Abstract
Online auction is one of the successful business models in electronic commerce. In recent years, online auction has once again become an important part in many people’s lives, setting new records in transaction amounts and bringing huge profits to the owners of auction websites. Customer satisfaction has become an important issue for online auction. The main purpose of this study is to explore the relationship between e-service quality and satisfaction by using a catastrophe model. The latent variables and manifest variables are defined in the Structural Equation Model (SEM). The catastrophe model was used to developing marketing. The empirical results demonstrated that design is the key factors to impact on consumers to choose the online art auction web site.
Introduction
Traditional marketing research often relies on linear assumptions: a fixed improvement in service quality is expected to yield a proportional increase in customer satisfaction. However, consumer behavior in competitive digital landscapes, such as online art auctions, frequently exhibits non-linear characteristics. Customers may remain loyal despite declining service quality until a critical threshold is reached, at which point they abruptly switch to a competitor.
This paper addresses the failure of linear models to capture these "discontinuous" jumps in behavior. By applying the Cusp Catastrophe Model (CCM), the researchers examine how the interaction between e-service quality (the normal factor) and switching barriers (the splitting factor) determines sudden shifts in consumer loyalty between major platforms like Yahoo.com and Ruten.com.
Methodology
The study employed the Cusp Catastrophe Model, a topological framework used to model systems where continuous input changes lead to discontinuous output changes. The model defines a three-dimensional response surface characterized by a "pleat" or fold.
- State Variable ($x$): Customer Satisfaction/Loyalty.
- Normal Factor ($v$): E-service quality (including site design, security, and information accuracy).
- Splitting Factor ($u$): Switching barriers (perceived costs of moving to a new platform).
Data was collected via a web-based survey in June 2010, targeting 295 users of Yahoo.com and Ruten.com in Taiwan. The researchers used GEMCAT II (GEneral multivariate methodology for estimating CATastrophe models) for estimation, providing a robust confirmatory analysis of the non-linear dynamics.
Results and Discussion: The "Pleat" of Loyalty
The core finding is that the relationship between service quality and satisfaction is not always smooth. When switching barriers are high, a "fold" develops in the data. This fold represents a zone of bistability where a customer's state (satisfied or dissatisfied) depends on their history—a phenomenon known as the Hysteresis Effect.
As illustrated in Figure 3, consumers on the upper surface (e.g., satisfied with Yahoo) may tolerate some decrease in service quality due to high switching barriers. However, once they cross the edge of the fold, they "jump" to the lower surface (dissatisfaction/switching). Conversely, to win back a lost customer, service quality must improve significantly beyond the point where they originally left.
Empirical results also highlighted that website design is the most critical component of e-service quality influencing these transitions.
Strategic Marketing Applications
The catastrophe model provides a sophisticated framework for market segmentation. Instead of treating all customers the same, managers can identify "convertible consumers"—those positioned near the cusp of the fold with medium switching barriers.
As shown in Figure 4, marketing strategies can be tailored based on a consumer's position on the surface:
- Path A vs. Path B: Path A represents a sudden jump in satisfaction/adoption, while Path B represents a more gradual, continuous change.
- Targeting: Marketing programs should focus on "convertible consumers" who are most susceptible to service quality improvements or incentives.
- Retention: To maintain market share, managers must simultaneously enhance e-service quality and invest in Customer Relationship Management (CRM) to raise switching barriers, effectively "locking in" satisfied users.
Conclusion
This research confirms that consumer behavior in the online art auction market is inherently non-linear. By utilizing the Cusp Catastrophe Model, firms can better understand the threshold effects that lead to sudden customer churn or adoption. While the study is limited by its specific geographic focus (Taiwan) and sector (art auctions), it offers a powerful alternative to traditional linear marketing models, emphasizing the critical roles of website design and switching costs in maintaining a competitive edge.
Summary updated