Dissertation
Understand consumer's product decisions when shopping by voice
Doctor of Philosophy (Ph.D.), Drexel University
May 2019
DOI:
https://doi.org/10.17918/kpjb-1b41
Abstract
Using deep learning and natural language processing technologies, many e-merchants nowadays start offering artificially intelligent virtual assistant on their website or through mobile applications to assist consumer's online shopping (e.g., Amazon Alexa, Domino's DRU assist). Compared with traditional online shopping that is navigated by point and click, shopping with virtual assistant resembles the dialogical communication process between humans, allowing consumers to place an order by talking or texting. My dissertation research explores how different types of interaction modes will influence consumer's product decisions while shopping online. I propose that engaging in voice-based interactions will promote consumers choosing more recommended products. This is because oral communication carries more social functions than text-based communication and click-based interaction, and this social inclination will evoke people's anthropomorphism tendency. Accordingly, they tend to think their communication partner is a social actor rather than a computer. As a result, they will be more likely to comply with the request by choosing the recommended product. In addition, I argue that this effect only exists in two-way dialogical communication, where people can get social responses from the computer reciprocally and diminishes in one-way monological communication, where the social responses from the computer are missing. This is because social interaction is two-way in nature, while one-way communication will create social distance and undermine the social bonds between communicators. Thus, it suggests that people rely on the entire interaction process to infer the computer's social identity. An online food ordering website is developed and utilized to test the hypotheses. This website manipulates both user's communication modality (i.e., speak vs. click vs. type) and computer's response modality (i.e., synthesized voice vs. interactive text vs. non-social response) to allow participants to interact with using different modalities. Five studies have been conducted with responses from more than 3,200 MTurk workers. Results from those studies lend support to the hypotheses.
Metrics
56 File views/ downloads
41 Record Views
Details
- Title
- Understand consumer's product decisions when shopping by voice
- Creators
- Zhen Yang - DU
- Contributors
- Yanliu Huang (Advisor) - Drexel University (1970-)
- Awarding Institution
- Drexel University
- Degree Awarded
- Doctor of Philosophy (Ph.D.)
- Publisher
- Drexel University; Philadelphia, Pennsylvania
- Number of pages
- 102 pages
- Resource Type
- Dissertation
- Language
- English
- Academic Unit
- Bennett S. LeBow College of Business; Marketing; Drexel University
- Other Identifier
- 9483; 991014632694204721