Next Article in Journal
Designing a Deep Q-Learning Model with Edge-Level Training for Multi-Level Task Offloading in Edge Computing Networks
Next Article in Special Issue
Investigating Media-User Interaction for Public Play Space in a Smart City
Previous Article in Journal
Series-Parallel Generative Adversarial Network Architecture for Translating from Fundus Structure Image to Fluorescence Angiography
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 12
Issue 20
10.3390/app122010674
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Can Voice Reviews Enhance Trust in Voice Shopping? The Effects of Voice Reviews on Trust and Purchase Intention in Voice Shopping

by
Jaeun Seo
1,
Daeho Lee
1,2,3 and
Inyoung Park
4,*
1
Department of Interaction Science, Sungkyunkwan University, Seoul 03063, Korea
2
Department of Human-Artificial Intelligence Interaction, Sungkyunkwan University, Seoul 03063, Korea
3
Department of Applied Artificial Intelligence, Sungkyunkwan University, Seoul 03063, Korea
4
Research Institute for Information and Communication, Ajou University, Suwon 16499, Korea
*
Author to whom correspondence should be addressed.
Appl. Sci. 2022, 12(20), 10674; https://doi.org/10.3390/app122010674
Submission received: 27 September 2022 / Revised: 17 October 2022 / Accepted: 19 October 2022 / Published: 21 October 2022
(This article belongs to the Special Issue Current Trends in Human-Computer Interaction(HCI))
Review Reports Versions Notes

Abstract

:
Despite the high expectations of the voice shopping market, the impact of reviews and product types on voice commerce has yet to be explored. The purpose of this study is to investigate the effect of reviews and product types on users’ trust and purchase intentions in voice shopping. We explore users’ trust for voice shopping, trust in the vendor and purchase intention in three different types of reviews (i.e., no review, review by rating, and review by feature) and product types (i.e., search goods, experience goods, and convenience goods). We found that review conditions had a significant effect on purchase intentions and trust in voice shopping, whereas product types did not. Even within the review conditions, only the review by rating condition showed a significant difference from the no review condition. This study contributes to consumers and marketers by demonstrating the importance of providing rating reviews which requires a low cognitive load in the audio-centric environment.

1. Introduction

With the growing popularity of voice assistants (i.e., smart speaker, e.g., Amazon’s Alexa, Google’s Siri, Microsoft’s Cortana), the size of voice shopping market is also rapidly increasing [1]. Voice shopping, which allows users to shop by talking with a voice assistant, offers several advantages in terms of user experience during the shopping process. For example, according to survey form Bert and Ava [2], voice shopping is hands-free, it makes multi-tasking more convenient, and it enables providing quick answers and results [2,3]. Due to such benefits, voice shopping is expected to become a USD 40 billion market across the United States and the United Kingdom by 2022 [1]. However, despite the high expectations of voice shopping market, the number of voice shopping users is increasing slower than expected. For instance, in 2018, only 2% of Amazon Echo members used voice shopping to purchase items [4]. In addition, among the consumers aged 18–44 and over 45, only 1.32% and 0.16%, respectively, preferred voice shopping over other shopping methods [2].
Note that perceived trust in the purchase process has a significant impact on shopping [5,6,7,8], while it is assumed that still many customers do not easily feel trust in the purchase process when voice shopping. Indeed, according to a report by eMarketer [9], users of voice shopping cannot obtain visual information of various products, due to the lack of a screen in voice assistants, and it lacks cues to judge goods before purchase, making it difficult to build trust in vendors and products. To overcome the problems caused by consumer’s lack of trust in their purchases, present online shopping provides reviews, and previous studies revealed that customers rely heavily on others’ reviews before making purchasing decisions [10,11]. Having the ability to consider various information and reviews about certain products can lower the level of perceived risk and help users make purchase decisions [12], whereas being unable to consider such information can cause difficulties in making purchase decisions. However, since current voice user interfaces (VUIs) rarely deliver reviews, little is known about the effect of such reviews on voice shopping. Moreover, for voice shopping service companies, determining how to provide proper information in voice shopping remains a significant challenge.
The current study investigates the effects of reviews in voice shopping; whether providing reviews in voice shopping improves purchasing intention and trust. Considering that the types of online reviews are organized by ratings or by features [13,14], we examined the three types of reviews in voice shopping (i.e., review by rating, review by feature, and no review). Since the VUI and online sites have different modalities, it is an important problem to recognize the type of review that fits the VUI. Specifically, voice shopping as a lean media [15], could be negatively impacted by providing too much information to users because it requires a higher cognitive load [16,17]. Thus, we expect that the types of reviews (i.e., no review, review by rating, and review by feature) which is suitable for voice shopping may differ from those for online commerce.
In addition, we consider the product types in the study. The review types that users desire are mainly dependent on the types of products that they are shopping for [10,18]. For example, users aiming to purchase search goods (e.g., a cell phone), as opposed to experience goods (e.g., a hotel), tend to consider detailed reviews to be more trustworthy compared to other types of reviews [18]. Additionally, convenience goods, which have a short purchase cycle, benefit more from electronic word-of-mouth (eWOM) through visually intensive social networking services (SNSs) (e.g., Instagram). This is in contrast to durable goods, which have a longer lead time before purchase, and benefit more from text-heavy SNSs (e.g., Twitter) [19]. Due to the different nature of the information that people desire depending on the product type, it is important to examine the review and product types suitable for the voice shopping context.
The present study which compares three types of review conditions (i.e., no review, review by rating, and review by feature) and three product types (i.e., search goods, experience goods, and convenience goods) could determine which combination of reviews and product types is the most suitable for the voice shopping. This study has implications for exploring which review types and product types are apt to voice shopping environments.

2. Literature Review

2.1. Effect of Voice Review on Trust and Purchase Intention

Voice shopping is a service that enables online shopping with voice through intelligent voice agent, and is known as “voice commerce,” “voice shopping,” or “v-commerce” [20]. According to Capgemini [21], consumers prefer voice shopping compared to web shopping because of its convenience (52%), ability to multi-task and do things hands free (48%), and its help for automating routine shopping (41%). In voice shopping, voice assistant delivers information about a product that a consumer wants to explore or purchase, but it is not able to offer information about many products at once because voice interface cannot convey multiple information (e.g., reviews from many people about the product) in a short period of time. Unlike the existing online shopping environment where consumers can actively explore various products, voice shopping is likely to provide a passive product search process focusing on limited information on recommending products delivered by voice agents. Therefore, for voice shopping consumers, it is a very important to receive trustful review in a format that reduces the cost of searching for information as much as possible.
For a long time, trust has long been considered one of the top elements in online transactions [5,6,7,22], and lack of trust was the reason behind e-commerce’s struggle with low usage rate in the past [23]. Not only does online trust enhance online shopping intention, but a low level of trust is known to prevent users from purchasing online, especially among female [24,25]. In addition, Hassanein and Head [26] refer that E-commerce which fails to gain customer’s trust is “doomed”.
Various studies also find that reviews can greatly affect the level of trust among users. For example, review has a positive impact on the level of trust in e-commerce [27]. Electronic word of mouth (eWOM) and perceived web quality has effect on user’s trust [28], and while dependent on user’s gender, eWOM still had effect on level of trust, and should be designed more carefully [24]. Following the evidence, we hypothesize that providing reviews in voice shopping would lead to higher level of trust than the current voice shopping format which is not giving any review information (e.g., Amazon Alexa. Nugu from 11st Street).
Additionally, different types of review also affect users differently. For example, reviews that are centered around product’s benefits have a higher recall rate than reviews that are centered around product’s attributes [29], and review content that is focused on product’s quality-related statements increases review helpfulness [30]. Following this, types of review provided in voice shopping might also influence users in different ways. In this study review by rating (i.e., suggest product’s overall rating of reviews) and review by feature (i.e., suggest list of product’s features and its average score) were investigated because they are both a type of review popularly used in large e-commerce such as Amazon and Coupang. While review by rating and review by feature is both a summary of product reviews, rating review only offers numerical information and does not specify which aspects were positive or negative [31], and feature review gives information about several product features with a detailed summary [12]. Additionally, review by rating could have a greater an impact on the users than review by feature as it is known to have impact on initial trust perceptions [32]. Therefore, in this study, two different types of review (i.e., review by rating and review by feature) are each compared with no review condition. The hypotheses are as follows.
Hypotheses H1.
A voice agent which provides reviews by rating leads higher level of trust in voice shopping and trust in vendor than a voice agent with no review.
Hypotheses H2.
A voice agent which provides reviews by feature leads higher level of trust in voice shopping and trust in vendor than a voice agent with no review.
Reviews are also known to affect user’s online purchase intention. A positive review can reinforce customer’s emotional trust and intention to shop online [33,34]. Similarly, being able to consider various information such as review can lower the level of perceived risk, and help users make purchase decisions [12]. Online reviews can also indirectly affect consumer’s purchase intention, by impacting consumer’s perception towards brand equity [35] or perceived value and trust [36] and thus ultimately affecting consumer’s purchase intention. We hypothesized that review in voice shopping context will also positively impact user’s purchase intention.
Hypotheses H3.
Purchase intention will be significantly higher when using a voice agent which provides reviews by rating compared to using a voice agent with no review.
Hypotheses H4.
Purchase intention will be significantly higher when using a voice agent which provides reviews by feature compared to using a voice agent with no review.

2.2. Types of Review and Product Type

Different types of reviews are known to have different effect on users depending on the type of product users are purchasing. Types of review that users seek for can be different depending on the types of products they are shopping for [8,10,37]. For example, users looking for search products seek for attribute-based reviews, while users looking for experience products search for experience-based reviews [38]. Additionally, negative reviews have a more significant effect for experience goods than for search goods, due to the higher accessibility of search-attributes information compared to that of experience-attributes [39].
Many studies explain such phenomena with various matching theories. One of theories is a cognitive fit theory. Cognitive fit theory defines ‘cognitive fit’ as the match between problem representation and task, and that the match between the two variables can lead to efficient problem solving [40]. Several studies tested cognitive fit theory in e-commerce context. Particularly, video formatted reviews were more effective than text reviews for experience goods compared to search goods [37]. The content of the review (i.e., attribute-centric and benefit-centric) also had different effect on the users depending on their expertise–novice or expert [10].
Another example is media richness theory, which defines media richness as a fixed property of a medium that can support the communication occurring in media [41]. When the richness of the medium is not appropriate for the complexity of the task, it is more likely that messages will be misunderstood [42]. Simple tasks, when given too much information, will very likely lead to higher complexity of decision process [43]. These findings were also found to be true in e-commerce context- complex products such as computer and automobiles showed better fit with websites that offered richer information while simple products did not show such conclusive results [41].
Review-product congruity was proposed as well. Under review-product matching conditions consumers not only showed deeper-level comprehension, but also spent more time doing so [44]. Following prior studies, we divide reviews depending on their contents (feature and ratings) and also pose research questions on the interaction effect of review types and product types.
RQ 1:
Do certain matches between the review types and product types show higher purchase intention and trust?

3. Methodology

3.1. Experiment

For this study, we designed a 3 (no review, review by rating, and review by feature) × 3 (search goods, experience goods, and convenience goods) experiment in which the participants were asked to purchase items from a voice shopping simulating website. The purchasing of items was under the simulation and did not require actual transaction of money.

3.2. Participants and Procedure

A total of 118 people (79 females and 39 males, M = 24.881, SD = 0.707) participated in the voice shopping simulation and completed the related surveys via the Internet. Recruitment notices for the experiment were posted on the college’s official website as well as the college community’s website. The willing participants were requested to contact the researcher in charge of the experiment, then were given links for their assigned review conditions. The participants were also informed that the entire procedure would take approximately 15 min. In order to prevent any bias, the purpose of the study was presented as “Usability testing of voice shopping.” In addition, each participant has compensated KRW 3000.
The details regarding the given links for the assigned review conditions are as follows. On the first page of the website, the participants were informed of the browser settings for the experiment (e.g., using the Chrome browser and allowing the website access to the microphone). When the participants clicked the “Next” button, the consent page was presented in which they read the consent form prior to the experiment. If the participants clicked the “Agree” button, then they were allowed to move on to the two pre-experiment pages consisting of the payment method form for their participation in the experiment (i.e., page 1) and the demographic information survey, which included two questions about their previous online and voice shopping experiences (i.e., page 2). Next, they submitted these forms on the explanation page, which also included the purpose of the study (i.e., “Usability testing of voice shopping) and simple voice commands used during the experiment. Moreover, the participants were required to pass a short quiz, which consisted of two multiple-choice questions regarding voice commands. Finally, they were asked to purchase three items (one from search goods, experience goods, and convenience goods, respectively) and complete a survey on the website after each purchase. After submitting the final survey, a “Thank you” message was presented on the screen to clarify that the experiment had concluded.

3.3. Materials

3.3.1. Voice Shopping Simulation Website

To simulate the most likely voice shopping experience, the experiment was conducted online via web. The website was designed to have minimum visual cues, except for cues a screen-less smart speakers often use as well such as cues to signify when the speaker is “listening” or when the speaker is “talking”.
The voice shopping script used on the web was created after careful observation of Amazon’s Alexa and SKT’s Nugu. Participants could follow the script on web by interacting with voice inputs and outputs. “Web Speech API” was used so the website could analyze users’ voice inputs using speech to text (STT) functions. Additionally, pre-made MP3 files from “https://freetts.com” (accessed on 15 October 2022) was used for voice shopping outputs. The simulated voice shopping website presented 12 pages in total to the users. The users also participated in the voice shopping simulation by using simple voice commands (see Appendix A). For example, if the system recognized command words, such as “Next” or “Purchase,” then it proceeded to play pre-made MP3 files according to the script. Pre-made file provides high-quality TTS voice that are similar to voice of conversational agent. Moreover, the experiment screens were designed to have minimal visual information in order to imitate an actual voice shopping experience with a smart speaker.
The whole experiment including the voice shopping stimulation and the 3 surveys took place online. Upon contacting our research team to join the experiment, participants were given links according to the review conditions they were assigned. To prevent any sort of bias, the purpose of the study was introduced as ‘usability testing of voice shopping’. When participants first visited the link, they were notified that the whole experiment would take 15 min and were informed of the browser setting needed for the experiment ((1) Have to use Chrome browser, (2) Have to allow the website access to the microphone).
After the notification, participants were informed of some voice commands that can be used during the experiment and were asked to answer a short quiz to verify that they understood all the commands and how to use them.

3.3.2. Representative Items for Product Type

In order to ensure that participants conceived all the product types as the experiment intended, 41 respondents participated in a pretest to select the representative of each product types. The respondents were asked to choose the most fitting option for each item. In the experiment, the participants were asked to purchase total of three items by purchasing one item from each of the product types; computer (search goods), milk (convenience goods), and tour (experience goods), and participated in survey after each purchase. These items were picked according to a pre-test (i.e., an online survey) prior to the experiment, that administered to 41 people to ensure that each item was actually perceived as the product type it represented.
The pre-test survey was adapted from previous studies [7,26,45]. The pre-test included a list of 18 items, from which the participants were asked to choose one of the following options for each item which each represented key attributes of one of the three product types: Option 1—search goods) It can be evaluated before purchase; Option 2—experience goods) It can only be evaluated after purchase; or Option 3—convenience goods) It can be purchased without much thought. Out of the 18 items, one item was chosen for each of the three product types, according to the option the items were voted highest for. In this case, the chosen items were computer (search goods: 73.2%), milk (convenience goods: 56.1%), and tour (experience goods: 73.2%). The specific contents of each items such as price and reviews were selected from e-commerce companies in South Korea such as 11 STREET and Coupang.

4. Measure

4.1. Post-Purchase Surveys

Each post-purchase survey included two manipulation checks that asked the participants about the following aspects: (1) The review condition that they were given; and (2) Which item they purchased. The survey results that failed either one of the manipulation checks were removed (42 out of the 348 results). The remainder of each survey consisted of questions based on a seven-point scale, ranging from strongly disagree (1) to strongly agree (7).

4.2. Dependent Variables

Table 1 shows each survey question and the related dependent variables. Users’ purchase intention, trust in voice shopping, and trust in the vendor are measured in the present study. Cronbach’s alpha coefficients and descriptive data of the surveys are also presented in Table 1. Based on the findings, all of the dependent variables Cronbach’s α scored above 0.80.

5. Results

3 × 3 MANOVA was conducted to analyze the effects of review type and product type on the dependent variables. The results from the MANOVA and Tukey (HSD) post hoc analysis indicated that while rating condition, compared to no review condition, had significantly higher purchase intention (F = 4.216, p = 0.016).
For this experiment, Jamovi (Version 1.0.8.0) and IBM SPSS (23.0) statistical software were used to analyze the data. We tested the effect of the review conditions on the purchase intention, trust in voice shopping, trust in vendor (H1-H4) and the effect of the product types on such conditions (RQ 1). According to the results, the effect of the review conditions (Wilks’ λ = 0.947, F(6, 590) = 2.73, p < 0.05) was significant, but not the effect of the product types conditions (Wilks’ λ = 0.977, F(6, 590) = 1.14, p > 0.1). Thus, product type did not have significant effect on purchase intention, trust in voice shopping, and trust in vendor and therefore RQ1 was not supported. The review conditions had a significant impact on users’ purchase intentions (F(2, 297) = 4.22, p < 0.05) and trust in voice shopping (F(2, 297) = 2.65, p < 0.1), but not on trust in the vendor (F(2, 297) = 2.28, p > 0.1).
Tukey’s post hoc test was also performed to determine the differences between the review conditions (see Table 2). There was no significant difference between the review by feature condition and the no review condition; both no review condition and review by feature condition did not have significant effect on purchase intention, trust in voice shopping, and trust in vendor. Hence, H2 and H4 are not supported. Meanwhile, the review by rating condition showed significantly better performance compared the no review condition for trust in voice shopping and purchase intention but not for trust in vendor. Therefore, partial of H1 and H3 were supported.

6. Discussion and Conclusions

To determine the effect of review and product types on users’ trust and purchase intention, this study analyzed the participants’ survey results after purchasing three items from the voice shopping simulation website. Overall, the review by rating condition was found to have a positively significant effect on purchase intention and trust in voice shopping. The details of this finding are presented in the following sub-sections.

6.1. Review Conditions

The review by rating condition significantly improved the users’ purchase intentions and trust in voice shopping, which is in line with previous e-commerce studies. Reviews had a significantly positive impact on the level of trust [24,27] and purchase intentions [12], which can explain the significant positive effect of the review by rating condition in the present study. However, no significant differences were found between no review and review by feature condition, which does not align with previous studies.
The insignificance between review by feature condition and no review condition could be due to the characteristics of VUI. According to related research, since VUIs require higher cognitive loads than GUIs [17]. Not only that, but previous study shows that long, complex audios had negative effect on participant’s working memory [50]. In the present study, the voice files of the review by feature condition were generally longer than those of the review by rating condition, which could be the reason why the former condition failed to improve the users’ voice shopping experiences even when it provided review information.
As for the trust variables in the voice shopping context, this study divided trust into two variables (trust in vendor and trust in voice shopping). Although there was no actual vendor in the experiment, there was a possibility that voice shopping itself could be perceived as a vendor as well as an e-commerce platform. However, the result shows that the review conditions did not have a significant impact on trust in the vendor, but the review by rating condition did have a positive effect on trust in voice shopping. Thus, the results indicated that the users only perceived voice shopping as an e-commerce platform, and not as a vendor.

6.2. Product Types

The results of this study also showed that product types did not have a significant effect. This is an important finding, since previous e-commerce studies show that product types have a significant effect on users’ behaviors [37,51]. A possible explanation could be that convenience goods can be perceived as both search and experience goods, and this hindered with participants’ point of view. However, even when data related to convenience goods were deleted, no significant results were found. A pairwise t-test was conducted to determine if there were any significant differences between search and experience goods, but no differences were found.
This could be explained by Nelson’s study in 1981 where he explained that search success rate is a key variable that helps users distinguish between search and experience goods [52]. Voice shopping is an audio-centric media and have limited amount of information it can provide to the users because audios are more transient compared to text and has to be segmented into small sections [53]. Not only that, but voice shopping is considered a lean media [15] and can show negative effect due to high cognitive load when too much information is given [16,17]. Due to such limitations, voice shopping could have failed to provide enough information to the users for there to be a clear distinction between search and experience goods.

6.3. Implications

The academic implications of this study are as follows. First, this study is centered around the type of information provided VUI context. Research in the field of voice assistants has centered on the personalization of smart assistants for communication with users. For example, Moon et al. [54] found that extroverted users tend to feel higher social attraction when all of the smart devices in the home use the same voice, which is in contrast to introverted users. Even in driving context, users’ unique personalities require different types of assistant characteristics, and when they are incorrectly matched, the acceptance rate is lower than that of the default setting [55]. Overall, the aforementioned studies generally focused on how people can feel more natural by personalizing their interactions with voice assistants [54,55,56]. While such studies voice assistant is important, one should also keep in mind that content in which the voice assistant provides is also important. This study shows that content of the information such as ratings or feature reviews can have significant effect on the users and highlights the needs for such studies centering around contents of VUI in the future.
Secondly, we found that prior studies in e-commerce area can be applied to voice shopping, when medium characteristics are considered. Although it is known that people who shop online strongly rely on other people’s reviews before making purchase decisions [10,11], current voice shopping platforms such as Amazon Alexa, Google Assistant, and 11STREET does not offer reviews to its users. Our study shows that presenting ratings review information significantly improved user’s trust level and PI, which is in line with prior e-commerce studies. However, review by feature condition did not have a significant effect on the users’ level of trust or PI. Voice shopping, due to its audio-centric characteristics, can have negative impact to the users when it provides long audio information [53]. Other findings in e-commerce area should also be tested in the future, with VUI characteristics in mind.
The practical implications of this study are as follows. First, this study sheds light on the types of reviews that should be provided in voice shopping in order to raise users’ purchase intentions, level of trust, and perceived usefulness. In this regard, reviews in voice shopping should be short and direct (like the review by rating condition), instead of presenting over-abundant amount information. Although further research is necessary regarding the approximate lengths of effective reviews, this finding can be used to create guidelines for reviews in the voice shopping context.
Second, this study found that the trust in the vendor variable was not affected, as opposed to the trust in voice shopping variable, which can be insightful for vendors aiming to join voice shopping platforms. However, no vendor cues were provided in the experiment, which could be the reason why the trust in the vendor variable did not show significant results in all of the review and product conditions. The implication of this finding is that companies that sell smart speakers and own e-commerce platforms, such as Amazon or 11STREET, could benefit from a higher level of trust in the voice shopping market.

7. Limitations and Future Research

There are several limitations in this study that are worth noting. First, this experiment was conducted online in order to reach as many participants as possible, and to overcome additional external obstacles. However, the fact that it was an online experiment could have hindered the participants’ concentration levels and impacted the results, even though careful measures (e.g., manipulation checks and quizzes) were included. It is hoped that the strict screening of the data that failed either of the two manipulation checks supplemented the findings.
Second, this experiment required a computer, which is not the most popular medium for voice shopping. In addition, the fact that a computer was used to set up the smart speaker for the experiment could have created bias among the participants. Thus, visiting a link seemed less likely to affect the participants, compared to other options. Moreover, due to some technical difficulties, the sequence of the items that the participants were asked to purchase was fixed (i.e., (1) computer; (2) milk; and (3) tour), which could have impacted the results.
Finally, this study revealed how providing voice reviews in voice shopping can increase trust in voice shopping but did not measure whether the newly suggested VUI increases user satisfaction and reduces frustration with the voice shopping experience. We thought measuring user satisfaction is apt to subjective feelings, so we measure the user experience through whether trust in shopping was improved instead of checking user satisfaction. However, not measuring user satisfaction for voice interfaces may be a limitation of this study, given that measuring user satisfaction for new VUI can help improve VUI development.
Although this study mimicked the latest voice shopping experience, other methods and VUI characteristics should be considered when designing a voice shopping platform. Because VUIs are not suitable for providing users with abundant information, the current flow might not be the best strategy for voice shopping. For the future study, finding new ways to promote voice shopping that best fits its medium could be beneficial. For example, voice shopping could make recommendations to users, instead of simply assisting them in a search for certain items. More specifically, if data from SNSs is used, then the voice shopping platform could remind users that an anniversary or birthday is approaching, after which it makes recommendations about presents that are currently popular, gifts purchased the most on certain holidays, and sales promotions for items previously viewed by the users. In this regard, finding new ways to influence people to purchase items through voice shopping could be beneficial, both theoretically and practically, since it can contribute to the study of human-voice assistant interactions and develop the voice shopping market.

Author Contributions

Conceptualization & Validation, J.S., D.L. and I.P.; Writing—original draft, J.S.; Writing—review & editing, J.S. and I.P. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by the MSIT (Ministry of Science and ICT), Korea, under the ITRC (Information Technology Research Center) support program (IITP-2021-0-02051) supervised by the IITP (Institute for Information & Communications Technology Planning & Evaluation).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Conflicts of Interest

The authors declare no conflict of interest.

Appendix A

Voice shopping script for milk (English translation of the script)
  • Review by Rating Condition
  • User: “Milk.”
  • Voice shopping: You want to buy milk! There are three recommendations. The first item is a 24-pack of 125 mL, healthy sterile milk, certified as nonantibiotic. The price is KRW 13,790. A total of 3085 people gave 4.5 out of 5 stars. If you want to hear the next product, please say “Next.” If you wish to purchase the product, then please say “Purchase.”
  • User: “Next.”
  • Voice shopping: The second item is a 24-pack of 190 mL, original sterilized milk. The price is KRW 14,600. A total of 5100 people gave 4.5 out of 5 stars. If you want to hear the next product, then please say “Next.” If you wish to purchase the product, then please say “Purchase.”
  • User: “Purchase.”
  • Voice shopping: Your order has been placed. Voice shopping will end now.
2.
Review by feature condition
  • User: “Milk”
  • Voice shopping: You want to buy milk! There are three recommendations. The first item is 125 mL sized 24 pack set of healthy sterile milk certified as non-antibiotic. The price is KRW 13,790. A total of 67% of the people said it was better than expected and 64% said it was fresher than expected. If you want to see the next product, please say next; if you wish to buy the product please say purchase.
  • User: “Next”
  • Voice shopping: Second item is 24 pack set of 190 mL, original sterilized milk. The price is KRW 14,600. In total, 78% of the people said it was savory. If you want to see the next product, please say next; if you wish to buy the product please say purchase.
  • User: “Purchase”.
  • Voice shopping: The order has been placed. Voice shopping will end now.
3.
No Review condition
  • User: “Milk”
  • Voice shopping: You want to buy milk! There are three recommendations. The first item is 125 mL sized 24 pack set of healthy sterile milk certified as non-antibiotic. The price is KRW 13,790. If you want to see the next product, please say next; if you wish to buy the product please say purchase.
  • User: “Next”
  • Voice shopping: Second item is 24 pack set of 190 mL, original sterilized milk. The price is KRW 14,600. If you want to see the next product, please say next; if you wish to buy the product please say purchase.
  • User: “Purchase”
  • Voice shopping: The order has been placed. Voice shopping will end now.

References

  1. OC&C Strategy Consultants. The Talking Shop the Rise of Voice Commerce. 2018. Available online: https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwi51dK0_uLpAhXQQN4KHbIZB4IQFjAAegQIBxAB&url=https%3A%2F%2Fwww.occstrategy.com%2Fmedia%2F1285%2Fthe-talking-shop_uk.pdf&usg=AOvVaw1W2vvcha7xOp84p3foMCjN (accessed on 26 September 2022).
  2. Bert and Ava. Voice Shopping Consumer Adoption Report June 2018. Available online: https://voicebot.ai/voice-shopping-report- (accessed on 15 October 2022).
  3. Rhee, C.; Choi, J. Effects of personalization and social role in voice shopping: An experimental study on product recommendation by a conversational voice agent. Comput. Hum. Behav. 2020, 109, 10635. [Google Scholar] [CrossRef]
  4. Anand, P. The Reality Behind Voice Shopping Hype. Available online: https://www.theinformation.com/articles/the-reality-behind-voice-shopping-hype?shared=85666fc9c540c93e (accessed on 26 September 2022).
  5. Pavlou, P.A.; Gefen, D. Building effective online marketplaces with institution-based trust. Inf. Syst. Res. 2004, 15, 37–59. [Google Scholar] [CrossRef] [Green Version]
  6. Zahedi, F.M.; Song, J. Dynamics of trust revision: Using health infomediaries. J. Manag. Inf. Syst. 2008, 24, 225–248. [Google Scholar] [CrossRef]
  7. Thakor, M.V.; Kumar, A. What is a professional service? A conceptual review and bi-national investigation. J. Serv. Mark. 2000, 14, 63–82. [Google Scholar] [CrossRef]
  8. Kim, M.J.; Chung, N.; Lee, C.K. The effect of perceived trust on electronic commerce: Shopping online for tourism products and services in South Korea. Tourism Manag. 2011, 32, 256–265. [Google Scholar] [CrossRef]
  9. eMarketer Editors. Purchases Via Smart Speakers Are Not Taking Off. eMarketer Lowers U.S. Forecast Amid Security Concerns and Competition from Smartphones. 2020. Available online: https://www.emarketer.com/content/purchases-via-smart-speakers-are-not-taking-off (accessed on 25 September 2020).
  10. Park, D.H.; Kim, S. The effects of consumer knowledge on message processing of electronic word-of-mouth via online consumer reviews. Electron. Commer. Res. Appl. 2008, 7, 399–410. [Google Scholar] [CrossRef] [Green Version]
  11. Schlosser, A.E. Can including pros and cons increase the helpfulness and persuasiveness of online reviews? The interactive effects of ratings and arguments. J. Consum. Psychol. 2011, 21, 226–239. [Google Scholar] [CrossRef]
  12. Yang, J.; Sarathy, R.; Lee, J. The effect of product review balance and volume on online Shoppers’ risk perception and purchase intention. Decis. Support Syst. 2016, 89, 66–76. [Google Scholar] [CrossRef]
  13. Liu, C.L.; Hsaio, W.H.; Lee, C.H.; Lu, G.C.; Jou, E. Movie rating and review summarization in mobile environment. IEEE Trans. Syst. Man Cybern. Part C Appl. Rev. 2011, 42, 397–407. [Google Scholar] [CrossRef]
  14. Raja, D.K.; Pushpa, S. Feature level review table generation for E-Commerce websites to produce qualitative rating of the products. Future Comput. Inform. J. 2017, 2, 118–124. [Google Scholar] [CrossRef]
  15. Kraus, D.; Reibenspiess, V.; Eckhardt, A. How voice can change customer satisfaction: A comparative analysis between e-commerce and voice commerce. In Wirtschaftsinformatik; University of Siegen: Siegen, Germany, 2019; pp. 1868–1879. [Google Scholar]
  16. Maity, M.; Dass, M. Consumer decision-making across modern and traditional channels: E-commerce, m-commerce, in-store. Decis. Support Syst. 2014, 61, 34–46. [Google Scholar] [CrossRef]
  17. Cohen, M.H. Surfing the Voice Web: Issues in the Design of a Voice Browser, in ASR2000-Automatic Speech Recognition: Challenges for the New Millenium ISCA Tutorial and Research Workshop (ITRW). 2000. Available online: https://www.isca-speech.org/archive_open/asr2000/asr0_236.html (accessed on 26 September 2022).
  18. Jiménez, F.R.; Mendoza, N.A. Too popular to ignore: The influence of online reviews on purchase intentions of search and experience products. J. Interact. Mark. 2013, 27, 226–235. [Google Scholar] [CrossRef]
  19. Choi, Y.K.; Seo, Y.; Yoon, S. E-WOM messaging on social media. Internet Res. 2017, 27, 495–505. [Google Scholar] [CrossRef]
  20. Whang, C. Voice Shopping: The Effect of the Consumer-Voice Assistant Parasocial Relationship on the Consumer’s Perception and Decision Making. 2018. Available online: https://conservancy.umn.edu/handle/11299/201041 (accessed on 15 October 2022).
  21. Capgemini. Conversational Commerce: Why Comsumers Are Embracing Vioce Assistant in Their Lives. 2018. Available online: https://www.capgemini.com/wp-content/uploads/2018/01/dti-conversational-commerce.pdf (accessed on 26 September 2022).
  22. Bawack, R.E.; Wamba, S.F.; Carillo, K.D.A. Exploring the role of personality, trust, and privacy in customer experience performance during voice shopping: Evidence from SEM and fuzzy set qualitative comparative analysis. Int. J. Inf. Manag. 2021, 58, 102309. [Google Scholar] [CrossRef]
  23. Grabner-Kräuter, S.; Kaluscha, E.A. Empirical research in online trust: A review and critical assessment. Int. J. Hum. Comput. Stud. 2003, 58, 783–812. [Google Scholar] [CrossRef]
  24. Awad, N.F.; Ragowsky, A. Establishing trust in electronic commerce through online word of mouth: An examination across genders. J. Manag. Inf. Syst. 2008, 24, 101–121. [Google Scholar] [CrossRef]
  25. Kim, D.; Benbasat, I. Trust-related arguments in internet stores: A framework for evaluation. J. Electron. Commer. Res. 2003, 4, 49–64. [Google Scholar]
  26. Hassanein, K.; Head, M.M. The Influence of Product Type on Online Trust; MeRC Working Paper No. 6; McMaster eBusiness Research Centre (MeRC), McMaster University: Hamilton, ON, Canada, 2004. [Google Scholar]
  27. Utz, S.; Kerkhof, P.; Van Den Bos, J. Consumers rule: How consumer reviews influence perceived trustworthiness of online stores. Electron. Commer. Res. Appl. 2012, 11, 49–58. [Google Scholar] [CrossRef]
  28. Al-Debei, M.M.; Akroush, M.N.; Ashouri, M.I. Consumer attitudes towards online shopping. Internet Res. 2015, 25, 707–733. [Google Scholar] [CrossRef]
  29. Wang, S.; Cunningham, N.R.; Eastin, M.S. The impact of eWOM message characteristics on the perceived effectiveness of online consumer reviews. J. Interact. Advert. 2015, 15, 151–159. [Google Scholar] [CrossRef]
  30. Siering, M.; Muntermann, J.; Rajagopalan, B. Explaining and predicting online review helpfulness: The role of content and reviewer-related signals. Decis. Support Syst. 2018, 108, 1–12. [Google Scholar] [CrossRef]
  31. Abulaish, M.; Doja, M.N.; Ahmad, T. Feature and opinion mining for customer review summarization. In Pattern Recognition and Machine Intelligence, Proceedings of the Third International Conference, PReMI 2009, New Delhi, India, 16–20 December 2009; Springer: Berlin/Heidelberg, Germany, 2009; pp. 219–224. [Google Scholar]
  32. Sebastianelli, R.; Tamimi, N. E-tailer website attributes and trust: Understanding the role of online reviews. Online Inf. Rev. 2018, 42, 506–519. [Google Scholar] [CrossRef]
  33. Cheung, C.M.; Lee, M.K.; Thadani, D.R. The impact of positive electronic word-of-mouth on consumer online purchasing decision. In Visioning and Enginerring the Knowledge Society. A Web Science Perspective; Lytras, M.D., Damiani, D., Carroll, J.M., Tennyson, R.D., Avison, D., Naeve, A., Dale, A., Lefrere, P., Tan, F., Sipior, J., et al., Eds.; Springer: Berlin/Heidelberg, Germany, 2009; Volume 5736, pp. 501–510. [Google Scholar]
  34. Yeon, J.; Park, I.; Lee, D. What creates trust and who gets loyalty in social commerce? J. Retail. Consum. Serv. 2019, 50, 138–144. [Google Scholar] [CrossRef]
  35. Hayes, R.A.; Carr, C.T. Does being social matter? Effects of enabled commenting on credibility and brand attitude in social media. J. Promot. Manag. 2015, 21, 371–390. [Google Scholar] [CrossRef]
  36. Ponte, E.B.; Carvajal-Trujillo, E.; Escobar-Rodríguez, T. Influence of trust and perceived value on the intention to purchase travel online: Integrating the effects of assurance on trust antecedents. Tour. Manag. 2015, 47, 286–302. [Google Scholar] [CrossRef]
  37. Xu, P.; Chen, L.; Santhanam, R. Will video be the next generation of e-commerce product reviews? Presentation format and the role of product type. Decis. Support Syst. 2015, 73, 85–96. [Google Scholar] [CrossRef]
  38. Luan, J.; Yao, Z.; Zhao, F.; Liu, H. Search product and experience product online reviews: An eye-tracking study on consumers’ review search behavior. Comput. Hum. Behav. 2016, 65, 420–430. [Google Scholar] [CrossRef]
  39. Park, C.; Lee, T.M. Information direction, website reputation and eWOM effect: A moderating role of product type. J. Bus. Res. 2009, 62, 61–67. [Google Scholar] [CrossRef]
  40. Vessey, I.; Galletta, D. Cognitive fit: An empirical study of information acquisition. Inf. Syst. Res. 1991, 2, 63–84. [Google Scholar] [CrossRef]
  41. Simon, S.J.; Peppas, S.C. An examination of media richness theory in product web site design: An empirical study. info 2004, 6, 270–281. [Google Scholar] [CrossRef]
  42. Trevino, L.K.; Daft, R.L.; Lengel, R.H. Understanding managers’ media choices: A symbolic interactionist. In Organizations and Communication Technology; SAGE Publications: Thousand Oaks, CA, USA, 1990; Volume 71. [Google Scholar]
  43. Daft, R.L.; Lengel, R.H. Information Richness. A New Approach to Managerial Behavior and Organization Design; Texas A and M Univ College Station College of Business Administration: College Station, TX, USA, 1983. [Google Scholar]
  44. Huang, L.; Tan, C.H.; Ke, W.; Wei, K.K. Comprehension and assessment of product reviews: A review-product congruity proposition. J. Manag. Inf. Syst. 2013, 30, 311–343. [Google Scholar] [CrossRef]
  45. Nakayama, M.; Sutcliffe, N.; Wan, Y. Has the web transformed experience goods into search goods? Electron. Mark. 2010, 20, 251–262. [Google Scholar] [CrossRef]
  46. Everard, A.; Galletta, D.F. How presentation flaws affect perceived site quality, trust, and intention to purchase from an online store. J. Manag. Inf. Syst. 2005, 22, 56–95. [Google Scholar] [CrossRef] [Green Version]
  47. Zhou, T. An empirical examination of continuance intention of mobile payment services. Decis. Support Syst. 2013, 54, 1085–1091. [Google Scholar] [CrossRef]
  48. Hong, I.B.; Cho, H. The impact of consumer trust on attitudinal loyalty and purchase intentions in B2C e-marketplaces: Intermediary trust vs. seller trust. Int. J. Inf. Manag. 2011, 31, 469–479. [Google Scholar] [CrossRef]
  49. Hsu, M.H.; Chuang, L.W.; Hsu, C.S. Understanding online shopping intention: The roles of four types of trust and their antecedents. Internet Res. 2014, 24, 332–352. [Google Scholar] [CrossRef]
  50. Leahy, W.; Sweller, J. Cognitive load theory, modality of presentation and the transient information effect. Appl. Cogn. Psychol. 2011, 25, 943–951. [Google Scholar] [CrossRef]
  51. Huang, P.; Lurie, N.H.; Mitra, S. Searching for experience on the web: An empirical examination of consumer behavior for search and experience goods. J. Mark. 2009, 73, 55–69. [Google Scholar] [CrossRef] [Green Version]
  52. Nelson, P.J. Consumer information and advertising, Economics of information. Soc. Dimens. Econ. 1981, 3, 42–77. [Google Scholar]
  53. Singh, A.M.; Marcus, N.; Ayres, P. The transient information effect: Investigating the impact of segmentation on spoken and written text. Appl. Cogn. Psychol. 2012, 26, 848–853. [Google Scholar] [CrossRef]
  54. Moon, Y.; Kim, K.J.; Shin, D.H. Voices of the internet of things: An exploration of multiple voice effects in smart homes. In Distributed, Ambient and Pervasive Interactions, Proceedings of the 4th International Conference, DAPI 2016, Held as Part of HCI International 2016, Toronto, ON, Canada, 17–22 July 2016; Springer: Cham, Switzerland, 2016; Springer: Cham, Switzerland, 2016; pp. 270–278. [Google Scholar]
  55. Braun, M.; Schubert, J.; Pfleging, B.; Alt, F. Improving driver emotions with affective strategies. Multimodal Technol. Interact. 2019, 3, 21. [Google Scholar] [CrossRef] [Green Version]
  56. Klaus, P.; Zaichkowsky, J. AI voice bots: A services marketing research Agenda. J. Serv. Mark. 2020, 34, 389–398. [Google Scholar] [CrossRef]
Table
Table 1. The Dependent Variables.
Table 1. The Dependent Variables.
VariablesMeasurementM (SD)Cronbach’s αSources
Purchase intentionI would return to voice shopping.2.938
(1.686)		0.860[46,47]
I would consider purchasing from voice shopping in the next three months.2.660
(1.694)
I would consider purchasing from voice shopping in the next year.3.007
(1.844)
If I could, then I would discontinue my use of voice shopping.4.562
(1.907)
If this product is significantly less expensive from voice shopping than from a well-known online merchant, then I would consider purchasing it from the former.5.131
(1.776)
Trust in voice shoppingVoice shopping is reliable.4.127
(1.548)		0.956[36]
Voice shopping is trustworthy.4.095
(1.562)
Voice shopping includes integrity.4.350
(1.536)
Trust in vendorEven if it not monitored, I would trust the vendor to perform the job correctly.3.772
(1.479)		0.968[48,49]
The vendor of this product is reliable.3.810
(1.481)
I trust the vendor of this product.3.791
(1.456)
The vendor of this product is trustworthy.3.810
(1.438)
Table
Table 2. Tukey’s Post hoc Test Results for the Review Conditions by Rating and Feature.
Table 2. Tukey’s Post hoc Test Results for the Review Conditions by Rating and Feature.
Dependent Variable(I)
Review Condition		(J)
Review Condition		Mean Difference
(I–J)		FSig.
Purchase IntentionsRatingFeature0.425 *4.2160.088
No0.542 ** 0.017
Trust in Voice ShoppingRatingFeature0.222.6510.563
No0.47 * 0.058
Trust in the VendorRatingFeature0.41382.2820.095
No0.1671 0.665
Note: * = significant at the 0.1 level; ** = significant at the 0.05 level.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Seo, J.; Lee, D.; Park, I. Can Voice Reviews Enhance Trust in Voice Shopping? The Effects of Voice Reviews on Trust and Purchase Intention in Voice Shopping. Appl. Sci. 2022, 12, 10674. https://doi.org/10.3390/app122010674

AMA Style

Seo J, Lee D, Park I. Can Voice Reviews Enhance Trust in Voice Shopping? The Effects of Voice Reviews on Trust and Purchase Intention in Voice Shopping. Applied Sciences. 2022; 12(20):10674. https://doi.org/10.3390/app122010674

Chicago/Turabian Style

Seo, Jaeun, Daeho Lee, and Inyoung Park. 2022. "Can Voice Reviews Enhance Trust in Voice Shopping? The Effects of Voice Reviews on Trust and Purchase Intention in Voice Shopping" Applied Sciences 12, no. 20: 10674. https://doi.org/10.3390/app122010674

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop