Most service organisations are, quite rightly, very interested in measuring the attitudes, beliefs and perceptions of their customers. For some departments (like marketing and sales), it is often one of their primary metrics. As a consequence, various research teams have attempted to devise measures that are comprehensive, reliable and robust. This short paper is a review of the major instruments available to practitioners and researchers.
Customer satisfaction is an assessment of how well a company’s products or services meet customer expectations. Customer satisfaction relates to a general and specific psychological evaluation of a customer’s experience of a product or service. It is well established that satisfied customers are key to long-term business success (McColl-Kennedy & Schneider, 2000) . The idea of the service-profit chain directly relates customer satisfaction to business success and profitability.
Customer satisfaction predicts customer retention, loyalty, and product repurchase. Customer satisfaction has also been suggested to have an impact on future product search activity, alterations in “hopping behaviour”, as well as trials of other available products in the sector. Keiningham and Vavra (2001) found that for every percentage increase in customer satisfaction, there is an average increase of 2.37% of return on investment. Further, when a customer is satisfied they spread this information, acting as marketers for the company.
Gitomer (1998) proposed that nearly one half of American business is built on “word-of-mouth” communication. Such findings demonstrate the importance of customer satisfaction. Whilst some of these ideas and findings are disputed, the measurement of customer reactions remains important. It is thus key to be able to assess these reactions, in order to determine methods of improving business effectiveness.
Measuring customer service is highly important, yet there is no universally accepted measurement scale. There are several theories proposing how we should approach the assessment of customer service. These include the expectancy-dis- confirmation approach, the performance-only approach, the technical and functional dichotomy approaches, the service quality vs. service satisfaction approach, and the attribute importance approach. Two of these approaches can be seen in Table 1.
However, measuring customer satisfaction is not simple partly because personal attitudes towards quality vary between individuals. What one may consider to be superior quality may be seen as average by another. Garvin (1984) concludes that “quality lies in the eyes of the beholder” (p. 41). However subjective this may be there is clearly considerable agreement between customers, particularly at extremes of good and bad service.
As assessing quality in service is such a complex task, over the years, service management literature has introduced a number of models to measure overall service quality (Seth, Deshmunkh, & Vrat, 2005; Senić & Senić, 2008) . The majority of these models are based on comparing what the customer expected from the service or product, and the actual service quality levels perceived. This idea was originally introduced by Grönroos (1982) . For instance, the service will be considered excellent if perceptions exceed expectations, but bad if it does not meet expectations (Vázquez et al., 2001) .
The aim of this paper is to review the various published measures of customer service.
2. The Measures
Three of the most common scales are documented below in Table 2. They seem to be widely used and reported in various academic papers.
Table 1. Two approaches to measurement.
Table 2. The most popular scales and their individual items.
Parasuraman et al. (1985) built on Grönroos’s idea, and formulated the five gaps model, which was the basis for a 22-item questionnaire assessing quality in service, the SERVQUAL (Parasuraman et al., 1988) (see Table 2). The five most critical dimensions of quality were identified as reliability, responsiveness, assurance, empathy and tangibility. To a large extent this model and measure which has been critiqued and updated has dominated the field. The reason is probably because the five dimensions are clear and important and the test is short and the items interpretable in many settings.
Many instruments have been introduced for measuring service quality since. The best known alternative to SERVQUAL is SERVPERF, constructed by Cronin and Taylor (1992) . Much discussion followed differentiating between the two models in order to conclude which one was more valid when predicting service quality.
A contrasting suggestion to Grönroos’s (1982) original idea of comparing expectations with experience is the idea that the measurement of service quality should only include customer perceptions (Caro & García, 2007) .
Grönroos (1982) suggests that service quality is comprised of two dimensions, functional and technical. The functional dimension relates to how a service is delivered, where the technical dimension refers to an output of the service, e.g., what the customer receives. These aspects of service quality have been found to impact on customer attitudes towards a brand as well as on future behavioural intentions (Dagger & Sweeney, 2006) . This suggests that perception of service quality is determined by functional quality, technical quality and corporate image: that is the image and reputation of the organisation for service in general. Fonesca (2009) supported this model, proposing that these three factors represent the main determinants of satisfaction. Therefore, it is argued, service quality leads to satisfaction directly.
However, when measuring satisfaction, many challenges occur (Maričić, 2008) . Satisfaction is a subjective measure of a customer’s perception of the quality of a product or service. The measurement of satisfaction also encompasses expectations of quality of a product or service. Bateson (1995) highlights the difficulties of measuring service quality due to the fact of its intangibility, heterogeneity, inseparability and perishability.
Table 3 describes a number of service quality and customer satisfaction models and scales. Each scale is comprised of different dimensions. Some of these are consistent across models, such as reliability (SERVQUAL, RSQS, SERVPERF, Weighted SERV-PERF, SERVQUAL-R), suggesting this is a key dimension of customer satisfaction and service quality. The scales vary in the number of dimensions included, from 2 - 10, varying in the different factors they think it is necessary to assess. Our research suggests that the SERVQUAL and the weighted SERV-PERF are the most used, however it is only possible to assess the published research in the area, and it may be that many of the other scales are widely employed throughout business yet not published in either the academic or trade literature.
These scales have not all been subjected to exacting and repeated psychometric assessment such as checking their dimensions through exploratory and confirmatory factor analysis, as well as the internal reliability (Cronbach’s Alpha) of those dimensions. More importantly perhaps in this area, there are few good norms so that those who use the measures can compare their data to that of a large group of representative service providers.
There are various ways to evaluate the different scales mentioned in the table. The can be divided into psychometric and practical criteria. Psychometricians are interested in such things as reliability, validity and dimensionality. They look for evidence that the various scale items factor or cluster into the various dimensions the authors proposed. They are also interested in the internal reliability of the questionnaire, but most of all the validity: that is the proof that the questionnaire data predicts actual behaviour like sales. Some scales have been put to the test while others have received much less attention. Whilst it is relatively cheap and easy to develop a model and questionnaire items it is much more expensive in terms of time and money to psychometrically evaluate the scale.
Table 3 shows that where measured the tests seem to have acceptable levels of internal reliability, though there is less work on test-retest reliability. Psy- chometricians would in this case be very interested in the factor structure of these measures using factor analysis and structural equation modelling to try to
Table 3. Research on these scales.
determine if there is a universal structure to the dimensions people use in the evaluation of services. So far there appears very little research of that kind. More importantly there appears no research to indicate which of the various scales in these measures relates most powerfully to the overall satisfaction of the customer
Further there is little evidence that the scales retain their properties when translated into different languages. Similarly, it would be of interest to examine demographic correlates of these scales to attempt to determine whether, for instance age, social class or gender were related to the different factors in customer evaluation.
There can be no validity without reliability but it appears that there is relatively little published evidence on the validity of these scales. This research is not easy and usually involves correlation customer service data and other salient data like sales, profit, store visits or repeat purchases over time to assess the expected relationship between these measures. More interesting is how much variance is accounted for by the customer service data compared to other issues, like price which may attract customers. This could answer the question of whether it is indeed a wise investment to spend money on attempts to improve customer service.
The second and not always closely related criterion is always practical. Table 3 shows some evidence of the applications of these scales. Practitioners want questions which are very short clear and easy for customers to complete and yet measure the dimensions they are really interested in. They are often happy to trade off length for psychometric reliability and add as well as remove items they think not relevant to them.
They also like to use the same measure for different purposes. Some are concerned that you may annoy customers by using such measures while other believe the opposite, namely that customers feel appreciative in being asked. Most seem to be used in the hospitality industries but that is changing.
There are inevitable gaps in Table 2 and Table 3. There may well be other measures that are in use and many other papers that attempt to assess these instruments. In that sense these tables are, and will always be preliminary, but as far as we know no one has attempted a review such as this.
Both those in pure and applied research want scales that are short, psychometrically sound and sensitive to the various facets/dimensions that make up the behaviour (customer service) that is assessed. The review has located what we believe are the most well known and used measures and the evidence for them. We hope to be able to update this report every so often as the literature and the use of these measures increases.