyalty programme in question.
8. REFERENCES
Babakus, E. and G.W. Boller, 1992, ‘An Empirical Assessment of the SERVQUAL Scale’, Journal of Business Research, Vol.24, No.3, pp.253-268.
Barletta, M., 2003, Marketing to Women: How to Understand, Reach and Increase Your Share of the World’s Largest Market Segment, Chicago: Deaborn Trade Publishing.
Bharadwaj, S.G., P.R. Varadarajan and J. Fahy, 1993, ‘Sustainable Competitive Advantage in Service Industries’, Journal of Marketing, Vol.57, No.4, pp.83-99.
Bisgaard, S., 1998, ‘The Role of Scientific Method in Quality Management’, in Boutellier, R. and W. Masing, eds., Qualitätsmanagement an der Schwelle zum 21. Jahrhundert. Festschrift für Hans Dieter Seghezzi zum 65. Geburtstag, München, Wien: Carl Hanser Verlag, pp.325-345.
Bolton, R.N., P.K. Kannan and M.D. Bramlett, 2000, ‘Implications of Loyalty Program Membership and Service Experiences for Customer Retention and Value’, Journal of the Academy of Marketing Science, Vol.28, No.1, pp.95-108.
Buttle, F.A., 1995, ‘What Future for SERVQUAL?’, in Bergadaà, M., Proceedings of the 24th European Marketing Academy Conference, Cergy-Pontoise: ESSEC, pp.211-230.
Carman, J.M., 1990, ‘Consumer Perceptions of Service Quality: An Assessment of the SERVQUAL Dimensions’, Journal of Retailing, Vol.66, No.2, pp.33-55.
Cole, W.E. and J.W. Mogab, 1995, The Economics of Total Quality Management: Clashing Paradigms in the Global Market, Cambridge (MA): Blackwell Publishers.
Cronin, J.J.Jr. and S.A. Taylor, 1994, ‘SERVPERF Versus SERVQUAL: Reconciling Performance-Based and Perceptions-Minus-Expectations Measurement of Service Quality’, Journal of Marketing, Vol.58, No.1, pp.125-131.
Cronin, J.J.Jr. and S.A. Taylor, 1992, ‘Measuring Service Quality: A Reexamination and Extension’, Journal of Marketing, Vol.56, No.3, pp.55-68.
Drummond, H., 1994, The Quality Movement, New Jersey: Nichols Publishing.
Easton, G.S. and S.L. Jarrell, 2000, ‘Patterns in the Deployment of Total Quality Management. An Analysis of 44 Leading Companies’, in Cole, R.E. and W.R. Scott, eds., The Quality Movement & Organization Theory, Thousand Oaks: Sage Publications, Inc., pp.89-130.
Fornell, C., 1992, ‘A National Customer Satisfaction Barometer: The Swedish Experience’, Journal of Marketing, Vol.56, No.1, pp.6-21.
Hagan, J., 1994, Management of Quality. Strategies to Improve Quality and the Bottom Line, Milwaukee: ASQC Quality Press.
Hair, J.F. jr, R.E. Anderson, R.L. Tatham and W.C. Black, 1998, Multivariate Data Analysis, Upper Saddle River: Prentice Hall.
Haller, S., 1998, Beurteilung von Dienstleistungsqualität. Dynamische Betrachtung des Qualitätsurteils im Weiterbildungsbereich, Wiesbaden: Deutscher Universitäts-Verlag and Gabler Verlag.
Hopf, F. and I. Ograjenšek, 1999, Dienstleistungskonzept Loyalty Partner, St. Gallen: University of St. Gallen. [unpublished]
Johns, N. and P. Tyas, 1996, ‘Use of Service Quality Gap Theory to Differentiate Between Foodservice Outlets’, The Service Industries Journal, Vol.16, No.3, pp.321-346.
Kasper, H., P. van Helsdingen and W. de Vries Jr., 1999, Services Marketing Management. An International Perspective, Chichester: John Wiley & Sons.
Krosnick, J.A. and L.R. Fabrigar, 1997, ‘Designing Rating Scales for Effective Measurement in Surveys’, in Lyberg, L., P. Biemer, M. Collins, E. de Leeuw, C. Dippo, N. Schwarz and D. Trewin, eds., Survey Measurement and Process Quality, New York: John Wiley & Sons, Inc., pp.141-164.
Llosa, S., J. Chandon and C. Orsingher, 1998, ‘An Empirical Study of SERVQUAL’s Dimensionality’, The Service Industries Journal, Vol.18, No.2, pp.16-44.
Lytle, R.S., P.W. Hom and M.P. Mokwa, 1998, ‘SERV*OR: A Managerial Measure of Organizational Service-Orientation’, Journal of Retailing, Vol.74, No.4, pp.455-489.
Marshall, R., 1999, ‘Will Vampire Customers Bite Back?’, Customer Loyalty Today, Vol.6, No.2, pp.12-13.
Ograjenšek, I., 2002, Business Statistics and Service Excellence: Applicability of Statistical Methods to Continuous Quality Improvement of Service Processes, Ljubljana: University of Ljubljana, Faculty of Economics. [doctoral thesis]
Parasuraman, A., V. Zeithaml and L.L. Berry, 1994, ‘Reassessment of Expectations as a Comparison Standard in Measuring Service Quality: Implications for Further Research’, Journal of Marketing, Vol.58, No.1, pp.111-124.
Parasuraman, A., V. Zeithaml and L.L. Berry, 1988, ‘SERVQUAL: A Multiple-Item Scale for Measuring Consumer Perceptions of Service Quality’, Journal of Retailing, Vol.64, No.1, pp.12-40.
Parasuraman, A., V. Zeithaml and L.L. Berry, 1985, ‘A Conceptual Model of Service Quality and Its Implications for Future Research’, Journal of Marketing, Vol.49, Nr.4, pp.41-50.
Peace, G.S., 1993, Taguchi Methods. A Hands-On Approach, Reading: Addison-Wesley Publishing Company.
Quinlan, M.L., 2003, Just Ask a Woman: Cracking the Code of What Women Want and How They Buy, New York: John Wiley & Sons, Inc.
Rayner, S., 1996, Customer Loyalty Schemes. Effective Implementation and Management, London: Financial Times Retail & Consumer Publishing.
Reichheld, F.F. and W.E. Sasser Jr., 1990, ‘Zero Defections: Quality Comes to Services’, Harvard Business Review, Vol.68, No.5, pp.105-111.
Smith, A.M., 1995, ‘Measuring Service Quality: Is SERVQUAL Now Redundant?’, Journal of Marketing Management, Vol.11, No.1-3, pp.257-276.
Swift, J.A., 1995, Introduction to Modern Statistical Quality Control and Management, Delray Beach: St. Lucie Press.
Teas, R.K., 1994, ‘Expectations as a Comparison Standard in Measuring Service Quality: An Assessment of a Reassessment’, Journal of Marketing, Vol.58, No.1, pp.132-139.
Teas, R.K., 1993, Consumer Expectations and the Measurement of Perceived Service Quality’, Journal of Professional Services Marketing, Vol.8, No.2, pp.33-54.
Teas, R.K., 1993a, ‘Expectations, Performance Evaluation, and Consumers’ Perceptions of Quality’, Journal of Marketing, Vol.57, No.4, pp.18-34.
Zeithaml, V.A., L.L. Berry and A. Parasuraman, 1993, ‘The Nature and Determinants of Customer Expectations of Service’, Journal of the Academy of Marketing Science, Vol.21, No.1, pp.1-12.
................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.