Voice Assistant Realism: How Perceived Realism of a Device’s Voice Affects Consumer Satisfaction and Affinity
Open Access
- Author:
- Hirt, Olivia
- Area of Honors:
- Marketing
- Degree:
- Bachelor of Science
- Document Type:
- Thesis
- Thesis Supervisors:
- Brett Christenson, Thesis Supervisor
Jennifer Chang Coupland, Thesis Honors Advisor - Keywords:
- marketing
sensory marketing
sound
voice
voice assistant
robot
human
timbre
recommendation satisfaction
affinity - Abstract:
- Voice assistants are becoming increasingly prominent in our world today, providing firms with an opportunity to capitalize on optimizing voice assistant design in order to better satisfy consumers, thus promoting success of the product and the firm. This paper is intended to provide insight into how firms can adapt the voice of the assistant by utilizing the concept of anthropomorphism, or the attribution of human characteristics to non-human agents (Epley, Waytz, & Cacioppo, 2007), to create positive connect with consumers. In my research, I measure two dependent variables: 1) consumer satisfaction with the assistant’s recommendation and 2), consumer affinity for the device itself. I perform two studies, in which I manipulate the level of human likeness or “realism” of the voice to create a range from robotic to human, then measure participants’ responses. Study 1 looks at how the varying levels of voice realism impacts consumer recommendation satisfaction and affinity. Study 2 considers whether the context of the task the assistant is performing is an influencing factor, measuring whether matching the level of voice realism to the seriousness of the task context has an influence on consumer response. The results suggest that consumers prefer humanlike or “realistic” voice assistant voice, with respondents’ recommendation satisfaction and affinity (specifically in terms of their perception of friendliness of the device) generally increasing as realism levels increased. They also indicate that task context does not impact this effect, showing a preference for realism no matter the task the assistant was performing. These results inform my ultimate recommendation for firms to design their voice assistants’ voices to be as realistic as possible to maximize positive consumer response overall.