How far can Wearable Augmented Reality Influence Customer

Shopping Behavior
Hamraz Javaheri Maryam Mirzaei Paul Lukowicz
[email protected] [email protected] DFKI
DFKI Kaiserslautern, Germany
Kaiserslautern, Germany [email protected]

ABSTRACT 1 INTRODUCTION
We investigate if providing shoppers with Augmented reality (AR) Augmented Reality (AR) is a technology that allows users to see
as a shopping tool can lead to an increase in purchase rate compared the combination of virtual objects and real-world in a way that
with conventional shopping applications. In a "simulated shopping" aligns the virtual objects with physical ones in real-time [7]. AR
study with two groups with a total number of 20 participants, a test provides a new opportunity to see a new reality with virtual objects
group used a wearable AR device (HoloLens) as a primary shopping superimposed, complementing real environments, hence providing
method while a control group used a tablet as a conventional 2D users with information that they cannot directly detect with their
shopping device. According to the surveys collected from partici- senses [2, 3, 8]. AR technology has been developed throughout
pants, 19 participants (95%) commented that the AR method was years and has allowed easy accessibility and proximity between
more joyful than conventional 2D method. Furthermore, all par- customers and products. This technological revolution became pos-
ticipants said that their experience with AR technology was more sible through some applications available on smart devices such as
realistic. 16 participants (80%) believed that the AR method was computers, mobile phones, tablets, and personal digital assistants
more influential than the conventional method, but the other four [4, 10, 11, 19]. Augmented Reality has a great potential to be used
participants (20%) said that neither of the methods had any impact in different areas such as education, medicine, architecture, interior
on their buying intentions. Although the mean number of prod- design, marketing, advertisement, and shopping [18].
ucts added to the basket by each person in HoloLens experience In past decades, the influences of technological advancements
and tablet experience was not significantly different (p=0.675), the on everyday routines could be easily observed. With smart devices
mean number of bought items was significantly higher in HoloLens entering people’s daily life, companies’ interest to invest in AR
experience compared to the tablet experience (p=0.004). In conclu- applications has been raised, resulting in an increase in the de-
sion, AR increased customer’s interest in shopping. While it had velopment of several applications and campaigns for e-commerce
no significant impact on adding products to the shopping cart, it purposes. Adidas and Lego are only two of many brands that use
affected the rate of purchase. AR as an advertisement tool to promote their products and brands.
This study is an initial experimental evaluation of the influence
CCS CONCEPTS that presenting a product through wearable AR has on the shop-
ping behaviour. Specifically we investigate whether using AR as a
• Human-centered computing → Empirical studies in HCI.
shopping tool can lead to an increase in purchase rate compared
with conventional shopping applications. We do so through a game
KEYWORDS where subject can use virtual money to do simulated purchases.
Augmented Reality, AR, Shopping, HoloLens While simulates shopping with virtual money is is not the same
as purchasing real products with real money, we argue that seeing
ACM Reference Format: a statistically significant change in behaviour is a valuable indica-
Hamraz Javaheri, Maryam Mirzaei, and Paul Lukowicz. 2020. How far can tion of what are likely to see in real world shopping scenarios. In
Wearable Augmented Reality Influence Customer Shopping Behavior. In Mo- particular it is significant step beyond mere surveys, which have
biQuitous 2020 - 17th EAI International Conference on Mobile and Ubiquitous dominated much previous work in this area.
Systems: Computing, Networking and Services (MobiQuitous ’20), Decem-
ber 7–9, 2020, Darmstadt, Germany. ACM, New York, NY, USA, 6 pages.
2 RELATED WORKS
https://doi.org/10.1145/3448891.3448930
Previously, almost all the studies related to AR relied on collecting
data from customers via surveys or doing interviews with experts.
In the study conducted by Silva et. al [15] in 2012, an AR application
Permission to make digital or hard copies of all or part of this work for personal or
classroom use is granted without fee provided that copies are not made or distributed was proposed that would have several features such as guiding,
for profit or commercial advantage and that copies bear this notice and the full citation detailed information, entertaining gaming possibilities, access to
on the first page. Copyrights for components of this work owned by others than ACM exclusive price-based promotions, and interactive AR displays to
must be honored. Abstracting with credit is permitted. To copy otherwise, or republish,
to post on servers or to redistribute to lists, requires prior specific permission and/or a understand if consumers would use an AR application on their
fee. Request permissions from [email protected]. smartphones while visiting a shopping mall. In 2013 using primary
MobiQuitous ’20, December 7–9, 2020, Darmstadt, Germany data collected through a web-based online survey, Correia et. al. [6]
© 2020 Association for Computing Machinery.
ACM ISBN 978-1-4503-8840-5/20/11. . . $15.00 tried to verify, if AR as a recently available technology for buyers
https://doi.org/10.1145/3448891.3448930 can be an appealing part in the purchasing process. They also tried

464
MobiQuitous ’20, December 7–9, 2020, Darmstadt, Germany H. Javaheri et al.

to investigate if AR can be used by the companies to reach the

customers effectively. Johnson et. al. [9] performed a study in 2014,
involving online questionnaires and semi-structured interviews
with experts in the field of AR technology to evaluate if AR could
be used as a sales/marketing strategy in fashion retailing industry.
In the same year, the first experimental study in a real retail envi-
ronment was done by Spreer et. al. [16] in collaboration with one
of the largest German bookstores and a leading AR development
company, to examine the issue of AR acceptance and multimedia
product presentation. In the following year, Stoyanova et al. [17]
compared three advertising and digital shopping platforms (abso-
lutely interactive, AR Marker-based and AR Markerless platforms)
to explore their effectiveness impressions and purchase intention of
customers. The independent and dependent variables were adver-
tising types and cognitive aspects, respectively. In an experimental
study in 2017 [12], it was shown that IKEA shopping-oriented AR
application was recognized as highly enjoyable, useful, and fur-
ther evoked higher purchase intentions than the website version.
In this study, the attitude toward the product was not found to
be the main influence, however, the engaging experience and the Figure 1: Application architecture
conveyed unique product knowledge were the prominent factors.
Furthermore, Cehovin et. al. [5], examined L’Oréal Genius Makeup
and the IKEA Catalog AR applications using observational experi-
ment followed an interview, to investigate how mobile augmented
reality impacts the consumers search and evaluation behavior in
the customer decision-making process. The most recent study was
done by Saifeddin Alimamy et. al. [1]. In this experiment, the in-
tervention group was randomly assigned to an experiment (using
META1 glasses) with a poster showing either a fake-branded smart-
phone (as a high involvement product) or a fake-branded burger
(as a low involvement product). An augmented video was used to
show the superimposition of virtual elements with the mentioned
posters. Results showed that co-creation through exposure to AR
reduces perceived financial risk for the high-involvement product (a) (b)
and reduces the perception of physical risk for the low involvement
product. The results also provided evidence that exposure to AR Figure 2: (a) first person view of AR application. (b) third
increases customer perceived trust and purchase intent. person view of a participant using AR application

3 MATERIAL AND METHOD

3.1 Study Design The participants were divided into tablet first group (would try
A simple shopping application was designed using Unity game first the tablet and then the HoloLens) and HoloLens first group
engine for both devices. The application interface was identical (who would try first the HoloLens and then tablet) so they could
for both versions with an exception that the AR version allows try both experiments to decrease the bias (Figure 3).
manipulating the product and visualizing it through the glass in the
real-world with actual physical dimensions. Thirty-three objects, 3.2 Participants
each with a value of less than 50 Euro and from different categories, In this study, the Convenience Sampling Method was used because it
were chosen to be presented as products for sale through this app. is considered the easiest, the cheapest and the least time-consuming
The application composed of 4 main parts that interact together to sampling method. The sample included subjects who were accessi-
simulate a shopping application similar to commercially available ble on the time interval of this study. After exclusion, 20 volunteers
ones (Figure 1). In the menu manager, it is possible to change the (10 males and 10 females) with a median age of 28.5 years (23–35
object, view a 3D model, open the shopping cart and add selected years) participated in this study which made the population limited
objects to it. Each product could be dragged, rotated, and added to generation Y. All participants had a higher university educa-
to the shopping cart. By entering the shopping cart, it is possible tion or were a student at the time of the survey. Written informed
to delete objects or to buy the selected ones. Finally, the selected consent was obtained from all participants prior to participating
objects would be saved as an order, and a reference number would in this study regarding collecting, analyzing and publishing the
appear which can be closed when it is not needed (Figure 2). anonymous data.

465
How far can Wearable Augmented Reality Influence Customer Shopping Behavior MobiQuitous ’20, December 7–9, 2020, Darmstadt, Germany

Figure 3: Study flow chart

3.3 Data Recording multiple-choice/checkbox questions and an open-ended question

The experiment took place in a room with basic furniture, including were included.
desk, chair, phone, and whiteboard, allowing participants to try and Itemized rating scales are used to measure consumer attitudes
place the offered objects freely. To collect primary data, multiple namely purchase intentions, attribute perceptions, and service
standardized English questionnaires were designed. Questions were satisfaction[13]. These ordinal scales are presumed to be convenient
mainly based on itemized rating scales (Likert rating scale, Semantic to investigate mentioned behaviours [14].
differential scale, and Comparative rating scale). In addition, a few

466
MobiQuitous ’20, December 7–9, 2020, Darmstadt, Germany H. Javaheri et al.

The questionnaires consisted of demographic questions (gender,

age, etc.,) followed by closed-ended questions dealing with the at-
titude towards the technology, questions about the experiments
in general, the designed shopping-simulated application, AR tech-
nology and used devices (tablet and HoloLens), questions about
comparison between two experiments and finally an open-ended
question regarding the participant personal opinion about useful
aspects of AR.

3.4 Statistical Analysis

The data collected through questionnaires were imported to an
Excel tabular format, and all the charts and graphs are created
through Microsoft office Excel. As well, SPSS version 23 (IBM Corp.
released 2015, IBM SPSS Statistics for Windows. Armonk, NY; IBM)
was used for statistical analyses. Continuous variables are shown as
mean ± standard deviation, and categorical variables are expressed
as numbers (percentages) unless otherwise indicated. Continuous
variables were compared between tablet first and HoloLens first
groups using the Mann-Whitney U-test, and Fisher’ exact test was
performed to compare categorical variables between the groups.
The number and the price of products which were added to cart
and bought by participants during two experiments were compared
using the paired sample t-Test. A two-sided p-value of less than
0.05 was considered statistically significant in all tests.
Figure 4: Comparison of the number and total price of items
4 RESULTS added to cart and bought in AR wearable device versus tablet
At the time of participation, 50% of the participants knew what AR experiment.
is and only 30% had experience using AR devices and applications.
All participants owned a smartphone and a Laptop, and 65% owned
also a tablet. All participants said that they would accept to install an significantly different (p=0.163). However, the participants on aver-
AR application on their smart device which shows the high rate of age paid significantly more to buy the items during the HoloLens
trust of participants to AR applications. Additionally, 90% said that experiment (110.9 € ± 102.4 €) compared to the tablet experiment
they would try such an application to use its advantages although (77.3 € ± 56.7 €, p=0.035) (Figure 4).
they do not like the appearance of the application. All participants According to surveys collected from participants, 19 participants
agreed that technology makes life easier, 70% agreed that they can (95%) commented that the AR method was more joyful than the
learn technology-related skills and 90% agreed that they follow conventional 2D method. Furthermore, all participants said that
technological updates. Most of the participants (n=15, 75%) had their experience with AR technology was more realistic than the
concerns regarding data security and safety using technology; only conventional method shopping experience. 16 participants (80%)
2 participants (10%) said that technology is safe regarding data believed that the AR method was more influential than the con-
security. ventional method, but the other four participants (20%) said that
To compare the participants’ shopping behavior the total number neither of the methods was influential (Figure 5).
and price of the products added to the cart and bought using two Eight participants (40%) said that the AR method was more un-
shopping experiments were compared. Of the total of 88 products derstandable, while 6 participants (30%) said that the conventional
added to the shopping cart, the participants bought 84 items (95.5%) method was more understandable and the other 6 participants (30%)
in the HoloLens experiment. However, during the tablet experience, believed that both methods are equally understandable and there
the participants bought only 56 (68.3%) of a total of 82 items added was no difference between them. The majority of the participants
to the shopping cart. This means the number of products added to (n=12, 60%) said that the conventional method was easier to inte-
the cart by each person in HoloLens experience and tablet experi- grate into. Four participants (20%) found AR technology an easier
ence was 4.4 ± 2.4 and 4.1 ± 3.2, respectively (p=0.675). Although method to integrate into it, and the other four participants (20%)
the mean number of added items was not significantly different believed that both methods were easy to integrate into.
between the two experiments, the mean number of bought items Eighteen participants (90%) found shopping experience with
was significantly higher in HoloLens experience (4.2 ± 2.3) com- HoloLens more joyful than the tablet and 16 participants (80%)
pared to the tablet experience (2.8 ± 1.8, p=0.004). The mean price stated that HoloLens was more useful for shopping in comparison
of the items added to the cart by each participant during HoloLens with the tablet. Majority of the participants (n=12, 60%) believed
(116.8 € ± 67.4 €) and tablet experiments (111.5 € ± 67.9 €) was not that tablet was a more comfortable device for shopping and 14

467
How far can Wearable Augmented Reality Influence Customer Shopping Behavior MobiQuitous ’20, December 7–9, 2020, Darmstadt, Germany

Figure 5: Comparison between two groups regarding influ- Figure 7: Surveys result showing participant’s preferred
ence rate of used method. shopping method in future.

While the average time spent on the HoloLens application was

13.54 minutes, the participants spent on average 5.4 minutes on
tablet application. It is observable that AR had an influence on the
average time spent on shopping. While it might have happened
due to user unfamiliarity with AR device, based on surveys, it is
convenient to say that it might be a result of increased interest and
attention of participants during the experiment with AR device.

5 DISCUSSION
This study tried to assess the effects of wearable AR on customers’
shopping behavior and attitude with experiments that allow par-
Figure 6: Comparison between devices used for experiment. ticipants to literally interact with the technology and device. The
experiments compared a custom-designed shopping application
used on HoloLens with the same application on a tablet. The assess-
participants (70%) claimed that tablet was easier to use for shopping ments were based on the results of the pre-designed questionnaires
purposes compared with HoloLens, which was an expected result that were filled by the participants during the experiment, and the
considering the fact that tablets are commonly a part of people’s data collected from the applications used in the study. Analyzing
daily life. On the other hand, 10 participants (50%) said that the the results, it was observed that although using a wearable AR
contents of the HoloLens were easier to handle in general, where shopping app did not have any major effect on the selection of
four participants (20%) thought that the contents of tablet are easier objects (which were added to the shopping cart), it significantly
to handle, and the rest mentioned that both devices are the same increased the final purchase amount of the participants compared
(Figure 6). to conventional shopping app on the tablet. Additionally, it was
Participants believed that the AR technology was most suitable observed that participants had more trust in what they saw us-
for decoration (n=20, 100%), then for furniture (n=18, 90%), home ing the AR shopping app since they could see and interact with a
appliances (n=60%), and electronics and computers (n=4, 20%), re- real-size 3D version of the objects. Another important finding was
spectively. None of the participants considered AR as a suitable the fact that participants believed AR technology is useful for buy-
technology for shopping cloths. Interestingly, 16 participants (80%) ing decoration, furniture, and home appliances while clothes are
stated that they would prefer to use AR technology for shopping in not a suitable product category for this technology. Although AR
the future. Other mentioned preferred methods for shopping were a shopping cannot replace shopping experience in a physical store,
physical store (n=9, 45%), online shopping via tablet or Smartphone joyfulness, realism, and usefulness in the wearable AR shopping
applications (n=8, 40%), and online shopping on a website (n=4, experiment made the participants interested in using this method
20%) (Figure 7 ). in the future. Furthermore, they could handle the content with the
The majority of the participants (n=16, 80%) agreed that AR HoloLens the way they liked which increased their buying inten-
assisted them in making a buying decision. 18 participants (90%) tion and shopping productivity. Therefore, wearable AR devices
acknowledged that AR influenced their understanding of the prod- seem to be a potentially useful shopping method for the future.
ucts. 19 participants (90%) also confirmed that AR can improve Based on the participants’ opinion, using the tablet was easier
shopping productivity. Additionally, 16 participants (80%) admitted and more comfortable which was expected due to the fact that wear-
that AR can be useful in buying what they want. The majority of able devices (especially the HoloLens which is not commercially
the participants (n=17, 85%) commented that AR increased their released yet) are not as common as tablets. All the participants
shopping intention. already have owned a smartphone and most of them have had a

468
MobiQuitous ’20, December 7–9, 2020, Darmstadt, Germany H. Javaheri et al.

tablet which made it easier for them to interact with the tablet Dissertation. University of Otago.
rather than the HoloLens. Hopefully by technology improvements [2] Ronald Azuma, Yohan Baillot, Reinhold Behringer, Steven Feiner, Simon Julier,
and Blair MacIntyre. 2001. Recent advances in augmented reality. IEEE computer
in the future wearable devices would be more popular and easier graphics and applications 21, 6 (2001), 34–47.
to use. [3] Ronald T Azuma. 1997. A survey of augmented reality. Presence: Teleoperators &
Virtual Environments 6, 4 (1997), 355–385.
There were some limitations to this study which should be con- [4] Mads Bodker, Greg Gimpel, and Jonas Hedman. 2009. Smart phones and their
sidered in the interpretation of the outcomes. The limitations of this substitutes: Task-medium fit and business models. In 2009 Eighth International
study include the number of products, sample size, and experiment Conference on Mobile Business. IEEE, 24–29.
[5] Federica Cehovin and Bernice Ruban. 2017. The Impact of Augmented Real-
environment. Despite trying to have a sufficient variety of products, ity Applications on Consumer Search and Evaluation Behavior. Kopenhagen:
a limited number of products and categories were presented in Copenhagen Business Scholl (2017).
the application in comparison to conventional online shops which [6] Luís Filipe de Oliveira Correia. 2013. Augmented reality and the interest for
consumers in their buying process. Ph.D. Dissertation.
might have an impact on participants’ buying decisions. Further- [7] Tobias Höllerer and Steve Feiner. 2004. Mobile augmented reality. Telegeoinfor-
more, all participants aged between 23 and 35 years old which is matics: Location-based computing and services 21 (2004).
[8] Dragos Daniel Iordache and Costin Pribeanu. 2009. A comparison of quantitative
not a large age target group. A Larger age group would provide and qualitative data from a formative usability evaluation of an augmented reality
more accurate and valid results; therefore, the influence of wear- learning scenario. Informatica Economica 13, 3 (2009), 67.
able AR on shopping behavior with a broader age group needs to [9] Megan Johnson. 2015. Augmented Reality as a Sales and Marketing Strategy in
Fashion Retailing. Ph.D. Dissertation. Cardiff Metropolitan University.
be investigated in upcoming studies. The study was performed in [10] SK Ong, ML Yuan, and AYC Nee. 2008. Augmented reality applications in man-
a fixed arranged room which counts as another limitation to the ufacturing: a survey. International journal of production research 46, 10 (2008),
study. To simulate a real shopping experience, another study should 2707–2742.
[11] Hanhoon Park and Jong-Il Park. 2010. Invisible marker–based augmented reality.
be carried out with a familiar experiment environment for each Intl. Journal of Human–Computer Interaction 26, 9 (2010), 829–848.
participant such as their own houses or workplace. [12] Krystof Raska and Tobias Richter. 2017. Influence of augmented reality on
purchase intention: The IKEA case.
[13] Gary J Russell. 2010. Itemized Rating Scales (Likert, Semantic Differential, and
6 CONCLUSION Stapel). Wiley International Encyclopedia of Marketing (2010).
[14] Mark Saunders, Philip Lewis, and Adrian Thornhill. 2007. Research methods.
In conclusion, despite the mentioned limitations, it was observed Business Students (2007).
that AR could increase customer’s interest in shopping and could [15] Filipe Lampreia Anes Estevens da Silva. 2012. Using augmented reality to enhance
increase the time spent on shopping (average 13.54 min on AR and the shopping mall experience. Ph.D. Dissertation.
[16] Philipp Spreer and Katrin Kallweit. 2014. Augmented reality in retail: assessing
5.4 min on the tablet). While it had no significant impact on adding the acceptance and potential for multimedia product presentation at the PoS.
products to the shopping cart, it significantly increased the rate of Transactions on Marketing Research 1, 1 (2014), 20–35.
[17] Jasmina Stoyanova, Pedro Quelhas Brito, Petia Georgieva, and Mariofanna Mi-
purchase (p=0.004). To investigate other impacts of AR on customer lanova. 2015. Comparison of consumer purchase intention between interactive
behavior, further studies with vaster participant groups should be and augmented reality shopping platforms through statistical analyses. In 2015
performed in the future. International Symposium on Innovations in Intelligent SysTems and Applications
(INISTA). IEEE, 1–8.
[18] DWF Van Krevelen and Ronald Poelman. 2010. A survey of augmented reality
ACKNOWLEDGMENTS technologies, applications and limitations. International journal of virtual reality
9, 2 (2010), 1–20.
This work was supported by the HumanE AI Net project. [19] Stefan Wiedenmaier, Olaf Oehme, Ludger Schmidt, and Holger Luczak. 2003. Aug-
mented reality (AR) for assembly processes design and experimental evaluation.
REFERENCES International journal of Human-Computer interaction 16, 3 (2003), 497–514.
[1] Saifeddin Al-Imamy. 2018. The effect of co-creation through exposure to augmented
reality on customer perceived risk, perceived trust and purchase intent. Ph.D.

469