Analysis of selling strategies in electronic marketplaces
Date of Completion
Business Administration, Marketing|Business Administration, Management
The seller behavior and selling strategies in electronic marketplaces have not been well studied in literature. In this dissertation, we classify sellers into new sellers and experienced sellers according to their feedback scores. We find that new sellers have higher listing error rates, lower auction success rates, and lower auction prices than experienced sellers. New sellers are also less likely to receive positive feedback and more likely to receive negative feedback rates, more easily become frustrated by auction failure, and get more fraud claims and suspensions than experienced sellers. We also find that there are structural differences in auction success and price determinants between new sellers and experienced sellers and between new items and used items. Based on these findings, we classify online auctions into four segments and find that sellers in these segments behave significantly differently. We also find that signaling strategies are relatively more important for auction success than marketing strategies. The time series analysis shows that sellers learn only most recent completed auctions, and these completed auctions have more impact on experienced sellers than on new sellers.^ In addition, the existing studies on selling strategies neither provided operational selling recommendations nor integrated multiple objectives of sellers. In this dissertation, we incorporate seller and product heterogeneity into our analytical framework and implement data mining analysis to explore operational selling recommendations. We use two data mining techniques, classification and regression tree (CART) and association analysis, to identify the critical factors along with their sequences for auction success and prices. We find different determinants for auction success and prices among these four auction segments. The CART also provides operational selling strategies. We propose that, by using expected auction prices with the classification and regression trees, sellers can integrate auction success with prices in their selling strategies. The association rules derived from the real business dataset provide useful tools for sellers to avoid auction failure and boost online transactions. Overall, this dissertation contributes to the literature by providing knowledge about seller behavior and effective selling strategies, which are essential for building effective and efficient electronic marketplaces.^
Tu, Yanbin, "Analysis of selling strategies in electronic marketplaces" (2006). Doctoral Dissertations. AAI3231251.