Try It On! Contingency Effects of Virtual Fitting Rooms

Abstract

A revolutionary application of the virtual reality technology in online retailing, virtual fitting room (VFR), has attracted attention of researchers and practitioners recently. However, it remains unclear whether and how VFR influences sales and post-sales outcomes based on the limited literature, and retailers hesitate in adopting the technology due to concerns about its profit prospects. In this research, we conduct two large-scale field experiments to test the causal effects of different VFR designs, and a lab experiment to unveil the underlying theoretical mechanisms. We find that, although VFR can have a sizeable positive effect on sales, it can be counterproductive when used improperly. Specifically, personalized VFR may not increase sales if used in combination with conventional product visualizations because self-discrepancy becomes salient under this condition. Moreover, VFR significantly influences post-sales outcomes, i.e., it enhances customer satisfaction and reduces product return rate. The findings provide distinct theoretical contributions and managerial implications.

Key words and phrases:

• virtual reality
• virtual fitting room
• online sales
• customer satisfaction
• product returns

Notes

1. In this study, we focus on the performances of online stores instead of offline stores. In traditional online stores without VFR, self-discrepancy cannot occur because, unlike in offline stores, customers cannot try the clothes on their own body [20]. Moreover, self-discrepancy cannot occur without simultaneous presence of CVD. Traditionally, CVD is prevalent in online stores (instead of offline stores) because it serves as an important vehicle for customers to visually evaluate each product without being able to examine it in person [30]. For these reasons, self-discrepancy is an important issue to online stores (rather than offline stores) when they provide VFR.

2. When expanding the observation window to a longer period, e.g., two months after treatment, we still observe consistent results regarding the effects of VFR.

3. The retailer employs a customer rating platform that was introduced by Alibaba and is prevalently adopted in China, where the seller gets an automatic “highly satisfied” rating if the customer does not provide a rating within 15 days of purchase. Hence, with the automatic ratings, the response rate for customer satisfaction is 100%. If we only count non-automatic customer-initiated ratings, the response rate is 43.94%. We observe consistent findings when estimating the treatment effect based on non-automatic ratings only.

4. The average period of product return (from the date of purchase to the date on which the retailer receives the returned product) was 9.3 days.

5. Although our analysis is on the product level, we also confirm that there are no systematic differences across customers in the pre-treatment versus post-treatment periods in Appendix A.

6. During the observation period (August 4^th - September 14^th, 2016), a one-time price promotion was offered on August 7^th-9^th for the Chinese Valentines’ Day, which was during the first-week of the pre-treatment period. For this price promotion, the retailer selected 100 products (10.93% of the treated group and 10.61% of the control group; t = 0.15, p = 0.93) and offered 40 off on orders of 229, 100 off on orders of 499, or 200 off on orders of 799 (in Chinese Yuan). Hence, we use a dummy variable (1 if product i was offered promotions in period t, 0 otherwise) to control for the effect of price promotion.

7. The retailer in Study 2 adopts a similar rating platform as that in Study 1, which automatically assigns a “highly satisfied” rating if the customer does not rate it within 15 days of purchase. Hence, the response rate with automatic ratings is 100%. When excluding non-automatic ratings from the sample, the response rate becomes 54.51% and the estimated treatment effect leads to consistent conclusions.

8. The average period of product return (from the date of purchase to the date on which the retailer receives the returned product) was 8.25 days.

9. The retailer offered a one-time store-wide summer promotion (i.e., 10% off on two items, 20% off on three items, or 30% off on four or more items) on August 26^th and 27^th 2017, which were during the last week (week 7) of the experiment period. Since this promotion applied equally to all products in week 7, we use a dummy indicator for week 7 to control for it in the model (labeled as “Promotion” in Table 5). We also control for the fixed effects of the other weeks in the model.

10. While the retailer in Study 1 mainly targets on younger women, the retailer in Study 2 serves a wider range of age groups. When re-estimating the model with the subsample of older customers (≥ 45yo) only, we find consistent results: the effect of standardized VFR is 0.597 (p < 0.001) and that of personalized VFR is 1.091 (p < 0.001).

11. For instance, traditional IIT or advertisements featuring highly attractive model images might lure more customers to buy the product but could reduce post-purchase satisfaction and increase product returns when customers realize that the product actually does not look so good on them after purchase.

12. In the high-end market, the inherent risk of purchase is greater because of the higher prices of the products. Hence, VFR might play a more important role as it helps reduce risk perception, and thus the sales-enhancing effect of VFR could be even more significant. In the low-end market, although the inherent risk of purchase may be lower, VFR can still significantly improve sales because it helps add excitement and fun to the shopping process of otherwise less-exciting products. Hence, in both high-end and low-end markets, the positive effect of VFR can still hold.

Additional information

Funding

Shuai Yang has received support from the Fundamental Research Funds for the Central Universities, the DHU Distinguished Young Professor Program (No. LZB2019001), and National Natural Science Foundation of China (Grant No. 71832001). Guiyang Xiong would like to acknowledge the support from the Roadmap Grant on Innovation from Whitman School of Management, Syracuse University.

Notes on contributors

Shuai Yang

Shuai Yang ([email protected]) is an Associate Professor and the Deputy Head of Business Administration Department at the Glorious Sun School of Business and Management, Donghua University, China. She received her Ph.D. from the University of Connecticut. Her research studies have been published in such journals as International Journal of Electronic Commerce, Journal of Electronic Commerce Research, and International Journal of Contemporary Hospitality Management, among others.

Guiyang Xiong

Guiyang Xiong ([email protected]; corresponding author) is an Assistant Professor at the Whitman School of Management, Syracuse University. He obtained his Ph.D. in Business from Emory University. He conducts empirical research of the application of information systems in digital marketing. His work has appeared in Journal of Marketing, Journal of Marketing Research, Marketing Science, Journal of the Academy of Marketing Science, and others.