Augmented Reality (AR) is a mixed reality that “combines real and virtual objects in a real environment, registers real and virtual objects with each other, and runs interactively, in three dimensions, and in real time” (Krevelen & Poelman, 2010).
AR has become increasingly popular in people’s lives -- it is in games like Pokemon Go, social media like Tik Tok, and online shopping platforms like the IKEA app. AR surveys also have huge potential in market research, as it allows the researchers to track the participants’ in-the-moment experiences. Aiming to test and validate the AR survey tool that HCD developed, we designed a pilot study to investigate the impact of the shopping environment on consumers’ in-the-moment perceptions about the stores they were asked to visit and the products they were tasked to find. Using this study, we learned about the development and use of the AR survey tool and explored how it can be improved and used in future research.
The Study
The study was conducted in November 2021. Participants (n=19) were asked to go to both Target and Walmart to complete this study. In each store, they rated their feelings toward the store pre- and post-shopping and toward three specific products -- Bounty Paper Towels, Noosa Yogurt, and Burt’s Bees Lip Balm. Their feelings were measured using the Self-Assessment Manikin (SAM), which breaks down people’s emotions into three dimensions: valence (sad-happy), arousal (calm-excited), and dominance (controlled-in-control) (Bradley & Lang, 1994). SAM consists of three pictorial scales that correspond to the three emotional dimensions. Since the scales do not contain words, SAM can be used regardless of cultural differences and biases introduced from verbal descriptions.
AR was incorporated in the study as the participants needed to look at the stores and the products through their phone cameras for 5 seconds before they rated their feelings, and they were to continue to look through the camera when they answered the survey questions. In this way, the survey could record the participants’ most vivid shopping experiences at the moment.
Study Results
We only looked at directional results of this study, which can be indicative of larger effects given the small sample size. We first used two-sample t-tests to compare the means of the SAM scores of each store pre- and post-shopping and among the products between stores. Results indicated that in general, (1) the shoppers’ emotional states went downward after their shopping, as most pre-shopping SAM scores were higher than the post-shopping SAM scores, and that (2) people felt more positive about Target than Walmart, as most pre- and post-shopping SAM scores were higher in Target than Walmart.
Then, we ran multiple regression models to investigate what factors explained the emotions the participants had toward the products. We ran three models as the three emotional dimensions were independent of one another. Below is a summary of the significant variables in each regression (α=.05). While there were many indicative results, the three most interesting ones were: (1) product type only matters to emotional valence, (2) the emotional dimensions can influence one another across time, and (3) the products’ perceived price only affect shoppers’ emotional arousal during shopping.
Reflect on the AR Tool
Through developing this study, we identified issues on the current AR tool and discussed the possible solutions to the issues as well as future research directions to improve the tool. Overall, we identified two main areas of focus: designing surveys on smartphones and adding AR surveys to the current consumer research.
1. Designing Surveys on Smartphones
Since AR is often incorporated on phones, it is important to create a smooth research experience on smartphones so that the participants’ in-the-moment experiences will not be interrupted.
Phone surveys cannot contain as much information as traditional surveys on computers because of the limited width of the phone screen. In this study, we changed the original 9-point SAM into 5-point scales and broke paragraphs into shorter sentences to accommodate. Furthermore, a horizontal scale can be hard to read on the phone because the font has to be small to fit the whole scale onto the screen. It could be better to put options vertically on the screen. We could also rotate the phone to landscape orientation to do surveys so that there is more information per page. But we need to consider people’s phone use habits and the convenience of holding a phone in landscape orientation because it often requires people to hold the phone with both hands.
Furthermore, people have to use the camera function if they use AR on their phones. However, using the phone camera in some situations will make the participants feel uneasy during their survey process. For instance, some participants in this study commented that it felt embarrassing to hold the camera high up in the air at a store. Or it felt uncomfortable to hold the camera in front of a fridge because their hands became very cold. The solution to this problem would be creating a more fluent AR experience by reducing the camera use time. To know where and how much we should reduce it, we need to understand if using AR will interrupt people’s in-the-moment experience, and by how much participants will mix their survey experience with their shopping experience.
2. Adding AR Surveys to the Current Consumer Research
Simplicity matters when we want to capture in-the-moment experiences. If the AR tool is too complicated to use, it will interrupt the shoppers' natural shopping experience. While all surveys need to be concise, AR surveys particularly need simplicity in wording and survey design since people spend limited time per page on their phones.
Furthermore, combining AR surveys with more traditional survey tools might generate less confounding results. In this study, many participants commented that they forgot about the meanings of the SAM images during the survey because they were used to interpreting text instead of pictorial instructions. Given that AR is already a novel survey tool, adding SAM scales to the same survey could give the participants too much information to process. It might be better to combine a novel survey tool with more traditional tools to reduce the participants’ burden.
Future research can investigate more innovative solutions to simplify the survey process. One direction could be making quick and easy responses to each survey question, like swiping left or right on Tinder. Another direction could be conjoint analysis, where people can choose from multiple options simultaneously instead of one option at a time.
Concluding Thoughts
While AR has huge potential in market research, its development process can be long and effortful. Thus, one important question to ask now is whether it is beneficial to incorporate AR into our current market research toolbox, considering that an immature AR survey may do more harm than good on collecting accurate information from the participants. More importantly, the advantage of using AR, i.e. capturing people’s in-context experiences, may not outweigh the disadvantage of introducing a new survey tool to the participants, as it takes time and effort for people to adopt new technologies.
Another aspect to consider is to find out the niche research questions that only AR surveys can answer. It is not worth it to switch to a new survey tool if the old ones can do the job. Given that the development of the AR survey tool has just started, it can be hard to think ahead and figure out what the mature AR survey tool can do.
References
Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: the self-assessment manikin and the semantic differential. Journal of behavior therapy and experimental psychiatry, 25(1), 49-59.
Van Krevelen, D. W. F., & Poelman, R. (2010). A survey of augmented reality technologies, applications and limitations. International journal of virtual reality, 9(2), 1-20.
