Understanding the Antecedents of Customer Loyalty by Applying Structural Equation Modeling

Understanding the Antecedents of Customer Loyalty by Applying Structural Equation Modeling

Gülhayat Gölbaşı Şimşek, Fatma Noyan Tekeli
DOI: 10.4018/978-1-4666-6635-1.ch025
(Individual Chapters)
No Current Special Offers


The objective of this chapter is to form a model of customer loyalty for supermarkets in the Turkish retailing sector, which investigates the extent to which customer loyalty is influenced by customer perceptions of service quality, customer perceptions of product quality, comparative price perceptions, discount perceptions, value perceptions, and customer satisfaction. Structural equation modeling has been used to analyze the data collected from 1530 customers of four major supermarket chains in Turkey. After building a measurement model for customer loyalty and its potential antecedents, the relationships are examined. The direct effects of customer satisfaction, comparative price perceptions, and discount perceptions on customer loyalty; value perceptions, comparative price perceptions, and service quality perceptions on customer satisfaction; comparative price perceptions, discount perceptions, product quality perceptions, and service quality perceptions on value perceptions; and discount perceptions and service quality perceptions on product quality perceptions are empirically supported.
Chapter Preview


The retailing sector in Turkey is witnessing an intensive competition. More than 145 supermarket chains having more than 10 stores are in permanent competition in Turkish retailing sector. The growth of market, which over 12089 stores existed in 2012, increased dramatically so that over 5000 stores entered into Turkish market in the period of 2012-2014. With this increasing competition in the sector, customer preferences for retailing practices are also increasing and changing fast. So, customer loyalty becomes a central focus for retailers to maintain market shares.

Understanding customer needs and the process making them loyal customers can provide subjective performance measures for retailers. These performance criteria such as information and knowledge ownership, efficient use of technology, customer portfolio, customer satisfaction, product/service quality, customer loyalty, assistance to community and environment which are more sophisticated, less concreted, and more difficult to measure, have taken the place of the traditional profitability criterion for retailers (Şimşek & Noyan, 2009). These performance criteria especially customer loyalty have become more important concepts in positioning problem in marketing. Initially, customer loyalty provides necessary and vital information to retailers about their own status in the market to plan the future and to take action for the deficiencies. In this respect, customer loyalty refers to retailer’s self-recognition.

The conventional definition of customer loyalty is customer’s intention to repurchase a particular product or services in the future repeatedly (Jones & Sasser, 1995). Customer loyalty can also be described incorporating its behavioral and attitudinal dimensions, which is a reciprocal action or coaction between customers’ relative attitudes towards a store or brand alongside their repurchase behaviors towards that brand or store (Dick & Basu, 1994). The more comprehensive characteristics of customer loyalty have been described as usually choosing the same store, frequent repurchase intention, a high impact of a positive word of mouth, and willingness or voluntariness to pay higher prices (Topçu & Uzundumlu, 2009).

According to Rhee and Bell (2002), store health largely depends on customer loyalty. Customer loyalty can be described as long-term or short-term loyalty. Long-term loyal customers not easily change their store and product choice while short-term loyal customers do not sustain when they find a better opportunity (Chang & Tu, 2005). Customer loyalty provides a desirable atmosphere for firms because it makes flow of profit steady, reduces marketing costs, raises referrals, and promotes price premiums. Loyal customers are not easily convinced by the promotional efforts of competitors (Reichheld, 1996, Sirohi et al., 1998; Reinartz & Kumar, 1999; Yi & Jeon, 2003). The most attractive property is that loyal customers generate a stable portfolio for the products or services of a retailer. As stated by Wallace et al. (2004), a small shift in customer retention rates can make a large difference in earnings, and this influence accelerates over time. Virtually, the cost of getting new customers are larger than the cost of keeping existing customers; the difference between the costs shows that improvement in existing customer loyalty is critical (Wright & Sparks,1999; Yi & La, 2004).

Increasing number of the both large-scale supermarkets chains and the stores belonging to them in the Turkish retailing sector, entails to emit their claims about providing better service and more economical products to customers by using their scale advantage (Okumuş & Temizler, 2006). Dramatic increase in the number of supermarkets has created an important competitive environment in terms of price advantage, product and service quality. In order to attract customers to their own supermarkets, managers have conducted marketing campaigns and expect that these campaigns really help in attracting customers to their own supermarkets. Retailers want to know and understand the extent to which customer loyalty is improved by their efforts.

Key Terms in this Chapter

Confirmatory Factor Analysis: Used to build measurement model, thus, the effects of measurement errors are lessened through multiple indicators for a construct (factor or latent variable). Confirmatory factor analysis is more restrictive form of exploratory factor analysis (EFA). EFA has some assumptions such that all factors to be loaded on all observed variables (indicators, items, or measures), all factors to be correlated or uncorrelated and, error terms to be uncorrelated. These assumptions are required for EFA estimation but they are not fair. In CFA, the parameters supported by the underlying theory are estimated and investigated rather than the required but arbitrary parameters in EFA.

Independent Clusters Model: A type of common factor models in which they are group factors defined by distinct cluster of indicators. Group factors can be correlated or uncorrelated. Each indicator is loaded by just one group factor.

Discriminant Validity: The extent to which a construct is distinct from the other constructs within the model. There are no collapsing factors explaining variance in the indicator variables for each construct.

Composite Reliability: Composite reliability (CR), ? , which has been also referred to as McDonald’s ? coefficient, is obtained by combining all of the true score variances and covariances in the composite of indicator variables related to constructs, and by dividing this sum by the total variance in the composite.

Average Variance Extracted: A measure to assess convergent validity. Similar to explained variance in EFA, AVE is the average amount of variance in indicator variables that a construct is managed to explain. AVE for each construct can be obtained by sum of squares of completely standardized factor loadings divided by this sum plus total of error variances for indicators. For the completely standardized solution all indicator and latent variables are scaled to have unit variance.

Structural Equation Modeling: A complicated analysis that envelopes and uses the concepts of regression analysis, path analysis and confirmatory factor analysis, which they can be regarded as special types of SEM. However, SEM is implied as confirmatory, rather than data driven exploratory modeling; thus, it is appropriate for model testing, rather than model development. SEM comprises of two components, a measurement model and a structural model.

Structural Model: Structural part of the SEM model expresses the relationships among latent variables and disturbances (errors in the equations). Structural model is able to estimate the relationships among latent variables, is able to test overall model in addition to individual paths, and is able to model disturbances.

Complete Chapter List

Search this Book: