패널 자료 분석을 활용한 유통업자 상표 이용도와 점포 충성도의 관계

The three most significant changes within the domestic retail business that have happened over the past 30 years are as follows. First, the increase in scale along with offering of multi-channel. Second, the power shift from that of the manufacturers to that of the retailers. Third, the development and strengthening of private label (PB) by retailers. Technology development, entry of various channels and changes in consumption patterns have led to infinite competition which has no channel boundaries. In such extreme competition, retailers are encouraged to invest heavily in development and expansion of PB as a strategic weapon to realize store differentiation and profitability at the same time. This is also an effective way to ultimately acquire customer loyalty. Although the introduction and development of PB in Korea are slower than in developed countries in Europe and North America, most channels including hypermarket, convenience store, TV home-shopping, supermarket, on-line shopping mall, and mobile commerce have actively introduced and developed PB in recent years, which requires research on PB from diverse perspectives. However, the proportion of research that confirms factors influencing PB performance is high but there is little research regarding the ultimate purpose of operating PB, despite that the retailers take the risks. In this study, we extended the scope of the research topic of PB and investigated the relationship between PB purchasing behavior of consumers and loyalty to the retailer. The moderating effect of the involvement of product category, the quality level of PB and customers’ tendency of multi-channel usage on the relationship between PB share and loyalty to the store have been identified as well. Unlike most preceding studies, we took a three-year FMCG purchasing data of 2,350 panelists among various channels, managed and accumulated by Nielsen Korea. Instead of conducting a survey which measures customers’ level of perception and purchase intention, we calculated store loyalty to a particular store as share of wallet (SOW) which shows the ratio of spending in a particular store to total spending of the household across the stores. We also conducted a questionnaire survey targeting the same panels and matched the questionnaire results with purchasing data to enhance the structure and completeness of the data set. The stores are set into 14 different categories, and the big 3 hypermarkets’ data are mainly used for analysis as they are the key players to enlarge PB market in Korea and they have more categories that offer PB. In addition, their PB shares are higher than average and they have on-line channels as well. The analysis unit of this study is the individual panels in a given year. Store loyalty to a particular store is defined as the share of wallet (SOW). In addition, the share of items purchased and the share of shopping trips have also been used as alternative indicators of store loyalty to observe the relationship between PB share and store loyalty from various perspectives. Since it was already proven from many previous researches that PB share and store loyalty have an interdependent relationship, we implemented the 2SLS (Two Stage Least Squares) Method to investigate the causal relationship between two variables (see [Equation 1 & 2]. As for competing hypotheses set to verify the direction between PB share and store loyalty, PB share has positive effect on store loyalty and store loyalty also has positive effect on PB share (see <Table 1>). We also investigated these relationships on the basis regarding the share of number of items purchased and the share of purchasing frequencies which are known as other measures for behavioral loyalty. In both cases, PB share has a positive effect on store loyalty(based on number β=0.112, t=3.588 / based on frequency β=0.112, t=3.539), as well as store loyalty positively effecting PB share(based on number β=0.107, t=4.849 / based on frequency β=0.108, t=5.010). Each store level produced the same results. We postulate that category involvement has a moderating effect on the relationship between PB share and store loyalty. However, no moderating effect between the two on either direction has been detected. The assumption is based on the conjecture that customers still lack knowledge or recognition about PB, despite it being said that the awareness of PB has been improved. In addition, it may be due to the context of this study, which is FMCG categories that are often considered to be purchased habitually. Observation of the analysis results at store level, the hypermarket L, compared with hypermarket E and H, shows not only stronger relationship between PB share and store loyalty but significant moderating effect of involvement on the direction of PB share to store loyalty. No moderating effect of PB quality has been found in either direction and it has been the same at the store level. It implies that premium PBs have not been actively spread and are not considered as alternatives of high-quality NBs yet. The implication is interpreted as the main reason for customers to buy PBs is the lower price than alternative NBs, even though many retailers offer diverse PBs in price and quality perspectives. Finally, the results also show that there are no moderating effects on customers’ tendency of multi-channel usage on the relationship between the main two variables in either direction. However, at the store level, hypermarket H and L show negative moderation effect on the direction of store loyalty to PB share. If the customers’ expectations and needs for the great advantage of on-line channels, which are competitive price and convenient searching, are taken into considerations, customers crossing from on-line to off-line channels to shop are much accustomed to search and purchase products that meets their needs. With this reason, while store loyalty is positively affecting PB share, customers who purchase more at on-line shops than off-line stores are likely to find more competitively-priced products than PBs, which might be the reason why this moderating variable has a negative effect.