TFM Timea Udvarhelyi-Unlocked

Universidad Politécnica
de Madrid
Escuela Técnica Superior de
Ingenieros Informáticos
Master in Human Computer

Interaction and Design
Master Thesis
Self-Service Kiosk Design Guidelines for a

Digital Retail Ecosystem
Author: Timea Udvarhelyi
Trento, March, 2023

This Master Thesis has been deposited in ETSI Informáticos de la Universidad
Politécnica de Madrid.
Master Thesis
Master in Human
Computer Interaction and
Design
Title: Self-Service Kiosk Design Guidelines for a Digital Retail Ecosystem

March, 2023
Author: Timea Udvarhelyi
Supervisor: Co-supervisor:
Luca Turchet, UniTrento Ricardo Imbert, UPM

Abstract
Self-service kiosks (SSK) have increasingly become ubiquitous in today’s society, particularly as the
automation of self-service technologies is rapidly transforming the now traditional user interactions.
Within the retail industry, omnichannel ecosystems have emerged as a powerful trend that can o↵er
customers unique brand experiences. However, incorporating self-service kiosks into these ecosys-
tems presents new challenges in terms of tailoring the technology to meet user needs and increase
engagement.
The main objective of this thesis is to explore how customer engagement can be enhanced within
a retail ecosystem through the use of self-service kiosks. As a result, a set of guidelines has been
developed that can be employed by experience design teams. To achieve these goals, a research
methodology has been established that incorporates qualitative methods such as a context analysis
involving naturalistic observation, a future casting workshop with a small representative sample of the
target audience, and an extensive interview process accompanied by usability testing that compares
two iterative versions of the kiosk design using some of the proposed guidelines. The research data
were analyzed using a thematic analysis methodology, which identified the main pain points and future
expectations of customers.
Furthermore, this study has identified several key design considerations for SSKs in the context of
omnichannel retail ecosystems. These considerations include the importance of encouraging usage,
providing flexibility to accommodate di↵erent user needs and prepare for the quickly emerging tech-
nological trends, displaying information in an informative and engaging way, personalizing the expe-
rience for the user, ensuring accessibility for all users, connecting with the broader ecosystem of the
retail environment, and optimizing the size and placement of the SSK. The research also highlights the
importance of considering the cognitive load and decision-making power of the user when designing
SSKs, in order to ensure that the user experience is smooth and seamless.
Overall, the results of this research suggest that self-service kiosks have great potential to enhance
customer engagement in omnichannel retail ecosystems when properly designed and implemented. By
providing a set of guidelines and a research methodology, this thesis aims to support experience design
teams in creating e↵ective self-service kiosks that meet the evolving needs of customers in today’s retail
landscape. The results are primarily dedicated to the digital retail design team of the host company
where the author’s internship was conducted.
i
ii
Acknowledgements
I would like to express my gratitude to my thesis supervisor, Professor Luca Turchet, who guided me
along the way with high expectations and the right amount of freedom to carry out my thesis. I would
especially like to thank him for his kindness and patience.
This project could not be done without the help of the digital retail team at the internship company.
The team o↵ered me support and welcomed my findings with an open mindset and much excitement.
Thank you for encouraging me throughout these past months.
I must express my profound gratitude to my parents for supporting me through my studies. Thank
you for making this study experience possible and being there for me at all times. I love you both
tremendously.
Finally, I must thank all my friends who shared this study experience with me and made my days
brighter and all my friends who supported me from a distance. I am grateful for the new relationships
and for the old ones. To all of you, I thank you from the bottom of my heart.
iii
iv
Abbreviations
BITM Bring It To Me (function)

CSCW Computer Supported Cooperative Work
CX Customer experience
GTM Grounded Theory Method
GUI Graphical User Interface
HCI Human-Computer Interaction
KPI Key Performance Indicators
NPS Net Promoter Score
RFID Radio Frequency Identification
SSK Self Service Kiosk
SUS System Usability Scale
TA Thematic Analysis
TAM Technology Acceptance Model
UCA Usability Context Analysis
UEQ User Experience Questionnaire
UGC User Generated Content
UI User Interface
UX User Experience
WCAG Web Content Accessibility Guidelines
Table 1: List of abbreviations used in the research work
v
vi
Contents
Abstract i
Acknowledgements iii
1 Introduction 1
1.1 The Host Company . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The Problem Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 The Self-Service Kiosk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Purpose and Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 The Research Question . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.6 The Proposed Research Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.7 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Related work 8
2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3 Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.1 Define . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3.2 Search, Select, and Analyse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
vii
viii CONTENTS
2.4 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4.1 A Brief History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4.2 Fields of Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.3 Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4.4 Accessibility and Inclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.4.5 Quality Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4.6 Self Service Technology Design Guidelines . . . . . . . . . . . . . . . . . . . . . 18
2.4.7 Digital Retail Ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.4.8 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3 Design and Redesign 24
3.1 Released Version . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.2 Current Retail Ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.3 Beta Version . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4 Methodology 35
4.1 Methodological Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.2.1 Context Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.2.2 Workshop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.2.3 Usability Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.2.4 System Usability Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.2.5 User Experience Questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2.6 E↵ectiveness and Efficiency Measurement . . . . . . . . . . . . . . . . . . . . . 49

CONTENTS ix
4.2.7 Thematic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
5 Research Process 50
5.1 Context Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.1.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.1.2 Data Collection Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.2 Workshop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.2.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.3 User Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.3.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
6 Results 65
6.1 Context Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.2 Workshop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.2.1 Idea Presentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.2.2 Emerging Themes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.3 User Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
6.3.1 Introductory Question Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
6.3.2 System Usability Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
6.3.3 User Experience Questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
6.3.4 E↵ectiveness and Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.3.5 Open Ended Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

x CONTENTS
6.4 Summary of Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.4.1 Shopping Behaviours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
6.4.2 Sizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.4.3 The Store Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
6.4.4 Functionalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.4.5 Connecting to the Store Ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . 83
7 Design guidelines 85
7.1 Encouragement of use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
7.2 Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.3 Flexibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
7.4 Informative Progress Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.5 Personalized Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.6 Providing Choices to Users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.7 Accessibility and Inclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
7.8 Ecosystem Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
7.9 Reducing the Cognitive Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
7.10 Size Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
8 Conclusion 97
8.1 Summary of Thesis Achievements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
8.2 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
8.3 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Bibliography 100
A Literature Review Process 109
B Context Analysis Template 113
C Workshop Script 117
D Personas 120
E Completed Artifacts 123
F User Testing Script 127
xi
xii
List of Tables
1 List of abbreviations used in the research work . . . . . . . . . . . . . . . . . . . . . . v
2.1 Define step outlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Recommendations for the SSK for physical access . . . . . . . . . . . . . . . . . . . . . 16
2.3 The evolution of quality measures for SST . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.1 Levels of experience defined by the Digital Retail team . . . . . . . . . . . . . . . . . . 40
5.1 Usability testing process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
6.1 SUS scores and their interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.2 SUS individual scores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.3 Prototype A’s thematic means . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.4 Prototype task completion rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
xiii
xiv
List of Figures
1.1 Research process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.1 An SSK device in the store . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2 Landing page showing an unfiltered catalogue of products . . . . . . . . . . . . . . . . 26
3.3 Product detail page example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.4 Store ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.5 Digital footwear wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.1 Interpretation scales of the SUS index . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.2 Percentile rank of the SUS score showing the average score . . . . . . . . . . . . . . . 47
4.3 UEQ scale structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
5.1 Relevant emerging technology presentation slide from the workshop . . . . . . . . . . . 52
5.2 The template of the final artifact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3 Future casting ideation exercise slide . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.4 Age distribution in the participant sample . . . . . . . . . . . . . . . . . . . . . . . . . 62
6.1 Distribution of preferred shopping channels among participants . . . . . . . . . . . . . 69
6.2 Emerging themes: motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
xv
6.3 UEQ comparison between prototype A and B . . . . . . . . . . . . . . . . . . . . . . . 73
6.4 Table of values for UEQ comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
6.5 Prototype A’s answer distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.6 Prototype B’s answer distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.7 Task Efficiency for the two prototypes . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.8 Grouped expectations from the SSK before and after interaction . . . . . . . . . . . . 79
6.9 Filtering panel of prototype B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6.10 Browse while you wait option in BITM function . . . . . . . . . . . . . . . . . . . . . . 84
7.1 Image from the starting video . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
7.2 Image from the starting video . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.3 Concept design of progress bar in fly-out component . . . . . . . . . . . . . . . . . . . 89
7.4 Browse while you wait or stay close to the kiosk choice . . . . . . . . . . . . . . . . . . 91
8.1 The revised process of the research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
F.1 The SUS questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
F.2 The UEQ template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
xvi
Chapter 1
Introduction
In this first chapter, the causality of the research topic and question will be described, briefly establish-
ing base terminology, illustrating the relation between research work and internship, and presenting
the proposed methodology along with an overview of how this work is shown in the following chapters.
1.1 The Host Company
Adidas is a multinational company that designs footwear, apparel, and accessories that are meant to
be worn for sporting activities and everyday wear. It was founded in 1949 in Herzogenaurach, Bavaria,
Germany by Adolf Dasler. In the present day, its three-stripe logo can be instantly recognized across
the world, making the brand a household name.
The motto of the company is ”Through sport, we have the power to change lives.”, which serves as
the driving power of the company, towards the goal of becoming the best sports company in the world
(Adidas, 2018). According to their global brand strategy, they prioritize two key elements: sustain-
ability and inclusivity to leverage their sports credibility.
Being the largest sportswear manufacturer in Europe and reportedly with almost 1000 concept stores
worldwide, Adidas has an extensive reach in the retail industry, which means that their digital tech-
nology is being seen and used by thousands of shoppers daily on a global scale. This goal of becoming
the best in the market is dependent on many factors, and among high-quality products and sports
expertise, the customer experience is also a deciding factor.
Digital experience teams work with great e↵ort to achieve the latter. In my capacity as a digital
1
2 Chapter 1. Introduction
experience designer intern, I have witnessed how the teams bring scalable solutions for an inclusive
and cohesive experience to all their digital products, i.e. mobile applications, web applications, and
digital touch-points. All the presented research was carried out within the role of user experience
designer within the host company.
1.2 The Problem Space
In the aftermath of COVID-19, user attitudes changed towards in-store shopping. While many cus-
tomers started adapting to the trend of online shopping, due to its perceived usefulness and ease of
online purchase (Soares et al., 2022), many shoppers are eager to return to the physical stores after
quarantine (D’Innocenzio, 2021). Not only user attitudes were a↵ected by the pandemic, but the
business ecosystem too (Saputra and Praningrum, 2022).
The digital retail team is working on creating and delivering a seamless, connected, engaging, and
meaningful in-store shopping experience that empowers people to reach their personal goals with the
help of in-store technologies. Their strategic ambition is to provide an omnichannel, best-in-class retail
experience that can create an emotional connection with the brand itself through service experiences
and products that people enjoy using and are happy to come back to. The team defines touch-points
as interactions or points of contact between a user and the organization and defines the omnichannel
experience as the experience a user has while moving across channels throughout their customer
journey.
When talking about digital signage in the commercial domain, in retail experience the used technologies
currently enable users to interact with the brand through storytelling, events, social media, and services
with the help of touchscreens. The most important customer-facing digital signage that is available
in the store is the self-service kiosk, the smart fitting room mirror, and depending on the market the
footwear wall tablet or the footwear wall. It is important to mention here, that personal smartphone
devices are also considered part of the digital ecosystem, due to the in-store mode that has been
recently developed.
These digital touchpoints are supposed to create a consistent customer experience (CX) that guides the
users throughout their journey of discovering, trying on, and purchasing products. However, currently,
the experience feels disrupted as the touchpoints do not have a consistent connection between them.
1.3. The Self-Service Kiosk 3
The disconnected experience is the root of the current challenge that this study aims to tackle. We
suppose that a seamless omnichannel experience would increase customer engagement and would serve
as additional motivation for in-store shopping.
1.3 The Self-Service Kiosk
The primary technology and device chosen for this research work is the self-service kiosk (SSK),
through which the retail experience’s challenges are going to be tackled.
To better understand this technology the perspective of the responsible team inside the company was
examined. The SSK has the purpose of informing and delighting consumers. It allows the user to
browse products and services that the brand o↵ers, not just in the store, but outside of it, too. As
a secondary purpose, it enables brand ambassadors to inform and guide users through the o↵ered
products, o↵ering storytelling to boost engagement. The main functionalities are as follows:
• Users can browse for products with the help of the filtering and sorting options
• Users can search for specific items through a brief description or the name of the product using
a text input with the help of a pop-up keyboard
• Users can get detailed information about a chosen product
• Users can check for online availability of a liked product
• Users can add to wishlist or to shopping bag liked products by scanning a QR code and opening
the Adidas app on their phones
• Users can buy products from the online store starting the flow on the SSK and finishing it on
their phones through the Adidas app
• Users can request footwear to be brought to them in the store in a matter of 3 to 5 minutes
• Users can call store associates for help through the SSK
In the literature self-service kiosks are categorized in the group of on-site self-service technologies
(Dabholkar and Bagozzi, 2002). These technologies include graphical user interface applications that
allow for independent service performance in order to benefit the customer (Meuter et al., 2000).
self-service kiosks are also described by their common traits. These are systems in public locations
(Lazar et al., 2019), that serve a large variety of users by requiring them to perform a part of the
service without assistance (Yen, 2005).
1.4 Purpose and Motivation
The literature finds evidence of a well-defined set of guidelines for both digital signage design and for
self-service technology design respectively. As the two fields of research remain disjunct, this study
aims to fill that gap and unify the relevant guidelines for a self-service kiosk (SSK) in a digital retail
ecosystem, where the kiosk acts as a touchpoint, becoming a digital signage technology.
The main purpose of this work is to investigate existing design challenges, design principles, and
relevant themes for the design process of the SSK and to produce scalable guidelines based on them.
A unique angle that is being highlighted in this study is the set of guidelines proposed to tackle
omnichannel ecosystem challenges, as it is becoming a growing trend (Lindgreen et al., 2021). The
guidelines originate from the found gap in the literature, a revision of principles that the design team
already has in place and the field research carried out during this project. From this, a checklist of
guidelines was created to benefit designers in their design and evaluation processes.
The study also aims to produce guidelines applicable on a global scale, as the kiosk’s constraints tend
to di↵er between markets.
As a personal motivation, I would like to mention that working with and designing for a self-service
kiosk, gave me a great opportunity to step away from what I consider more traditional design, towards
a di↵erent context and type of device that naturally had di↵erent challenges and design guidelines.
Not only was this the first occasion working in the field of e-commerce, but also the first opportunity
to work with a team of professionals in a professional setting.
1.5 The Research Question
In order to tackle the defined challenges, the study aims to answer the following research questions:
1. How can the self-service kiosk increase customer engagement in an omnichannel retail shopping
environment?
1.6. The Proposed Research Process 5
(a) What are the needs and expectations of customers in an omnichannel retail ecosystem?
(b) How can an SSK influence the in-store shopping experience of the customer?
2. Which design guidelines should be followed to achieve touchpoint connectivity?
(a) How might we enhance the current experience, which is rooted in the web interface and
translated to various size displays surpassing PC screen sizes, by designing with touch-led
interaction at its core?
(b) How can this experience be inclusive and accessible?
1.6 The Proposed Research Process
The experience in digital retail is defined by the team using four distinct factors i.e. user, context,
experience, and device. The process of creating a quality user experience starts with identifying the
users and understanding their journeys, then the context needs to be understood. The third step is
outlining the purpose of the experience and lastly, recognizing a device that is most suitable to host
the experience.
Similarly to this process, an individual research plan was constructed and proposed that consists of
both desk research and filed research, which takes advantage of the author’s position situated as an
intern at the Adidas company. As a by-product of this study, all practical findings were tailored to
suit the use case of the self-service kiosk that can be found in the Adidas flagship stores.
Figure 1.1 shows an overview of the research that was carried out.
First, the desk research was accomplished by obtaining a holistic understanding of the fields of usage,
similar case study methodologies, and guidelines in digital signage ecosystem and self-service technol-
ogy respectively that led to the identification of the gap in the literature. These findings were later
used in the context analysis setup, the future casting workshop setup, and the guideline formulation.
Next, a context analysis was performed, providing important details on the SSK’s relation with con-
sumers, shopping assistants, and other stakeholders, as well as data about the role of the device in
the shopping journey and other contextual information. This information will be presented mainly in
chapter 3, as it o↵ers important contextual data that will help the reader understand the system of
the SSK better. Its findings were useful for the guideline formulation and the user study setup.
Figure 1.1: Research process
As the next step, a workshop was conducted to assess the future expectations of the primary tar-
get group. This helped to form an understanding of how users think about the future of shopping
in-store as well as how can emerging technologies influence their perspective about shopping habits.
The workshop was not limited to the SSK only but considered the entire ecosystem of touchpoints
so that purpose and perceived usefulness can be evaluated without bias caused by the necessity of
incorporating the SSK in the system. This was important for the objective of the research ensuring
connectivity between touch-points. As part of the process, touch-points can be redesigned and repur-
posed as well to ensure the overall seamless experience of the consumers. Similarly to the context
analysis, the workshop contributed to the guideline formulation and the user study setup.
Following this, a user study was conducted to discover the di↵erences between how users perceive
the original version and the new versions of the design. The study consisted of usability tests and
semi-structured questionnaires regarding the time convenience, perceived usefulness, and satisfaction
level of the users. This step served as an evaluation of the already established design guidelines and
1.7. Thesis Outline 7
helped with discovering new ones. Finlay, the user study helped finalize the list of design guidelines
for the SSK.
1.7 Thesis Outline
This thesis has been organized into 7 chapters. This section outlines the description of each chapter:
• In Chapter 2, the state-of-the-art related to self-service technology, digital signage, and retail
ecosystem are reviewed.
• In Chapter 3, the fundamentals of the self-service kiosk design are described along with a pre-
sentation of the digital touchpoint ecosystem.
• In Chapter 4, the methodological framework and the utilised methods in the field research are
introduced.
• In Chapter 5 the procedure of the field research is unveiled.
• In Chapter 6, the results of the research can be found.
• In Chapter 7, the unified design guidelines are discussed.
• In Chapter 8, the final conclusions are shown with the encountered work limitations as well as
future work possibilities.

Chapter 2
Related work
In this chapter, the process behind the literature review is presented along with the grouped findings
in the form of a general synthesis. These findings describe the work that has been done so far on the
state of design principles for self-service kiosks, with mentions of application fields, found methodology,
and arising concepts.
2.1 Overview
As one of the first steps in the research work, a thorough literature review was conducted. In this
process guidelines for high-quality literature were considered and a review approach was chosen based
on the initial research of relevant work. The topic was carefully defined and a suitable search was
conducted, after which a grouping process was done to then present the related information in a
thematic manner.
The literature shows the main areas of application with used methodologies, describes the research
domain, the evolution of the measured attributes, and the most important guidelines found in similar
areas.
2.2 Framework
In order to achieve a high-quality literature review the primary purpose of the review was determined.
This literature review aims to describe related research in the chosen field and shows how this work
8
2.2. Framework 9
extends this or addresses a gap in work in the field. This work inevitably also introduces relevant
terminology and provides definitions to clarify how terms are used in the context of the dissertation
(Randolph, 2009). More precisely, the research aims to address the gap in the modern retail ecosystem
concerned with omnichannel connectivity from the primary perspective of the self-service kiosk.
After initial research based on work given as an example, the approach of adopting a Grounded
Theory methodology was taken. Grounded Theory is defined as discovering theory from data (Glaser
and Strauss, 1967). It is a qualitative research method that involves the systematic collection and
analysis of data. It has the advantage of building a rationale for a given research work (Coyne and
Cowley, 2006), identifying important knowledge gaps that are worth inspecting (Creswell, 1998), and
showing the manner in which concepts were studied so far (McMenamin, 2006).
The two authors, B. Glaser, and A. Strauss were the first to develop this methodology and later
took di↵erent stances on how the literature review should be positioned in the methodology. There
are many nuances to Grounded Theory as it generated various discussions among researchers. The
three distinct stances about the literature review’s place in Grounded Theory Methodology (GTM)
are as followed: Glaserian GTM (Glaser, 1978), Straussian GTM (Corbin and Strauss, 2015) and
constructivist GTM (Charmaz, 2014). Glaserian GTM argues that the data collection should come
prior to the literature review. The argument is based on the premise that the literature could bias
the data collection, the analysis, and the theory development (Thornberg and Dunne, 2019). The
Straussian and constructivist perspectives both suggest conducting the literature review at an early
stage of the research work. Straussian GTM takes into account the knowledge and personal experience
of the researcher prior to the literature review. The research deals with interpreted data rather than
raw data (Thornberg and Dunne, 2019). This literature review falls in line with the latter stance. The
discovered information is viewed in context with the previous knowledge of the researcher: a real-life
encounter with the discussed technology, design guidelines studied, initial benchmarking shared by
the host company, a prior search to identify relevant concepts, and various discussions with experts
who have worked with self serve technologies before.
Lastly, a framework proposed by Wolfswinkel et al. was selected for the literature review to be based
on. The authors o↵er a guide with the purpose of enhancing the value of the review by coming up with
an accurate and concept-centric review (Wolfswinkel et al., 2013). This guide o↵ers structure and a
set of actions to all researchers, but it is most appreciated by those who have limited experience in the
field of research. It is also stated that justified deviations from the guidelines are allowed. The five
10 Chapter 2. Related work
stages of the method are as follows: define, search, select, analyze, and present. In the define stage
the researcher is instructed to decide on the criteria of inclusion or exclusion of the evaluated paper
in question. Some given examples of criteria include determining the time period of the publication
and the impact factor of a publication outlet. There is often the case that this criterion is revised
and it can be relaxed or further limited. The next step is to identify the research field. An area
nested within a discipline would be considered a narrow field and familiarity with the constituting
fields is considered advantageous. As a third step of this initial stage, the appropriate outlets and
databases are chosen. Possibly the most influential step in the define stage is determining the search
terms. There are no specified limitations for a number of research terms, however, it is suggested to
take advantage of the di↵erent functionalities of the search engines. It is also instructed to show all
the used search terms to ensure the truthfulness of the research. In the search stage the navigation
through the chosen data set takes place. The process demands documentation and mapping of which
search terms result in which findings.
The select stage requires a thorough filtering process that ensures the quality of the final data sample.
After filtering out doubles, so that each paper appears only once, a sampling is required based on
the title and the abstract followed by a sampling based on the full text. In the occurrence of forward
and backward citations, some additional papers will be selected to be cut out of the final relevant
literature.
The analyze stage aims to provide a system to a so far unstructured stack of studies. The proposed
approach suggests picking a random paper and reading and highlighting any insights that can have
a connection with the review’s scope and research question. Every highlighted instance serves as a
relevant ’excerpt’. Three di↵erent types of coding are suggested to be used together: ’open coding’,
’axial coding’, and ’selective coding’. ’Open coding’ is described as the process that transforms previ-
ously gathered ’excerpts’ into ’concepts’ that capture ideas from the excerpted data and are mutually
exclusive. These concepts are best described as umbrella terms, a classification on a higher level.
These are accompanied by meta-insights and other relevant personal comments about the findings.
Taking advantage of Grounded Theory, in this step open coding enables the conceptualizing and ar-
ticulating of the often hidden aspects of a set of excerpts to ultimately identify a set of categories and
a ’bird’s eye view’ of the findings. This process needs to be repeated for every paper that is part of
the final sample. ’Axial coding’ is meant to uncover interrelations between categories with the help
of ’comparative analysis’. ’Selective coding’ acts as a filtering process for the identified categories.
2.3. Process 11
During this step, the researcher theorizes based on their own judgment.
In the present stage all the fathered relevant knowledge is structurally described and discussed, keeping
in mind that the findings themselves shape the structure of the presentation.
2.3 Process
After an initial search on the topic of the self-service kiosks and a thorough discussion with the digital
retail team, the outline of the broader topic started to take shape. The broad topic was defined as a
contribution to the design process in the field of digital signage in a retail setting. In this case, there
was an evaluation of the state of the SSK and its connection to other digital signage in the flagship
store. This landscape is representative of the global market. After the research question and the
supporting research questions were defined, related concepts were listed based on preexisting design
knowledge.
The next step was to incorporate the selected framework (Wolfswinkel et al., 2013) as precisely as the
use case allowed.
2.3.1 Define
In the define stage first, the criteria of inclusion were defined: in the literature review topic-relevant
papers were accepted i.e. conference articles, book chapters, and journals. The original criteria about
the publication year, which was determined as the interval from 2 years prior to the time of the research
up until the time of the research, was extended to allow fundamental literature to be reviewed. This
extension was made due to the early publication years of essential books that serve as the baseline of
the later conducted research and case studies in the selected sample and the publications that allowed
for a more holistic understanding of the literature and its evolution. When finding a cited paper inside
a literature sample, the acceptance criteria, especially the publication year, were not considered. A
further deciding criterion was the paper’s availability in English, which a↵ected a small percentage of
the literature. In this case, the field of the use case was noted. Lastly, each and every paper needed
to be accessible either through public access or academic credentials.
The field of research contains HCI and Information Science, as they are the main fields of interest
both in the academia and outside of it, followed by Psychology in the capacity of reflection and
Defining the inclusion Topic relevant scientific papers, conference articles, book
criteria chapters published in the past 10 years, available in En-
glish and accessible trough institution credentials
Identifying the field re- HCI, Information Science, Psychology, Digital Retail
search
Determining the ap- Scopus, ScienceDirect, IEEE Xplore, emerald insights,
propriate sources ACM Digital Library, University repositories
Defining the search self service kiosk + design guidelines
terms self-service kiosk + design principles
self-service kiosk + user experience
self-service kiosk + usability
self-service kiosk + design guidelines + accessibility
self-service kiosk + design guidelines + inclusion
smart retail experience + design guidelines + intercon-
nection
smart signage + design guidelines
digital retail + user experience
omnichannel retail + user experience
omnichannel retail experience + design guidelines
Table 2.1: Define step outlines
argumentation in field of user-facing technologies, and Digital retail as it is the field of application
pertinent for the research.
Appropriate sources include Scopus, one of the most commonly known databases that also presents
each paper in a conveniently structured manner, and IEEE Xplore, which is a digital library that
encompasses relevant scientific and technological content, and emerald insights, due to its accessibility
to students, ACM Digital Library, which has numerous open articles that are considered foundational
by researchers and scholars. Furthermore, University repositories were included, as an evaluation of
relevancy toward the field of digital signage and the subtext of finding similar perspectives about
innovation to this research work.
The search terms were defined based on an initial web search and affinity mapping together with the
Digital Retail team. The scope of this exercise was to identify reoccurring keywords and to evaluate the
significance of the final keyword set. The most important ones are as follows: self-service kiosk, design
guidelines, user experience, accessibility, inclusion, interconnection, digital retail and omnichannel.
These terms were combined as shown in table 2.1. This presented list is the final version of the
combination of search terms. During the process of searching a slight redefinition was made.
2.4. Findings 13
2.3.2 Search, Select, and Analyse
The next step was the Search stage, where all the combinations of search terms were used in order.
The process was as linear as possible, with the exception of the deviation where new keywords were
found and later added or subtracted from the list of terms. All the searches were done using Google
Scholar and were selected based on the inclusion criteria after the title, year of publication, keywords,
abstract, and language were examined. During the search, the admission criteria were reevaluated
and the terms were organized based on importance. The year of the publication, tho still relevant
became less significant. It later permitted the discovery of research trends in the literature. Each
search was carried out on the first two pages of Google scholar, usually stopping reading further when
no keywords were spotted in the title or the abstract.
This initial search provided 70 papers as result. All the results were documented in a table format,
where the papers got an assigned number that later was used in the mapping process that was con-
ducted with the help of an affinity map. This map was used to identify general concepts and to show
the relations between the papers.
The analysis started in parallel with the selection process as pares were excluded when no relevance
was found after the first scan. Papers were analyzed linearly instead of random, as it was more
structured for the researcher in this way. Highlights and findings were assessed on the first read and
relevant objectives were found on the second read. Every word, sentence, or paragraph that was
underlined represents a relevant ‘excerpt’. In the end, 50 papers were left and considered. However,
when inspecting the literature, additional papers were found that further solidified certain aspects of
the research. The papers underwent varying degrees of analysis and the relevant concepts were placed
in the affinity map that can be seen in appendix A’s last page.
2.4 Findings
2.4.1 A Brief History
There are multiple related terms that play an important role in the evolution of the self-service kiosks
as we know them today. This machine was invented with the purpose of improving time and cost
efficiency(MFG, 2019). The oldest predecessor of the kiosk is the vending machine. This was the
original form of unassisted retailing. It was introduced in the 1800s and it has helped popularise
self-service as a model and a market trend (Kasavana, 2008). The first interactive self-service kiosk
was developed in 1977 at the University of Illinois by student Murray Lappe. The aim of the kiosk was
to inform students and visitors about campus information such as bus schedules (Kaur and Malhotra,
2018). This interaction meant the first encounter with a personal computer for many people. The
first network of interactive kiosks was developed in 1985 and it contained more than 600 kiosks. The
network’s purpose was to enable item searching through the entire stock of the store and indicate
locations where the searched item was available. This case is the first known application in the retail
industry. The first commercial kiosk dates back to 1991. It was on display at a computer exposition
called ”COMDEX” in Las Vegas. This machine had access to the internet and its purpose was to locate
missing children (Hampshire and Sanford, 2009). By the year 2000, approximately 15% of companies
used some form of self-service technology. Recently the three most popular self-service technology
platforms were recognized as such: vending, kiosks, and Web applications (Kasavana, 2008).
As a result of information technology innovation and a modular shift in service provision, in the present
day, numerous industries have incorporated self-service solutions in their customer-facing procedures.
Among the first to adopt this technology was the retail and the banking industry (Park et al., 2021).
Early adopter grocery stores used self-service technology(SST) in the early 1990s. The popularity of
SST in the hospitality industry is due to its ability to decrease labor intensity for employees, cost
for service providers, and time for customers. The solution is considered convenient according to the
literature (Law et al., 2019).
2.4.2 Fields of Application
In the healthcare industry, there were many fields of applications covered. Their common aim is to
provide a self-service technology that monitors data from the patients, such as signs of wellness, and
supports measurement processes in an efficient and inclusive way (Lyu et al., 2015). A more precise
use case is exploring the field of pediatrics (Gökgür, 2020). A case study was found for iris-capture
methods evaluating the usability of the biometric technology (Sirotin et al., 2016). There are many
applications in the travel, tourism, and hospitality (TTH) sector (Saputra and Praningrum, 2022) as
it is relevant for them to stay competitive. There are scientific investigations found that utilize and
evaluate self-service technology for many use cases such as restaurants(Park et al., 2021), supermarkets
(Wang and Han, 2022), airlines (Abdelaziz et al., 2010; Chang and Yang, 2008), museums (Aziz et al.,
2.4. Findings 15
2020), libraries (Garcia, 2019), banks (Curran and Meuter, 2005) and in the virtual realm too (Bolton
et al., 2018; Kral et al., 2022). In the retail industry, SST quality measures are examined for self-
checkout machines (Lee et al., 2013), exploring perceived quality measures.
2.4.3 Challenges
Diverse papers list design challenges that apply to any self-service technology. There have also been
industry-specific challenges identified in the process of the litterateur review. Describing the general
context in which the kiosk is used,
General design challenges emerging from the nature of the use of the SSK are listed as follows: time to
complete the task, flexibility of input and output, environmental noise, ease of learning/ obviousness,
and privacy (Lazar et al., 2019). The users of the kiosk generally have time-dependent conditions,
where they are potentially agitated and there can be numerous customers queuing behind them.
Kiosks must consider a wide range of users: they must be inclusive and accessible. Because of the
infrequency of use, these machines need to be obvious to use. In the case of personal data being
requested by the system, designers must be precocious about how they display sensitive information.
Lastly, environmental noise must be considered. In the case of sound used for the SSK, a best practice
in design is to equip the interface with audio control options.
In healthcare, a specific design challenge is obtaining a higher measurement accuracy through the
self-service kiosk and optimizing the interaction between the user and the system (Lyu et al., 2015).
In their design process, the authors present an approach where they generalize a solution meant
for an individual, recognizing the caused difficulty of public-use design. Their evaluation was made
by checking whether the solution meets the need of individuals with a predefined ”error tolerance”.
Another described kiosk in the literature operates with additional features such as smart voice and
the visual assistant function (Gökgür, 2020), to tackle solving the problem of optimized interaction.
These kiosks have several medical devices and sensors integrated into the system.
The public nature of self-service kiosks suggests that social context may also have an important role
in understanding the experience of users when interacting with such products. Yet, this is hardly
reflected in the development of self-service kiosks in which usefulness and functionality are still the
basic, and usually the only concerns (Günay et al., 2014).

2.4.4 Accessibility and Inclusion
For low-vision users, it is recommended that the kiosk has a large size, a high contrast, and inter-
nal illumination. For blind or visually impaired people, locating the kiosk can be made easier by
implementing a ”contactless smart card” reader (Gill, 1997).
Physical access constraints are also considered in the literature (Veijalainen, 2017). The lowest point of
any interactive element should not be lower than 0.7 meters and the highest point should not surpass
1.2 meters from the ground (Gill, 1997). These measurements also have a recommended degree of tilt
from the vertical that is displayed in the table 2.2. This consideration is also beneficial for people with
low vision For the radius around the kiosk minimum of 1.5 meters is favored for a fully nonrestrictive
interaction.
There were great e↵orts made to unify accessibility guidelines for self-service kiosks (Lazar et al., 2019).
Eight sets of guidelines were combined and updated in the study, which resulted in the proposition of
their own unified list of guidelines.
Kiosk’s screen height Angle interval from vertical

0 m - 0.9 m 60 - 90
0.9 m - 1.1 m 30 - 60
1.1 m - 1.3 m 0 - 30
Table 2.2: Recommendations for the SSK for physical access
For the concept of inclusion, even tho more research work was found that establishes a strong relation
between the SSK and the elderly (Mattheiss et al., 2011) or older adults (Chung and Park, 2021),
there was a study that positioned its work around the younger generation i.e. children between the
ages of 5 to 12 (Gökgür, 2020). In the study, gamification is used to improve the user experience in
the healthcare sector. The self-service technology described here has the scope of enabling patients to
execute di↵erent medical measurements and warning the user if a consultation is needed. It is used
by both patients and pediatricians. The study shows the positive e↵ects of including gamification for
vital sign measurements using the self-service health kiosk. The literature shows a significant task
completion time decrease for older users when the task is performed sitting down instead of the usual
standing position (Chung and Park, 2021). Another study is investigating the role of gender in the
quality evaluation of self-service kiosks (Lee et al., 2013). It is important to note here that gender
was perceived as a binary factor in the study, di↵erentiating between male and female users. The
2.4. Findings 17
main finding of the study shows that the need for interaction influenced the perceived quality of the
service more for females than for males. Technology anxiety, however, did not show a significant
di↵erence in perception between the two genders. An unexpected find throughout the literature
review was the study describing a socially inclusive storytelling workshop (Hayashi et al., 2014) on
the topic of universal usability for SSK technology. The user group includes di↵erent age categories,
education levels, impairments, and levels of technology knowledge i.e. ”blind, low vision, deaf, low
literate, elderly, young, and digital experts or not”. The participants with no impairments were either
blindfolded or were listening to music. The work proposes guidelines for the participatory practice itself
and o↵ers guidelines for possible communication issues that might arise in such a setting. The study
adopts the popular design guidelines for self-service kiosks as its base (Maguire, 1999). These guidelines
are later described in this literature review. The most important concepts of inclusivity were grouped
in seven points: equivalent use, flexibility in use, simplicity and familiarity, perceptible information,
tolerance for error, low physical e↵ort and size and space for approach and use. (Veijalainen, 2017)
2.4.5 Quality Measures
Over the years the design of the self-service kiosk has evolved, therefore the evaluation and challenge
scenery changed as well. There is an expansion in the number of methodologies and a constant
reiteration of the set of concepts used for accurate assessment. In figure 2.3 the changing conceptualized
attitudes are illustrated to show the emerging trends in perceived quality. These guidelines proposed
in the early literature already address the general design challenges described above.
Many researchers support the argument that the perceived quality of the service is similar to user
attitude (Cronin Jr and Taylor, 1992; Parasuraman et al., 1988). This connection later inspired the
theoretical foundation for measuring SST quality: a systematic service quality measurement framework
(Dabholkar, 1996). Subsequently, the framework was used and iterated on by various researchers.
Originally there were two methods proposed: the attribute-based model, with five relevant attributes,
and the overall e↵ect model. The attribute-based model focused on speed of delivery, ease of use,
reliability, enjoyment and control. It was used with all five attributes (Shamdasani et al., 2008) and
also with fewer conceptualized attributes to fit various use cases (Lee et al., 2013), leaving out the
speed of delivery. This study treats service quality for the SSK as ”a multi-dimensional construct” and
in addition considers technology anxiety (Meuter et al., 2003) and the need for interaction (Dabholkar,
1996) when evaluating quality.

Year Author Attributes

1996 (Dabholkar, 1996) speed of delivery, ease of use (e↵ort and complexity),
reliability (accuracy), enjoyment, control
2002 (Dabholkar and ease of use, performance, fun
Bagozzi, 2002)
2005 (Yen, 2005) efficiency, ease of use, performance, perceived control,
and convenience
2009 (Lee et al., 2009) reliability, personal attention, comfort, features
2011 (Ding et al., 2011) perceived control, service convenience, customer ser-
vice, service fulfillment
2011 (Lin and Hsieh, functionality, enjoyment, security, assurance, design,
2011) convenience, customization
2013 (Lee et al., 2013) time convenience, ease of use, control, enjoyment
Table 2.3: The evolution of quality measures for SST
In the fast food restaurant context, an SST quality scale was developed (Dabholkar and Bagozzi,
2002), considering ease of use, performance and fun.
For the o↵-site SST, a next wave of evaluation proposed efficiency, ease of use, performance, perceived
control and convenience (Yen, 2005).
For the retail industry reliability, personal attention, comfort, and features were conceptualised to be
measured (Lee et al., 2009). This was followed by the development of e-SELFQUAL (Ding et al.,
2011). The popular framework with SSTQUAL proposed in 2011 (Lin and Hsieh, 2011) o↵ers more
complete coverage in evaluating the self-service technology’s perceived quality. It is the framework to
transcend industries and context. This framework was incorporated in the context of airport service
(Singh, 2018) and restaurants too (Park et al., 2021). In the latest study, both the customer’s and
the service provider’s perspectives were considered, thus filling the gap of only customer-facing studies
being conducted.
2.4.6 Self Service Technology Design Guidelines
The work of M. Maguire (Maguire, 1999) serves as the baseline for design guidelines for the SSK, and
more broadly for the information kiosks. In his work, the author describes fundamental guidelines i.e.
defining user requirements, location, and encouraging use, physical access, introduction and instruc-
2.4. Findings 19
tions, language selection, privacy, help, input in general and considering di↵erent hardware, output
in general, and considering di↵erent types of sensory information, structure and navigation, menus,
customization, stakeholders and testing. In the following segment, the most important factors will be
specified with influential takeaways.
• Defining user requirements - In this first crucial step the purpose of the kiosk is defined, as
well as the intended user group by understanding their characteristics and their task goals by
writing specific task scenarios. This user-centered process will provide clarity about user needs.
Although it is not explicitly named in the study, the process of creating personas and writing
user stories that can be edge cases or typical occurrences is described.
• Location - Since the decision of the user to interact with the kiosk is instant, it is advantageous
if it was already established in the user’s mind. For this reason, it is advised to place the kiosk
in the natural flow of the user’s movement and at a point of need (May, 1993). Maguire also
draws a relation between perceived usefulness and the likelihood of usage, assuring a positive
connection between the two. Positioning should also consider the direction of the natural light
sources to avoid lowering the visibility of the content displayed on the screen.
• Encourage use - Here the interface design, especially graphical elements is considered. General
clarity, familiarity, and attractive appearance have the power to influence the user’s decision to
approach the kiosk. An example of best practice is also given: self-running demonstrations
should be implemented to help intuitive interaction.
• Privacy - The need for privacy depends on the type of displayed information. In general, users
do not wish to draw attention to themselves when they are interacting with the system. This
problem also points back to how the kiosk is positioned in the store: placing the kiosk facing
the windows should be avoided.
• Help - There are two main suggestions here. First, a one-click operation to switch on accessibility
mode is proposed. Second, a one-click operation to get any other type of help, such as calling a
store associate, is advised.
• Input to the system - The study indicates that the type of information required for the input
should be specific and easy to understand.

• Output from to the system - In this section, the author discusses the di↵erent types of
outputs. For text output, the minimum size was defined to be 16 points. The font requirements
demand simple, easy-to-read fonts and the avoidance of sans serif. Input and output text is
proposed to be di↵erent from one another to imply the di↵erence between them. The contrast
between text and background should be high, no matter if the background is light and the text
is dark or the other way around. Talking about the use of language, ”computer terms” should
be avoided, as well as terms that are difficult to understand by the average person who might
use the system. For color usage, there should be no more than 5 color codes used. Color coding
in the text is also not recommended. System response should occur within 3 seconds. If that
is not the case, feedback must be provided about the status of the system. Images should be
used carefully so that they do not reduce the impact of the text. For an image to be seen in
the default size, the user should not have to scroll or zoom out. Lastly, music can be used as a
background to set a mood or as an attention grabber. In these cases, a playing length should
be shown.
• Structure and navigation - The author describes the relevance of a home page that acts as
the singular starting point in the kiosk navigation flow. The next topic is the establishment
of a logical hierarchy in the navigation and also a correct representation of it through visual
components i.e. a main menu. In the process of navigation, the user should always recognize the
visual representation of the path present on the screen. Each screen should have a dedicated,
clear title. User controls such as start, restart, finish, back, next, OK, select, cancel, and exit
should be provided along the way. A system auto-reset functionality is in place when no input
is made in ”a few minutes”. A precise time is not determined.
• Testing - In order to understand and evaluate user reactions discussion groups, interviews, and
user-based tests are proposed. For the user testing process, the author remarks on a ”fair pro-
portion” of users that have little to no previous experience with the interactive system. Elderly
and disabled people should be involved in the testing, as well as children. For a thorough user
test, the most important scenarios should be formulated into user tasks. Along with performance
measures, user comments should also be recorded.
In the researcher’s personal impression, the above-presented guidelines instinctively entered the com-
mon sense of experience designers since the publication. I would also argue that the placement of
the SSK needs to be strategic and it needs to consider the overall context of use, it should be easily
2.4. Findings 21
noticeable, or at least its presence should be indicated upon entering the location in which it is rele-
vant. In the section describing the testing guidelines it is important to remark that since the year of
publication, the kiosk interaction has become more intuitive.
Throughout the decades of research in HCI, many evaluation systems were recognized and redefined.
Besides recognizing inclusion and e↵ectiveness as the standard objectives for kiosk design (Maguire,
1999), M. Maguire, the renowned lecturer at Loughborough Design School, has also laid down the fun-
damental ergonomic principles for this topic (Maguire, 2014) and paid special attention to accessibility
(Maguire, 2001).
More recently, the focus shifted towards usability and user experience (Siebenhandl et al., 2013). Later
studies (Tala, 2016; Veijalainen, 2017) incorporated the general rules of thumb of heuristics (Nielsen,
1994, 2005) that are found favorable in the design and evaluation process. These guidelines were
originally meant for web interfaces, that would appear on a personal computer screen, however, they
were widely adopted due to their general aspect that allowed for a broad interpretation. The heuristics
can be found in the appendix B.
2.4.7 Digital Retail Ecosystem
There are no unified design guidelines when it comes to the digital retail ecosystem. However as
omnichannel retailing is becoming more popular (Lindgreen et al., 2021), many papers approach
the ecosystem from diverse angles. The importance of the adaption to an omnichannel approach to
retailing in order to remain competitive and meet the changing needs of consumers in today’s digital
age is discussed in a systematic literature review (Asmare and Zewdie, 2022).
Studies found that the presence of digital signage increased overall store satisfaction and reduced
perceived wait time compared to the blank wall condition (Garaus and Wagner, 2019).
A paper proposes a new conceptualization and measurement scale for omnichannel retailer brand
experience, which takes into account the various touchpoints through which customers interact with
a retailer’s brand with the motivation of widening the scope and diversifying the evaluated channels.
The scale consists of six dimensions: sensory experience, emotional experience, cognitive experience,
social experience, brand experience continuity, and brand experience uniqueness (Frasquet-Deltoro
et al., 2021).
Another study is establishing guidelines for the store layout, the technology integration, and the
CX for the future evolution of the physical retail store. (Alexander and Cano, 2019). The authors
recommend integrating various technologies such as mobile apps, beacons, and AR/VR to provide
customers with a personalized and seamless shopping experience. They also suggest using customer
data to tailor the experience. Creating an immersive and engaging environment that fosters customer
loyalty is emphasized, including events and activities that build a sense of community, and exceptional
customer service in-store and online. The same authors also propose the store of the future is proposed,
considering the store as a whole as part of the omnichannel retail ecosystem (Alexander and Cano,
2020).
A book found in the literature provides a practical guide for designing user interfaces based on princi-
ples from cognitive psychology (Johnson, 2020). Topics such as visual perception, attention, memory,
and decision-making are covered. Understanding how people make decisions and remember informa-
tion can help designers create interfaces and experiences that are more e↵ective and user-friendly.
As a result of COVID-19, many companies implemented risk and disaster protocols in their business
(Saputra and Praningrum, 2022). This also means IoT and RAISA integration in their business ecosys-
tem. The paper discusses AI-Based SST as a future field of exploration. This emerging technology
use is predicted to potentially bring tremendous changes to how SST operates (Chen et al., 2021).
2.4.8 Methodology
After carefully scanning the literature the methods with the highest recurrence were defined.
Grounded theory has been used in a wide range of fields, including sociology, psychology, nursing, and
business, among others, to explore complex phenomena and develop rich, detailed theories that are
grounded in the data (Kaptelinin et al., 2014; Demirkan and Spohrer, 2014; Kelly et al., 2019; Hayashi
et al., 2014). It is no surprise as it is a well-known qualitative research method that aims to develop
a theory based on the analysis of data, without any preconceived notions or hypotheses.
Throughout the literature there are numerous examples of studies using thematic analysis (Gökgür,
2020; Garaus and Wagner, 2019; Alexander and Cano, 2020). More than a quarter of the examined
papers mention thematic analysis.
The literature presents a wide variety of methods in their research. Results are often based on
2.4. Findings 23
systematic literature reviews (Alexander and Cano, 2019; Vakulenko et al., 2018) and case studies
(Syrjänen et al., 2012; McKenzie et al., 2022).
User testing is also a reappearing method (Park et al., 2021; Galdolage, 2020; McKenzie et al., 2022).
This method can be used for many purposes such as to evaluate the e↵ectiveness of audiovisual
communication between customers and service providers in remote kiosks (Syrjänen et al., 2012), to
assess the impact of privacy partitions and chairs on the usability and user experience of self-service
kiosks (Chung and Park, 2021), or to explore the key challenges and potential solutions in the design
of self-service technologies (Kaptelinin et al., 2014).
With the purpose to design an inclusive and accessible kiosk a participatory design workshop was
conducted with participants from diverse backgrounds and capabilities (Hayashi et al., 2014; Cinto,
2016). The activity included people with disabilities, older adults, and non-native speakers of the
language used in the kiosks. In another case the participants worked together to identify and prioritize
design principles for improving the user experience of self-service technologies (Kaptelinin et al., 2014;
Garcia, 2019).
Studies use methodologies such as survey-based empirical study (Frasquet-Deltoro et al., 2021), lean
testing (Gökgür, 2020), ethnography (Jones et al., 2010) and design science research which is a research
methodology that aims to develop and evaluate artifacts or innovative solutions that address complex
and practical problems (Nica et al., 2022).

Chapter 3
Design and Redesign
This chapter has the objective to illustrate the state of the art of the SSK presenting its usage and
purpose, context, device-related information, and the overall user experience with the main UI screens.
The currently released version and the beta version in its prototype stage are going to be presented.
3.1 Released Version
The self-service kiosk is part of the organizational environment of the Adidas store. It is a touchpoint
currently available in the biggest stores in the key markets. It is a system currently resembling the
website of Adidas, which is a webshop for apparel and footwear products, as well as accessories.
The following information was obtained through desk research and a conducted context analysis.
Figure 3.1 shows how the SSK is displayed on the store. These devices are located in the stores being
attached to the wall at a fixed height depending on the size of the screens. The device sizes vary
based on the market, however, the sample device and the ones in the test store are 65 inches long
in their diameter with a 16:9 ratio, with a vertical placement, fixed to the wall with a height of 40
centimeters from the ground. It is an Elo device with a 4K touchscreen display built for commercial
use. Regarding the touch technology and capacity, the device has PCAP (TouchPro® Projected
Capacitive) - 40 Touch and Infrared - 20 Touch, with an active area of 142.848 cm x 80.352 cm and a
resolution of 3840 x 2160 @ 60 Hz with 12 additional supported resolutions. The response time-total
of the device is 8 ms typically and it can weigh up to 53.77 kg, depending on how many modules are
attached to it.
24
3.1. Released Version 25
Figure 3.1: An SSK device in the store
The SSK found in the dutch market is not tilted, since it is attached vertically to the wall. We note
that some markets (such as the Russian market) operate with smaller devices that are tilted. These
devices are generally as high as the face of the table, so the user can conveniently reach the entire
interface. The interaction is happening with touch input from the user or by scanning with the help
of a secondary device: the cell phone. The output is only visual, as the SSK does not emit haptic or
audio output at the moment of the context analysis.
The main purpose that the SSK is fulfilling is to provide a wider variety of items that are not available
in the store but can be purchased online. The user can browse or search through the whole product
catalogue and receive information about all the products. Online purchasing is also possible through
the SSK together with the smartphone. Customers can also request shoes to be brought to them to the
physical location of the SSK so they can try them on in the store. Assistance can also be requested
26 Chapter 3. Design and Redesign
with the “call store assistant” function if the user would like to switch to human interaction. An
important aspect is that the interaction with the SSK can be transferred to the user’s mobile phone
at certain moments in the flow, such as the ordering process for a more secure experience.
When my research began, there were already some pain points discovered from user research conducted
so far. The digital retail team pointed out that the store’s touchpoints are not connected enough with
each other, therefore the provided shopping experience is not immersive enough and the quality of the
user experience does not reach the desired level. This led to including another scope of this research:
discovering the expectations and the perceived usefulness of the SSK.
Figure 3.2: Landing page showing an unfiltered catalogue of products
In figure 3.2 the landing page of the current version can be seen as a static image. Starting from
the top element, the full-width picture indicates a commercial video that is constantly playing on the
top of the screen. The section below is the interactive part of the screen, starting with the filtering
indicator showing the current group of products that is shown. Below there are the product cards
placed in a grid system showing information such as the main image of the products, name of the
product, price, available models, and brand subcategory. By clicking a product card, the user will
navigate to a product detail page with the selected product. This grid section is one that the user
can scroll through, allowing users to explore further in the catalogue. The displayed product list will
change automatically after a new filter is applied or removed. On the right side of the interactive part
of the UI, there is a navigation bar present with the functions: home, search, shopping bag, bring it to
me, and call store associate. The home button will redirect the user to this page without filters being
applied. The search will open a keyboard and a search bar allowing users to search for products by
name or description such as OZWEEGO or blue originals shoes. The shopping bag will show what
items were added to the virtual shopping bag in this session of using the SSK. The BITM button will
show if there are any requests in the queue. The call store assistant button will call a person working
in the store, who can help the users with many di↵erent questions. In this case, the landing page is
the same as the product listing page (PLP).
In figure 3.3 we can see an example of a product detail page or PDP, that shows all the necessary
information about a chosen product. In this example, instead of the video on the top, we can see the
hero image of the red T-shirt on a full-width display with a gallery indication on the bottom part of
the image, hinting towards the other images available of the product. Other main sections are the
shirt model information - this shirt is available in two colors and the product information contains a
description, the hero image, product specifications, and reviews. On the right side of the page, the
most important information for the user is summarized. First the collection and rating of the shirt,
below the final price(with a discount in this example), and a size selector with British-market sizing
and a size guide. Under this section, the main functionalities can be seen. Namely the Add to bag
and call store associate functions. The first function will add the selected item to the virtual shopping
bag of the session after an available size was selected from the size selector above. The second button
results in a shopping assistant or store associate coming to the physical location of the device to answer
any question or request that can be done in the store.
Validation Based on Existing Guidelines from the Literature
Based on Maguire’s guidelines (Maguire, 1999) identified during the literature review, the following
conditions were observed:

Figure 3.3: Product detail page example
• Location: there are 3 SSKs located in the store where the observation took place. 2 on the
ground floor and one on the 1st floor. The main SSK is on the right side of the entrance.
According to the shopping assistants, it is the most used one, as it is on the path of the customer
when exploring the store. Right next to the counter where customers finish their shopping
journey. There is one SSK placed in the women’s sportswear section with a di↵erent starting
screen, showing products from the same category as the products around it. The SSK on the 1st
floor is only showing upcoming events and it is used for advertisement more than interaction.
The research only focuses on the main SSK.
• Encouraging use: The current version’s default display screen is the landing page, indicating
the content reachable through the SSK. However, no additional encouragement is present within
the released version.

• Physical access: There are no objects blocking the way around the kiosk. The ergonomics of
the kiosk respect the restrictions of accessibility discussed in the literature review.
• Introduction and instruction: this evaluation criterion is outdated in the opinion of the
author, as there should not be a need for instructions on how to use the kiosk for the users. The
design should be intuitive enough. An encouraging use should be sufficient to interact with the
machine.
• Language selection: The current version of the SSK does not have a language selector. This
leads to many tourists abandoning the SSK in an early phase and searching for a shopping
assistant instead.
• Privacy: For security and privacy reasons the shopping journey is always transported to the
user’s phone before the moment of payment and the introduction of other personal information.
Target Group
In general, Adidas aims to be inclusive and targets all sports lovers and urban fashion lovers. Dis-
tinctions are made between user groups in many ways, according to the purpose of the categorization.
For example, users can be defined by their gender, shopping habits, level of loyalty, or their age and
belonging to a certain generation. Users are also defined by their location, since the company has
a global reach, there is a di↵erentiation among markets. Another di↵erentiation is the channel of
reach and purchase. Consumers might be shopping in Adidas-only stores or through retailers, both
online and o✏ine. It can also be that users will be categorized within a certain context, such as
in-store shopping: we di↵erentiate between tourists and locals, family shoppers and individual shop-
pers, and athletes and fashion enthusiasts. All these categories can be represented through numerous
subcategories.
The target group can be defined as the group of users who visit Adidas stores, regardless of the
frequency of visits, gender, age, and physical attributes. The design team assumes that all users have
minimal understanding of technology operation and they all own cell phones.
One of the main factors of distinction between user groups is based on shopping motivation. Users
are grouped into four motivational archetypes. With this customer segmentation, people can fit into
one or more categories, according to the internal report. The four segments are as follows:
1. Autonomous enthusiast: This archetype of a person is confident and independent. They take
their shopping seriously and their style is important to them. They base the experience on their
own research and past experiences, they know their size and style well, but are still open to
suggestions and accepting help if necessary. They would usually seek non ”regular” advice.
2. Optimistic Explorer: The optimistic explorer is a spontaneous buyer, that likes to spend
time browsing and interacting with the store personnel. When the optimistic explorer enters the
Adidas store, the atmosphere needs to feel ‘right’ - the attitude of store personnel, the vibe, and
the way products are presented. They prefer delight over efficiency and want a social experience
while shopping.
3. Prudent planner: The prudent planner is an independent shopper that knows what they want.
This type of person is looking for efficiency and expertise. This type of person does their research
before going to the store and requires little to no assistance. This person has a di↵erent view
of the brand and the store then the autonomous enthusiast. They prioritize the brand’s reliable
sports credibility over the brand image.
4. Advice seeker: The advice seeker likes to be reassured and make a choice together. They do
extensive research online in advance of a store visit. They rely both on their own research and
the expert’s advice.
If we were to illustrate these archetypes in relation to each other we could see that both the autonomous
enthusiast and the optimistic explorer are open to exploration, while the prudent planner and the
advice seeker want to control their shopping process. The autonomous enthusiast and the prudent
planner are both transactional and task-oriented, while the others, the optimistic explorer and the
advice seeker, are relational and people-oriented.
3.2 Current Retail Ecosystem
The digital retail ecosystem that is being researched in this work is a collection of digital touchpoints
in a commercial store of Adidas.
The store ecosystem can vary between stores and countries slightly, but looking at the Dutch market
we can identify the main touchpoints for the ecosystem. In figure 3.4 the current ecosystem of the
store is shown with the possible communication links indicated.

3.2. Current Retail Ecosystem 31
Figure 3.4: Store ecosystem
First, the Product in the store refers to an item being readily available in stock in the store. These
products are equipped with an RFID tag.
Product online refers to the market’s stock available in the interval of a week if ordered. Stores usually
will have limited stock available in the store but will have more items (with di↵erent sizes and models
available) online.
Figure 3.4 shows Smart mirror in fitting room on the left side, being connected to in-store products.
The smart mirror is located in the fitting room area in the stores. As its main functionality, it is able to
identify products brought inside the fitting room with the help of RFID technology. The information
about the products is then shown on a part of the mirror surface, allowing the customers to get
informed about the price, available sizes, and models, as well as used materials and their benefits.
A new functionality is being currently developed that allows consumers to request di↵erent sizes or
models of the already liked products. With the help of the shopping assistants, these products will be
brought to the fitting room, allowing the customer to stay in the fitting room in the meantime.
The digital footwear wall works with the same base technology of the RFID. This screen is the size of
a wall that is situated in the store. It is a digital signage technology that acts like a shoe shelf that can
be seen in retail stores. When a customer lifts a shoe, the background of the wall changes, displaying
Figure 3.5: Digital footwear wall
information about the shoe. The information includes the price, name of the shoe, materials, and a
QR code that the customer can scan for further information. This digital signage is also shown in
figure 3.5. When the shoe is lifted, the shelf feels the change in the shoe’s location and displays the
correct information. If two di↵erent shoes were to be switched on the shelf and lifted again, the correct
information would be shown again, as the shelves react to the changes with the RFID technology.
Customer’s mobile phone can be connected to the ecosystem as well. There are many QR codes located
in the store that redirect users to the Adidas flagship app, which has a special in-store mode. Users
can request shoes to be brought to them with the “bring it to me” functionality, also referred to as
BITM. The in-store mode is also automatically activated if the user allows the Bluetooth connection
on their phone and turns on the right settings. This mode has additional information about in-store
services such as gift wrapping, item penalization with lettering and/or emojis, a bookable slot with
experts and a running track simulator to be given advice about running gear, pop-up yoga sessions,
and other Easter eggs. The connection to the other touchpoints happens through the app triggered
by pointing the phone’s camera to a displayed QR code in the store.
So far we looked at customer-facing systems, but it is important to note that there are many systems
that the shop assistants work with. These systems help allocate products and check their availability
and are hosted on employee devices which are mainly tablets. Orders that have the pickup address in
3.3. Beta Version 33
the store are also connected to these systems. Figure 3.4 refers to this as the employee tablet.
Finally, the Self-service kiosk is connected to the online product catalogue through a database and
connected to the customer’s phone through the scannable QR codes to continue the shopping journey
from the SSK onto the phone.
When it comes to the communication between touchpoints, most of the connections are established
through online channels connecting interfaces to a digital database. An interesting new technology
is used to connect products to interfaces through RFID technology. Radio Frequency Identification
or RFID is a technology that helps the interconnectivity of devices. The name refers to a wireless
system made up of two components: tags and readers. The tag is a small chip that is detected by the
reader which emits radio waves through its antennas and receives signals back from the tag. These
tags usually communicate their identity and other related information. The amount of information a
tag can store varies from one identification number to several pages of data. The tags can be active
or passive depending on whether or not they are powered by batteries. Readers can be mobile or even
mounted on a post or into the architecture of a cabinet, room, or building. Its main usages include
inventory control, equipment tracking, personnel tracking, patient monitoring, ensuring that patients
receive the correct medications and medical devices, and providing data for electronic medical records
systems. This technology is mainly applied in medical fields, however, retail does not shy away from
it either (U.S. Food and Drug Administration). All the physical products in the Adidas stores have
an RFID chip in the price tag of the product, which is removed after purchase. The RFID chip has
no functionality or any kind of impact outside of the store.
3.3 Beta Version
As this data is confidential, I will only describe the new direction of the designs, without providing a
visual representation.
The beta version of the SSK tries to drift away from the resemblance to the web store, it includes more
information about in-store benefits and is aiming to provide a more personalized experience than the
original version. This version only changes the UI design aspects and UX design concepts. Hardware
constraints do not change, nor does the attached sensory technology (in this iteration). The new
design intends to help users with browsing and tries to minimize the need of using the search option,
as the search requires typing on a touchscreen keyboard which users found inconvenient, - previous
research shows. The new design o↵ers predefined categories that help the users navigate through the
browsing process and a complex filtering component that allows multiple filtering selections from each
filter. The new design also has a video tutorial before the interaction begins that intends to encourage
consumers to use it. This video is playing when no one is using the SSK, and it shows the benefits
of using it by informing about the main functionalities. The landing page of the beta version is no
longer the unfiltered store catalogue, but instead, it shows the 3 main shopping categories grouped
as female, male, and kids. The landing page also showcases the services o↵ered in the store such as
the BITM function and the professional advice given by experts on topics such as running the gift
wrapping service, and item penalization.

Chapter 4
Methodology
This chapter illustrates the chosen methodological framework by explaining the used methods and
taking a look at the paradigm and the data type as well. In the second part of the chapter, the used
data collection methods are presented with the used tools.
4.1 Methodological Framework
HCI research is considered to incorporate diverse research methods even in one research work. The
recognized success in academic publication in the field advocates for this phenomenon (Lazar et al.,
2017). The majority of HCI research falls under the category of empirical contribution or artifact con-
tribution. Empirical work presents either qualitative or quantitative data collected by various methods
found in the literature. Artifact contribution presents the design and development of new artifacts
such as interfaces and mock-ups. Other types of contribution include methodological, theoretical, data
set, survey, and opinion contribution (Lazar et al., 2017).
Digital technologies can be viewed in the domain of HCI in two distinct ways. These can be considered
either socio-technical systems from the concept of social practice - in which case the researcher usually
takes an interpretivist epistemological stance or cognitive artifacts that are extensions of the human
mind - in which case the research usually takes a positivist epistemological stance (Card et al., 1983;
Nygaard and Bergo, 1975). While interpretivism uses exclusively qualitative methods, positivism uses
quantitative methods allowing qualitative methods too. Defining other di↵erences between the two
stances, interpretivism treats the user as a partner in the research and design team, while positivism
35
36 Chapter 4. Methodology
positions the user as an evaluator for every design decision. In the latter case, the role of the user
becomes important in a later step of the design process. Lastly, when describing the cultural back-
ground of the stances, interpretivism is rooted in sociology, anthropology, or ethnomethodology, while
positivism leans toward cognitive psychology or cognitive science.
Looking back at the methodological framework of the desk research part of this work, the litera-
ture review, and the acknowledgment of its nature is important to note. From an epistemological
perspective, Grounded Theory is considered an interpretive method along with meta-ethnography,
meta-synthesis, meta-triangulation, and meta-narratives (Paré et al., 2015). As it was discussed in
the academic context, the interpretivist stance in HCI utilizes studies in naturalistic settings and ac-
tion research approach. In its qualitative analysis along with Grounded Theory, thematic analysis is
mentioned. Methods for collecting qualitative data consist of unstructured observations, and verbal
reports such as in-depth interviews and surveys. As implied by Grounded Theory, interpretivism in
general considers that the researcher is inseparable from reality, the knowledge about the world is
intentionally seen through lived personal experiences. When discussing the reliability of interpretivist
research it is recognized that there are implications caused by subjectivity(Weber, 2004).
Positivism uses controlled experiments that are usually conducted in a laboratory. More recently
controlled in-the-wild studies also qualify as positivist data collection. This stance collects qualitative
data that is either objective or subjective. The former can include performance measures such as
e↵ectiveness and efficiency and behaviors with the help of logs and sensors. Statistics are also in this
category. The latter consists of validated questionnaires. Qualitative data is often quantified and
it includes grid-based observations and coded verbal reports. On a conceptual level, it detects the
researcher as a separate entity from reality and perceives the human mind as separate from objective
reality. Looking at the factor of validity positivism provides certainty. It assures that the collected
data truly measures reality. It is also reliable because the research can be reproduced without changing
the outcome (Weber, 2004).
It is also important to mention that a mixed approach or triangulation is present among the occurring
methods. This mixed approach or triangulation, as the name suggests, utilizes both quantitative and
qualitative methods and data, and the analysis is conducted based on the same research question with
both approaches.
As an outcome of the investigation on how to best construct a sound academic framework, the shifting
4.1. Methodological Framework 37
waves of HCI paradigms were discovered and will be briefly discussed in this paragraph. There are
three main paradigms described in the field of HCI (Harrison et al., 2007). The authors note that these
three paradigms are not the only possible constructions of paradigms in said field and their coexistence
is also possible. In the paper, the two recognized types of epistemology (interpretivism and positivism)
are not directly linked to the described paradigms, further proving the ambiguity surrounding the epis-
temological distinction between paradigms. Laying out chronologically the three paradigms are human
factors, cognitive revolution and situated perspective. Categorizing by the metaphor of interaction and
the central goal of interaction, the first one defines interaction as man-machine coupling with the pur-
pose of optimizing the fit between the two, the second as information communication with the goal of
optimizing the accuracy and efficiency of information transfer, and the third one (situated perspective)
as ”phenomenologically situated” with the purpose of o↵ering support for situated action in the world.
This third paradigm examines questions such as ”What existing situated activities in the world should
technologies support?” and ”How can interaction be supported without too many constraints about
what a computer can do or understand?”. The third proposed paradigm also discusses its stance on
truth and value: an argument must be formed to describe the relationship between the collected data
of the study and the information the researcher seeks to discover. Based on the ubiquitous aspect
of creating an omnichannel ecosystem in a shopping environment and the relevance of the context
in this use case implies that this work will follow the guidelines of the third paradigm: the situated
perspective.
Answering the question of how this work provides a contribution to the scientific world of HCI,
the study o↵ers a solid methodology on design challenge definition, design evaluation, and processes
towards creating unified guidelines. This works also aims to set an example for future researchers
on how to put into practice design methods for usability testing and user expectation assessment in
the digital retail field. This study also shows how to integrate research work in a corporate setting,
aligning with a design team.
In conclusion, this research work takes an interpretivist stance in HCI, using qualitative data gathered
through user-centered data collection methods with a situated perspective. The aim of the research is
to provide an empirical contribution. Based on the literature gap in the field of self-service technology
in an omnichannel digital signage ecosystem, this research work aims to define unified design guidelines
in order to enhance the user experience of customers in a retail setting.

4.2 Methodology
Using qualitative research methods such as observations, interviews, focus groups, and workshops
became increasingly popular in recent years. So much so that over 30% of CHI and almost 50% of
CSCW conference papers used it as a primary method between the years 2016-2018 (McDonald et al.,
2019). These methods provide valuable data about the user experience and can be used to uncover
the underlying motivations, behaviors, and attitudes of users (Dourish, 2004; Nielsen, 1994) along
with their needs and motivations (Lazar et al., 2017). This gathered and evaluated data helps define
the next iteration in the design process (Lazar and Norcio, 2006) and serves as a measuring tool for
efficiency and overall usability (Kirakowski, 2002).
Qualitative data can be characterized as subjective since it aims to understand an individual experience
and perspective, naturalistic, flexible, iterative, and holistic (Lincoln and Guba, 1985).
In order to increase the reliability of the later presented findings, data- and methodological triangula-
tion (Creswell, 1998) was used, to gain multiple perspectives about the topic at hand. We note here
that triangulation is not a guarantee of reliability but it increases the trustworthiness of the results.
In the process, the method of context analysis, workshop and user study was used with the help of
user testing. All methods captured multiple points of data based on observation and interviews.
In order to adequately analyze the data gathered with the methods listed below, there were specific
measurement processes and analysis techniques chosen that led to the results discussed later in this
thesis. The System Usability Scale, the User experience Questionnaire, and the E↵ectiveness and
Efficiency Measures were pointedly used for the User Study Interviews, while the Thematic Analysis
was used also for the Literature Review and the Workshop. All the chosen methodology was presented
to and evaluated by the team before the activity began, providing feedback and suggestions to ensure
the quality of the research process.
4.2.1 Context Analysis
Not only do the features and the interaction styles a↵ect the overall usability of the product, but the
characteristics and tasks of the users do too along with the organizational and physical environment
in which the interaction between humans and machine takes place. Based on the guidelines o↵ered
by Thomas, C. and Bevan, N. (Thomas and Bevan, 1996), the context in this study refers to all the
4.2. Methodology 39
factors which have an e↵ect on the usability of the product excluding the features of the SSK itself. A
Usability context analysis is carried out to ensure a shared understanding of the specification of the
overall Context of Use of a product. It consists of 3 main steps which involve gathering information
about the product and its intended Context of Use.
The context study is composed as follows:
1. Describe the Product and its use
(a) describe the Product to be tested – by completing the Product Report
(b) describes the Context of Use – by completing the Context Report
2. Identify critical components that could a↵ect usability
3. Plan the evaluation
(a) specify the Context of Evaluation
(b) produces a concise Evaluation Plan
(c) specify any usability measures and criteria
This research focused on the first two steps to conduct the usability context analysis, as the evaluation
will be carried out with user tests. The used template can be found in appendix B.
Observation is a great additional practice when it comes to forming a holistic view. There are two main
types of observation: the controlled and naturalistic observation. Controlled observation usually is
performed in a moderated setting and focuses on revealing quantitative data. Naturalistic observation
or observation in ”the wild” is the less structured version of the observation. This method originates
from anthropology and ethnography, where researchers observed and recorded people’s and even ani-
mals’ behavior in their natural settings (Pruitt and Adlin, 2010). In UX, observation in the wild has
become an essential tool for understanding user behavior and identifying pain points in products or
services. The benefit of this technique is that the user does not communicate with the observer and it
is not aware that is being observed, therefore it is not biased by the taught of being observed. Also,
when consumers meet a product in the store (and use the system in real life) – they are much more
likely to encounter the frustrations (and benefits) of real-life use than if they were in a lab following
a set of instructions.
In order to get an accurate understanding of the context, a naturalistic observation was also carried
out, where consumers were observed in the Adidas store.
Interaction level An experience designed for a single interaction such as a

consumer checking the in-store stock in the app.
Journey level The experience a consumer has when researching, trying
on, ordering, and purchasing products.
Relationship level An Adidas member using Adidas products and apps,
maintaining a relationship and receiving support and
recognition for achieving their goals related to sports and
the brand.
Table 4.1: Levels of experience defined by the Digital Retail team
The digital retail team defines 3 levels of experience for the context analysis as seen in table 4.1.
The interaction level is considered the atomic level of the interactions as it consists of one digital
interaction. The journey level is made up of multiple interactions, that can be present on various
channels, building towards a larger goal. The relationship level is defined as many interactions and
journeys over time. On this level, customers are considered members and the relationship between
them and the brand is based on loyalty.
4.2.2 Workshop
Workshops are a popular method in user experience design used for many di↵erent purposes. They
encourage collaboration between participants that are usually di↵erent stakeholders, team members,
or even end users. Workshops serve as a tool that helps design teams collect new ideas that are
grouped together and in some cases even evaluated. They can be great for problem-solving, feature
prioritization, and gaining insights into user needs and pain points. It is a great tool to encourage
active collaboration and establish a shared base-knowledge about the topic of the workshop. It is an
e↵ective way of including stakeholders in the design process and improving collaboration (de Greef
et al., 2016). Workshops as we know them now, date back to the early 20th century in the world of
academia where students and teachers collaborated creatively on design projects.
In UX there are many types of workshops such as design sprints, user research, future expectation
discovery, and usability testing. Design sprints, for example, are a fast way of validating design ideas
and iterating on designs (Tang and Liu, 2017). They are also considered optimal for involving users
in the design process and gaining insights from them, where they become co-creators and o↵er their
4.2. Methodology 41
own perspective about experience design (Morrison and Bell, 2010). Tho there are many types of
workshops, we also take note of the similarities: all these workshops have a structured format with
assigned time intervals for the included activities. There is one or more person, (usually designers)
facilitating and the participants are usually a relatively small group to allow personal opinion sharing
and comfortable collaboration. Active participation is required from all participants and the outcome
oftentimes is an artifact summarizing the ideas or concepts discussed throughout the activity.
Workshops are also present in other domains such as architecture, urban planning, and software
development.
Future Casting Workshop
A subcategory of the workshop that will be further explored in this thesis work is the future-casting
workshop. Future casting workshops are one of the most creative and daring types of workshops in
UX design. They let participants think imaginatively about the future after being presented with a
common base knowledge that serves as the foundation of assumptions later on in the activity. This base
knowledge usually consists of technological advancements, social and cultural changes, and economic
trends. The goal of future casting workshops is to identify potential challenges and opportunities that
may arise in the future and to design solutions that will be able to adapt and respond to these changes.
These workshops can also help designers to identify new opportunities for innovation and to develop
products and services that will be able to adapt to changing user needs and preferences.
This practice has been around for a while and it is widely popular, as emerging trends in technology
are rapidly changing and each iteration can result in more daring outcomes. Future casting workshops
are a great way of preparing for the upcoming changes, having already mapped out potential paths
that technological advancements can take.
The process that was followed for the thesis work is based on the proposition of Steve Brown, who is
a keynote speaker, writer, strategist, and executive coach.
According to the author, this workshop is designed to help organizations with building successful
strategies based on understanding the trends that are and will shape the future of business, technology,
and human behavior. The process involves guiding people to explore and understand the potential
possibilities in a future time frame and then helping them to generate ideas for new products, services,
or experiences that they would like to create for that envisioned future. This allows individuals to
gain insight into what could be possible in the future and develop creative solutions that may meet
the needs of future users or customers (Futurist, 2015).
In the method seven essential steps are described that will be described in the following section.
1. Expert testimony: In this first step experts bring insights about a wide range of disciplines
and share them with the group so that they have a shared knowledge base and help participants
get into the right mindset for further exercises. These must always include social trends, major
technological trends relevant to the domain, business, and ecosystem trends (such as government
regulations).
2. Synthesis: Here participants reflect on the received information and think about what this
means to them and to the organization that they represent. All the shared knowledge is trans-
formed into key ideas.
3. Personas: Participants receive personas which are tools that help users shift their perspective
and think of a di↵erent or more holistic user group, depending on the intentions of the workshop
facilitator and the research group. Personas in general are great tools to empathize and are built
on ethnographic studies or real people. This helps balance out the participants’ unconscious
bias that would gravitate toward designing for themselves rather than a larger (more inclusive)
group.
4. Rapid future casting: This step can be defined as the core of the future casting workshop. This
is a converging step in the design process, where participants work with the previously gathered
insights to come up with various ideas. In rapid future casting, small teams of participants work
together to come up with ideas for future experiences that they hope will better serve the needs
of their assigned persona. Two main categories of trends are selected by the participants, from
the Expert testimony phase and iterated upon, forcing groups to think in new directions. Once
each team has created enough ideas from the rapid future casting phase, they pick the most
promising idea as their focus for the rest of the process.
5. Science-fiction prototyping: This step helps battle-testing the final ideas from the previous
steps. Teams imagine a time further ahead on a timeline, where their idea is already implemented
and working. Each group defines three acts in this step. First, a brief description is given of how
the idea works by an example of a person using the imagined system (or feature). Second, an
4.2. Methodology 43
incident is described where something goes wrong (for example an equipment failure or privacy
violation). The third act is a description of how this previously discussed issue is solved. This
step is based on scientific facts instead of pure science fiction and helps teams to understand
edge cases and critical thinking. It also helps with paying attention to operational details. The
output should be a list of improvements and considerations the team can apply back to their
planned experience and system.
6. Back casting: This process dates back to the early 1990s when John B. Robinson first outlined
this idea at the University of Waterloo. This is a planning method that starts in the future by
asking What do we need to start working on in the last iteration, before the product is launched?
and ends in the present with the question What do we need to start working on tomorrow?.
This is a step-by-step process, where the amount of years a step consists of is defined by the
workshop planner. Considerations can include ideas like a technology that needs to be developed,
infrastructure that needs to be deployed, the talent that needs to be acquired, or partnerships
that need to be forged. In the end, this exercise provides a road map of how to deliver the
proposed experience. This step is essential if working with di↵erent stakeholders that will most
likely work on these projects. It helps with making the first step tangible and giving a sense of
direction.
7. Reporting: This step happens after the end of the workshop and it is up to the researcher
to collect and summarise the ideas that were created previously. The facilitator summarizes all
the major findings of the future casting session in the report, ensuring that all the value of the
workshop is documented.
This method was used mainly because the fact that the system felt out of touch with users’ needs and
also because of a need of discovering consumer expectations and future visions, to better understand
their behaviors and wants. These reasons are also listed in the section When should you use a future-
casting workshop?
4.2.3 Usability Testing
Usability is another trusted and well-known practice in UX design. It is a highly adaptive method
to di↵erent circumstances and cases. Usability testing, also known as user testing is a method that
results in the enhanced user experience of a service or product. The method consists of defining the
main objective of the test and the scenarios that will be carried out, selecting a representative group
of users or potential users, and observing and recording the interaction between users and the product
or service to identify usability, accessibility, and e↵ectiveness issues. The goal is to later improve the
product or service by correcting the issues.
Usability tests can be moderated or unmoderated, in-person or online. Depending on the time and
resources that the researcher has, they can choose which type of test suits them the best. It is
recommended to run at least five user tests to discover the majority of the usability issues and it is
unlikely to discover many more, also the best practice suggests testing with small groups using an
iterative test-and-design methodology, according to the objective of the research (Virzi, 1992; Nielsen,
1994).
In the process of usability testing the tasks were developed in form of prototypes followed by the
system usability scale, the user experience questionnaire, and open-ended questions. The time of
completion and the number of errors were also measured only to gain a more holistic view of the
di↵erences between the prototypes.
The method is becoming more informal as facilitators are relying more on qualitative data rather than
quantitative data. More time is spent on identifying usability problems rather than justifying their
existence by measuring errors and task completion rates and times. The main reason for reducing the
formality of user tests is the acceptance of the value of usability testing. The method is more useful for
identifying problems rather than validating products. This realization also justifies the small sample
size needed for a high-quality result in usability testing (Dumas et al., 1999).
4.2.4 System Usability Scale
Usability can only be defined as a general quality of the appropriateness to a purpose of an artifact
or system rather than a quality existing in an absolute sense.
The System Usability Scale, often abbreviated as SUS, is a scale of ten questions that represents a
global view of objective assessment of a system’s usability (Brooke et al., 1996).
In appendix F the questionnaire is shown. We note here that the word cumbersome in statement 8
was replaced verbally with the word difficult as the researcher was aware, from previous research, that
the original word is not known to many non-native English speakers.

4.2. Methodology 45
The SUS questionnaire is a 5-point Likert scale, meaning that it is a psychometric response scale that
gathers data about the degree of agreement with a set of statements (Bertram, 2007). In the analysis
process of the Likert scale responses can be interpreted as a whole or as individual responses, which are
called ordinal data. We note the potential danger of assigning numerical values to the given answers.
Moving forward, for the purpose of the research, this scale was treated as interval data, therefore
assuming that participants who answered the questionnaire perceive the di↵erences between answers
as equal units. In the 5-point SUS questionnaire answers are: strongly disagree, disagree, neutral,
agree and strongly agree. For example, we assume that the di↵erence between agree and strongly agree
is perceived the same way as the di↵erence between agree and neutral.
The SUS scale yields a score between 0 and 100. For its calculation first, the degrees of agreement
are mapped to numerical values from 1 to 5, where strongly disagree is 1 point and strongly agree is
5 points. Then, the following equation is used for each participant’s answer:
Si = (Sio 5) + (25 Sie ) ⇤ 2.5
Where Si is the final score of participant i out of n participants, Sio is the Sum of the points of
the odd-numbered questions and Sie is the Sum of the points of the even-numbered questions for
participant number i.
Pn
i=1 Si
S=
n
This leads to obtaining individual scores. The final score is collected by calculating the average value
of the individual scores.
Due to the popularity of this method over the past 35 years in which it was used over 10,000 times
with hundreds of products, the score can be interpreted in at least five di↵erent ways. Past usages
made it possible for the SUS score to be grasped relatively to previous products’ perceived usability
(MeasuringU, 2016).
The figure 4.1 above connects the numerical value of the SUS index to the di↵erent scales of interpre-
tation. The top line shows a scale with 3 categories relating to the Net Promoter Score (NPS). The
NPS displays how likely are answer-givers to recommend the system to friends. Promoter scores start
Figure 4.1: Interpretation scales of the SUS index
at a SUS of 81, and detractors are associated with a SUS of up to 53. A score between 54 and 80
shows the uncertainty of this approach and is marked as passive. For this interval, further research
was advised.
The second line in the figure shows the scale of acceptability, where acceptable corresponds to a score
higher than 70 and unacceptable to lower than 50 (Bangor et al., 2008). The score between 50 and 70
was defined as marginally acceptable.
The third line allocates adjectives to the scores, where worst imaginable is a score below 20, between
21 and 50 the adjective poor was used, a score between 51 and 70 was OK, however, suggestions were
made to change this adjective to fair. A score between 71 and 84 is considered good, between 80 and
85 it is considered excellent and above 86 it is the best imaginable.
The grading system treats everything under a score of 51 as a failing grade, an F. A score between 52
and 62 is a D, between 63 and 72 it is a C, between 73 and 78 it is a B and everything above is an A.
Percentiles are not included in the figure, but it is stated that the average SUS is considered to be 68
as shown in figure 4.2.
This method was chosen with the scope of examining the presence of an increase in the perceived
usability of the SSK system, where the measurement is based on a comparison between scores. The two
measured scores represent the two di↵erent prototypes, where prototype B contains design guidelines
based on the conducted literature review and the context analysis conducted by the author, and the
limited user research gathered by the digital retail team.

4.2. Methodology 47
Figure 4.2: Percentile rank of the SUS score showing the average score
4.2.5 User Experience Questionnaire
User questionnaires are generally used in combination with other data collection methods to yield
irritable results. The User Experience Questionnaire (UEQ) was designed with the intention of com-
parison. The original version was created in 2005 with the help of a data analytical approach to
provide practical relevance to the created scales. The Questionnaire used is made up of 52 attributes
grouped in 26 semantic di↵erential pairs, where the two are the opposite of each other. The pairs are
randomized in order and their thematic meaning(Laugwitz et al., 2008).
Unlike the SUS, the UEQ is a 7-point Likert scale structure that is composed of 6 di↵erent scales. The
scales are the following: attractiveness, perspicuity, efficiency, dependability, stimulation, novelty. In
figure 4.3 the categorizations can be seen. And appendix F shows the used template.
Attractiveness is referred to as the pure valence dimension, measuring an overall impression of the
product. It has 6 items on the questionnaire, while all others have 4.
Perspicuity is a goal-directed or pragmatic quality aspect along with efficiency and dependability.
Stimulation and novelty are hedonic qualities.
Perspicuity is a measure of how easy it is to familiarize yourself with the system. Efficiency refers to a
smooth task performance without unnecessary e↵ort. Dependability looks at how in control the user
feels while using the system. Stimulation looks at whether the system is motivating and exciting to
use. Novelty is the measure of creativity and innovation. Whether the product catches the interest of
the users.
Figure 4.3: UEQ scale structure
Calculating the Scale means the raw data is extracted from the questionnaire by allocating the points
from 1 to 7 to each adjective pair. Since the order of the data is randomized in the questionnaire
to minimize answer tendencies, the next step is to transform the data to a common order where
negative terms are moved to the left. From each participant’s total answer, the 6 di↵erent means are
calculated, where each means corresponds to a di↵erent scale (attractiveness, perspicuity, efficiency,
dependability, stimulation, and novelty). Next, the variance and the standard deviation are calculated
with a confidence interval for each scale. We note that this method does not yield a final singular
score, as it cannot be interpreted on its own.
This method was chosen because the use case matches the scenario in which the questionnaire is
usually used. The UEQ is typically used for comparing improvements by measuring the UX of a
new version, it is also used to determine the areas of improvement (Schrepp et al., 2014). Other
use cases include comparing direct competitors in the market and measuring an accurate level of UX
in a product. This method also allows for further constant monitoring of the same tasks with later
versions, which can benefit the digital retail team in monitoring their progress in the future.
4.2. Methodology 49
4.2.6 E↵ectiveness and Efficiency Measurement
Usability is a quality attribute that measures how easy user interfaces are to use. Usability is a manda-
tory requirement for any user interface and system. When we talk about usability, we are actually
referring to several defining factors. Depending on the source these factors vary in number. The
author agrees the most with the seven-attribute definition of usability. The seven attributes to make
up usability are e↵ectiveness, efficiency, safety, utility, learnability, memorability and satisfaction.
Most attributes are measured through quantitative data, but e↵ectiveness and efficiency are on the
quantitative side.
”Efficiency is doing things right; e↵ectiveness is doing the right things.” according to Peter Drucker,
a management consultant.
In this study, e↵ectiveness is defined as the degree to which something is successful in producing a
desired result. This research measures it with the help of a binary statement: Was the participant
able to complete the task?
Efficiency is defined as the ability to accomplish something with the least amount of wasted time
and e↵ort. This is measured with the help of counting mistakes and measuring the time it takes for
participants to complete a given task.
4.2.7 Thematic Analysis
Thematic Analysis or TA is an accessible, flexible, and increasingly popular method of qualitative
data analysis. It is used to identify patterns in the data. This method is used in scientific fields
involved with studying human behavior. This method requires coding the data, identifying themes,
and producing a final report. This method requires careful attention to detail and a rigorous approach
to ensure that the themes identified are valid and reliable (Guest et al., 2011).
Chapter 5
Research Process
In this section, we are bridging the methodology to the measured results by describing how the
researcher approached the used methods to connect them to the later defined design guidelines. Details
are provided for each method’s preparation and facilitation processes.
5.1 Context Analysis
Context is important as it has the power to change, add or remove some features of the SSK. A
big di↵erence between the SSK and other digital products that the company has, such as the app is
that the user has immediate access to the clothing items and footwear once in the store, therefore
descriptive information loses importance for the SSK.
5.1.1 Data Collection
Based on the product and context report described in the methodology, a specific template was
constructed that provided all the necessary information for further measurements. This template is
available in the appendix B.
5.1.2 Data Collection Process
For the data collection of the context analysis, the tasks can be split into two main parts: the first
part took place in the store and the second part took place in the office.
50
5.2. Workshop 51
The researcher spent 3 hours in the store, in which the observations and the unstructured interviews
with store associates took place. The observation took place on 10.12.2022 (a Saturday), 2 weeks after
the store’s opening date. The location of the store is adidas Brand Center Amsterdam, Kalverstraat 87,
1012 PA Amsterdam, Netherlands. The note-taking was not only based on the previously constructed
templates but also on the instinct of the researcher to take notes about interesting happenings and
observed context, as naturalistic observation usually happens. This helped with journey-level and
relationship-level information.
The second part of the analysis was conducted in the office, where the researcher had unlimited time
to interact with the SSK, without disturbing the shoppers in the store. This was a great opportunity
to assess the SSK on an interaction level. At this point, a cognitive walk-through was also done with
the SSK to map out usability issues with the current version.
The results of the context analysis are the context report and initial usability issue findings. Outcomes
are included in chapters 1, 3, and 6.
5.2 Workshop
This workshop aims to uncover future expectations from an in-store shopping experience from the
perspective of Gen Z and young Millennials as the digital retail team is mapping out work for both
the immediate and the long-term future.
For this activity participants were selected with a non-probability sampling technique based on their
availability. The sampling is finished when the total amount of participants (sample saturation) and/or
the time limit (time saturation) are reached (Martı́nez-Mesa et al., 2016). This same technique was
used for the user testing participant selection as well. The di↵erence was that for the workshop only the
group of 15 interns was asked and the selection happened based on the time slot that was convenient
for both the facilitator and the majority of the participants.
The workshop was an in-person, moderated group session that took place on 08.12.2022 at the Adidas
office with the location: Hoogoorddreef 9A, 1101 BA Amsterdam, Netherlands.
The 8 participants were in a 2:6 ratio with more female participants attending than male participants.
As the workshop included interns only the age range was between 20 and 27. Interns were not related
to the development of the SSK design in their line of work. Some participants had data analysis
52 Chapter 5. Research Process
backgrounds while others came from a marketing or finance background. Participants were of di↵erent
nationalities: Chinese, Dutch, Indian, Italian, Portuguese, and Romanian.
As a creative narrative technique, Science Fiction Prototyping is a popular method in the commercial
sector. Engineers and designers use prototypes to determine the details of the designs and the complex
interactions of emerging technologies. Prototypes serve as a gateway between feature requirements
and the finished product (Johnson, 2011).
Next, the helping materials will be presented that were used for successful workshop facilitation.
Technological Trends
As part of the expert testimony, the facilitator explored emerging technologies and selected the ones
considered the most relevant to the aim of the workshop. These were later presented in the activity
with a visual representation shown in figure 5.1.
Figure 5.1: Relevant emerging technology presentation slide from the workshop
Starting with the vision of the brand the interactive wall was presented. This wall acts as a social
media enabler, allowing users to take pictures, search for inspiration and experience an immersive
environment in the store. Participants were not told that this idea already belongs to the brand.
5.2. Workshop 53
Next, the robot shopping assistant was shown, and its main functionalities were described. The robot
o↵ers in-store guidance, catalogue browsing, and operation through voice command and touchscreen.
The Kinect game is a device similar to the SSK in that it has similar hardware constraints and an
additional motion sensor that enables responding to user movements. The purpose of the game is to
educate about recycling, as the game itself is cleaning up a beach by virtually grabbing plastic bottles
and other garbage from the beach and throwing it away in the correct bin.
The Tilt brush developed by Google, first released in 2016, is an artistic tool allowing artists to create
three-dimensional artwork and even designs (clothing, footwear, or any other product) in a virtual
reality setting using light painting through a headset and controllers. this is now an open-source tool,
free for anyone to use with the right equipment. Overall, Tilt Brush is an innovative and powerful
tool that showcases the creative potential of virtual reality technology (Google LLC, 2021).
A hologram is a three-dimensional image that appears to be floating in space, created using laser
technology. Holograms can be viewed without the need for special glasses or other equipment, and
they o↵er a level of realism and interactivity that is not possible with traditional 2D images. Holograms
are also becoming more present in the retail sector and they have the potential to revolutionize many
other industries as well. In the example picture, a shoe design is represented with a digital twin
and show allowing the user to access additional information such as material composition to the user
(LamasaTech, 2021).
The metaverse is an entirely online world shared by millions of people. Experts speculate that the
metaverse could become the next sensation after the internet, creating a new kind of virtual space
where people work, play, socialize and even shop. It is important to note that issues like privacy,
security, and inequality in the metaverse will need to be addressed as the technology develops (Rouse,
2021).
AI-generated picture technology is a form of artificial intelligence that can create realistic images
from scratch using a descriptive prompt in form of input text. This technology uses a system of two
neural networks that work together to generate new images: one is the generator and the other one
is the evaluator. While AI-generated pictures have great potential, there are also concerns about
their potential misuse, such as creating fake images. An example of this technology is the AI DALL-E
developed by OpenAI. It uses a deep learning algorithm called a generative adversarial network (GAN)
to create images from textual descriptions (OpenAI, 2021).

The interactive floor is a type of display technology that allows users to interact with a projected image
or video on the floor surface. The technology uses a special sensor system to track the movement of
people or objects on the floor, allowing them to interact with the image in real-time, and a projector
to display the image or video on the floor surface. Some interactive floors may also have sound or
touch sensors built in to enhance the interactive experience. In the retail setting, this can become
an additional touchpoint in the ecosystem with the potential of guiding the user through the store or
informing them about products or even just enabling a more immersive store experience.
In the presentation drone delivery, face recognition technology, Bluetooth beacons, and 3D printing
technology were also included. All the participants were familiar with these technologies on a base
level.
Social Trends
When it comes to social trends we can see an increase in the demand and creation of user-generated
content or UGC. This content in the e-commerce setting usually means product reviews, social media
pictures videos, and comments. We primarily look at visual UGC. This can be categorized into four
di↵erent groups for Adidas in the following way: influencer-generated content, professional athlete-
generated content, employee-generated content, and regular consumer-generated content.
Influencer and professional athlete-generated content is usually higher quality visual content and o↵ers
images earlier in the release cycle due to exclusive early-date access to specific products, while employee
and regular consumer-generated content tends to be more realistic looking and comes later in the drop
cycle.
Other social trends include the customized social media shopping experience, where the recommenda-
tions are generated based on like-history. The channels of marketing are diversifying on social media,
using marketing on di↵erent platforms such as Facebook, Pinterest, Instagram, Twitter, and TikTok.
On these channels, low-priced products still dominate the selling rates. These channels also use social
polls and quizzes to collect data (Hub, 2021).

5.2. Workshop 55
Ecosystem Trends
The researcher selected three important ecosystem trends that were presented to the participants.
First, businesses are striving to become more sustainable in their processes and with their products
and services. Consumers are becoming more environmentally conscious, and retailers are responding
by prioritizing sustainability in their operations. This includes reducing waste, sourcing sustainable
materials, and promoting ethical manufacturing practices. Second, digital twins which is a technology
that involves creating a digital replica of a physical object such as a clothing item, a store, or even a
process. Using the digital twin can be a cost-efficient testing tool to simulate and analyze the behavior
of the physical system in real-time, enabling improved understanding, monitoring, and optimization
of the system’s performance (Glaessgen and Stargel, 2012).
Personas
Building personas is a great way to communicate empathy not just from designers to developers but
also from designers to any stakeholder. This tool can help bridge the communication gap between two
di↵erent perspectives. It is also used in cooperative design in which case, the usage of personas helps
designers, first and foremost, to understand the user needs more in-depth (Grudin, 1989).
For this workshop 6 personas were created to represent more marginalized user groups and to bring out
ideas that could help with inclusive design. Namely: Ignacio, Anne, Jordan, Helen, Peter, and Morgan.
This practice also encourages participants to shape their points of view and identify similarities and
di↵erences between themselves and the fictional person.
The personas were diversified by age - having fictional people between the ages 17 and 52, gender
identity, background - expat, tourist or local, occupation - student, sportsman, radio host, recruiter,
interior designer, family situation -single, married, with children, interest in the brand - attracted
to the sports credibility, liking the brand association, being interested in the urban streetwear and
shoe collector. Some of the personas are more autonomous decision-makers, while others prefer addi-
tional opinions or recommendations. And while certain personas are efficiency-oriented, others value
the experience factor in their shopping journey. Inspiration was taken from the participatory work-
shop including diverse participants (Hayashi et al., 2014). This workshop tried simulating a similar
environment through the personas being diversified.

All personas have di↵erent goals and pain points. While Ignacio has a language barrier, Anne su↵ers
from loss of central vision, Jordan is on a journey to discovering their sense of self and sense of fashion,
Helen is shopping with 2 young children and is in a rush, and Peter currently has a broken foot and
Morgan is deaf.
The personas were based on the context analysis user group categorization but the researcher took the
liberty to add notes about di↵erent personality traits in one hand to diversify the user group toward
inclusiveness and to add personality traits that help the participants view these characters as real
people. The used personas are included in appendix D.
The Final Artifact
The final artifact (figure 5.2) was constructed to summarise the insights on the top row by having the
synthesized ideas in the left corner and the given printed persona under it. The top right corner hosted
the crazy 8 ideation post-it notes. The science fiction exercise was in the middle and the backcasting
was illustrated with a timeline on the bottom of the A3-sized artifact.
In the preparation phase, participants were asked in person to join the workshop and then a written
confirmation was required where the objective and the duration of the activities were clearly stated.
A room was reserved at the heard quarter office and a checklist was prepared on what tools will be
needed for the workshop.
For the activities post-it notes, pens, and stickers were provided. The helping materials, such as the
personas, the final artifact template, and the schedule (for the facilitator) were printed out in advance.
A small budget was granted for the research that was used to provide sweet and salty snacks, dried
fruits, and beverages for the participants. These were catered to the participants during the workshop.
On the day of the event, the reserved room was prepared by organizing the necessary tools, and snacks
and making sure that all the technological equipment was working. All activities were evaluated
beforehand with the digital retail team. And the results were presented to them once they were
structured into a report. There was no video or audio recording done during the workshop, however,
the researcher took notes during and immediately after the event.
5.2. Workshop 57
Figure 5.2: The template of the final artifact
The workshop followed precisely the methodology of the future casting workshops proposed in chapter
4. The entire schedule can be found in appendix C. Participants were not given the name of the
exercise, but the facilitator wrapped the exercises in a time-traveling storyline.

First, everyone was greeted at the agreed spot 5 minutes before the starting time of the activity. They
were guided to the prepared meeting room and took their places. A short welcome and an overview
of the agenda were given.
To construct a rounded 2-hour workshop first an icebreaker was introduced preceding the future casting
activities. The icebreaker had the objective to put the participants in the mindset of time traveling
and science fiction. They were asked to share with the group and the facilitator their favorite TV show
or movie that has futuristic technology in it. As an example, the animated series Dexter’s laboratory
was given to the participants by the facilitator.
Next, the participants were grouped into teams of two based on how they set down, and the topic
of the workshop was introduced through a small description and answering questions from the par-
ticipants. Since only 8 participants confirmed and only 7 actually showed up to the event, 4 groups
were created, where one person worked alone. After that, relevant trends forecasting was done where
first information about the usages and variations of the SSK was given followed by a general emerging
technology presentation (see figure 5.1) where each technology was briefly discussed.
In the synthesis step, the participants interpreted the given information that was previously presented
to them by answering the questions How do you think this a↵ects the company, your department,
and your team? What would be a good response action? Secondly, how does this a↵ect you as a
user? The results were written on post-it notes that were later placed on the final artifact as the first
important piece of information. The users took on 2 di↵erent thinking hats: as employees and later
more importantly as users. This was mainly done to take advantage of the nonconformist situation
where the users had insights into the inner workings of the company and could orient ideas toward
the company strategy.
The following exercise was the persona distribution and getting familiar with them. This was called
(Choose your fighter) in the agenda slide.
The core exercise was next where participants ideated about in-store shopping experiences in the year
2047 while keeping in mind 2 factors from the emerging trends and the pain point and character of
their given persona. The exercise was presented with the slide in figure 5.3.
Once the groups were ready, they chose the best idea and developed it further in the next step
expending on the answers given with additional contextual information. This next thinking exercise
took participants further into the future where their idea was already developed. Here they answered
5.2. Workshop 59
Figure 5.3: Future casting ideation exercise slide
the questions:
1. What is your chosen digital signage? What does it look like?
2. Where is it situated?
3. Where is the customer situated?
4. What kind of interaction is happening?
5. What problem is being solved?
In the second part of this exercise the groups had to come up with a problem that their idea is
encountering and in the final part of the exercise they had to fix it, therefore reaching the final idea
that is now covering an additional edge case.
The final exercise before the presentation was backcasting, where the participants looked at how these
ideas could be iterated backward and what kind of intermediate steps needed to be taken to achieve
the desired idea.
The presentations were shown to the digital retail team and the researcher who facilitated the work-
shop. In the presentation, the final idea was described after the persona pain point and the most
important steps in the back casting timeline. This workshop agenda was realized for a second time by
the digital retail team involving di↵erent stakeholders, using the resources provided by the author.
5.3 User Study
The process of the user study requires a lot of planning and preparation. The facilitation of the tests
is also timely and so is the data analysis that follows. The researcher chose to conduct the user tests
in person in a moderated way.
First, the objective was defined, and the participant selection was started. Next, the details of the
evaluation were defined and time slots with participants were scheduled. After that, the prototypes
were designed and a pilot test was conducted resulting in some small changes in the task presentation
and the logistical construction of the user test.
In the usability testing process, the main focus was the connectivity of the SSK to the digital retail
ecosystem. For this purpose, the two di↵erent versions of the SSK were tested in form of prototypes.
The two di↵erent versions of the systems are:
• the prototype based on the current version of the SSK that was live at the time of the testing
and
• the prototype based on the unreleased version that already implemented some of the design
principles proposed to the team
These prototypes present di↵erent interaction flows to connect users to other digital touchpoints that
are also part of the ecosystem. In the process of validating perceived usefulness and added value to
the customer, the test also provided data for other design decisions that are comparable between the
two prototypes, as the design process of implementing changes is non-linear due to the business KPIs
that the team has to take into consideration.
As all participants are internal employees of the company, the problem of confidentiality did not occur,
as they are all bound by company contracts to treat sensitive information accordingly. All participants
5.3. User Study 61
were informed about the type of test that they were about to participate in as well as the way that
the collected data is going to be processed along with the granted anonymity in the reporting. All
participants agreed to be a part of the thesis research as well as to be audio-recorded and to provide
some personal data.
In the process of the user test, audio recording and observation were used to ensure data trustwor-
thiness. All participants had a written version of the questions to avoid deviation from the original
script, with a note to open questions where the researcher asked further unscripted questions based
on the previously given answers. Templates were provided for the di↵erent types of data that were
collected both in a physical and a digital form. Prototype hot spots were turned o↵ for the testing to
not influence the user’s behavior.
Participants were approached in person and via Teams to be invited to the user research, the message
stated the objective of the research without mentioning the measured attributes or the tasks that were
later performed specifically. After the participants chose a convenient time slot out of the available
times using a date picker software, a calendar invite was sent out with further information about the
place and required equipment (all participants needed to have their mobile phone with them) for the
test.
16 participants were selected using the convenience sampling method due to the relatively short time
to conduct the interviews. These participants represented a significant part of the targeted user group.
Out of the 16 participants 8 were female and 8 were male, for an equal distribution. Half of each group
(4 female and 4 male participants) started with prototype A and the other half encountered prototype
B first.
The age distribution can be seen in the chart in figure 5.4 with an average of 26.56 which is rep-
resentative of the younger part of the target population. Looking at the ethnic background of the
participants, the countries of origin include Brazil, Canada, France, Italy, India, Netherlands, Philip-
pines, Poland, Portugal, Spain, and Sweden. This ensured that the product being tested is inclusive
and accessible to users from various backgrounds.
The user tests were carried out in person in the Amsterdam Headquarter office of Adidas in a dedi-
cated area that had a sample device used for testing. The participants first received a non-disclosure
Figure 5.4: Age distribution in the participant sample
agreement and gave informed consent for the collected data to be used in this thesis. Additional
information about the scope of the study was given upon request after the testing process has ended,
to ensure a non-biased state of the participants. Participants were informed about their right to stop
the process at any given moment and were walked through the agenda before beginning the data
collection process. The script used for the interviews can be found in appendix F.
As the above table illustrates, first the introductory questions were asked to gather demographic
data and contextual information about the participants’ shopping habits toward Adidas. Next, they
performed the predefined tasks that led them through the main flow of the upper funnel shopping
experience with the first prototype using the thinking-out-loud process.
Tasks 1 was defined as such: You saw someone wearing a dark red T-shirt, with white stripes on the
side of the sleeve, and with a big Originals logo in the middle, that you really liked. You want to see if
it is available in the store, and buy it. If the shirt is available, you want to try it on, if not you want
to add it to your online shopping list. In this scenario you are looking for a women’s shirt, size 12.
Task 2: You are looking for the shoes that are called ”Vegan superstar shoes” that you want to try
on. In this scenario you are looking for a women’s shoe, size 8.
They were then asked about their experiences in a moderated way describing the perceived usefulness,
5.3. User Study 63
Nr Activity Note
1. Welcome and pro- Breaking the ice with the participants and informed
tocol information consent
2. Introductory Collecting demographic data and easing into the topic
question
3. Scenario presen- Introducing the think out loud process and giving con-
tation text
4. Task 1, Prototype Reading the instructions, participants performs the
A task while the facilitator observes, takes notes and mea-
sures the time, and counts errors
5. Task 2, Prototype The previous step is repeated with the new task
A
6. Questions about SUS, UEQ, and open-ended questions are asked in this
prototype A order
7. Task 1, Prototype Repeating step 4, with prototype B
B
8. Task 2, Prototype Repeating step 5, with prototype B
B
9. Questions about SUS, UEQ, and open-ended questions are asked in this
prototype B order
10. Finishing ques- Asking comparison questions, a quick look into future
tions expectations, and asking participants to express any
thought they couldn’t express so far
Table 5.1: Usability testing process
assessed usability, ease of use, efficiency, and e↵ectiveness as well as their biggest pain points and their
preferred interactions. This data was captured in the data-gathering templates.
As the methodology suggested, the SUS and the UEQ these Likert- type scales were filled first and only
after were the open-ended questions asked. The process was then repeated with the second prototype
followed by the same questions. In the wrap-up questions participants were asked to compare the two
experiences.
To assure a balanced data gathering half of the users were faced with prototype A first and prototype
B second, and the other half encountered prototype B first and prototype A second. All participants
were given the exact same tasks and shown the same images that helped them identify the products in
their tasks. This was done to mimic a real-life situation where the customer engages in research that
predates the shopping involving two di↵erent levels of research. While one is based on a previously
seen item that someone was wearing before, the second item was more familiar to the user as the name
of the product was known. The pictures were shown as the user test did not intend to measure the
memory and recognition capabilities of the user. The entire usability testing was estimated to be 1
hour long, the testing sessions were carried out in an interval of two weeks.
All the gathered data was transcribed in a Google Sheet file for a structured data representation and
in a FigJam file for concept categorization.

Chapter 6
Results
In this chapter, the obtained results are presented and interpreted.
6.1 Context Analysis
Most elements of the context analysis (based on the template found in annex B) were already discussed
in previous chapters. In chapter 1 the product is described. In chapter 3 the specifications are stated
alongside the observations on ergonomics and user types.
Here we will still mention the heuristics evaluation performed on the published version of the kiosk.
Heuristics
1. Visibility of system status: In the current version, the Bring it to me option and the Call store
assistant option show that there are multiple steps in the process of the service, but it is unclear
which step is happening in a given moment as the progression bar does not have labels.
2. Match between system and the real world: Here the digital store catalogue does not align with
the clothes displayed in the store.
3. User control and freedom: In this case, there is no UNDO option for filtering in the browsing
flow. We also note that after 10 seconds of the SSK being abandoned by the user, the starting
screen and the default settings appear.
65
66 Chapter 6. Results
4. Consistency and standards: The layout of the pages is consistent, and functionalities are intu-
itive. The wording of the shopping terms however leads to confusion.
5. Error prevention: Error-prone conditions are limited in the browsing process, as the filtering
categories have predefined options and service cancellation is possible in the case of committing
to an action by mistake.
6. Recognition rather than recall: Product information is always visible on the PDP and a short
summary on the PLP. Users do not need to remember additional information about the system.
7. Flexibility and efficiency of use: In the released version all the filters are on the same hierarchy
level. Shortcuts are suggested in the search functionality. A shortcut of paying for just one item
is directly available on the PDP.
8. Aesthetic and minimalist design: The UI of the SSK is first and foremost designed with func-
tionality in mind. The layout is clean and organized according to the European convention of
top to bottom and left to right. Dialogues do not contain confusing information for the user.
9. Help users recognize, diagnose, and recover from errors: Error messages, such as an empty search
result, are expressed using plain language. In this case, no suggestions are given.
10. Help and document: There is no documentation available for the users, however, a help function
is built in which calls a store associate.
6.2 Workshop
The artifacts provided by the participants can be found in the appendix E. A summary of the findings
will be presented below.
6.2.1 Idea Presentation
Participants came up with various ideas about the future of shopping and the technologies that will
be present in the store. The first group proposed a Robot shopping assistant with a to-do list and
customized suggestions. This robot can exist both online and in the store. The second group pitched
a large robot located in the middle of the store that scans the customers to determine their sizes. The
6.2. Workshop 67
interaction happens with a voice-assisted AI. The third group advocated for the concept of the size
ID. This could be additional information hosted on the app profile. This contains all the preferred
fit and size information for the user and even for other people they are shopping for. Based on this
information recommendations can be made that will fit the customer perfectly. The fourth group
came up with the idea to make the shopping experience smoother. The shopping experience is done
with the help of the cellphone. The liked items that are on the customer’s phone will be directly
brought to the fitting room upon arrival at the store. Fitting room waiting time will be decreased
since the queue is automatized. Size requests, additional item requests, and payment processes will
be happening in the fitting room.
6.2.2 Emerging Themes
After the presentations and a brief discussion about each idea, the following themes were identified.
1. The most important tasks were browsing, try-on, and payment activities. All the ideas had
touched upon the browsing and the try-on process, but only some thought about the payment
method too.
2. Participants agreed on wanting to increase the efficiency of the shopping process: ideas were
given on how to save time at the fitting room, changing size, and at the payment. We can
conclude that efficiency is a primary need for the vast majority of the user group and most
probably for most consumers too.
3. Participants gave a lot of importance to the size and fit technology. While one group proposed a
body scan for the accuracy of sizes, another group thought about how to store this information.
Another group recognized that a pain point is that the fitting room must be exited and entered
again for a new trial, due to choosing the wrong size or not knowing if the size that generally
fits the user will fit in the case of this particular product too.
4. Ideas revolved around robot assistants rather than thinking about how human shopping assis-
tants could help in this situation. One group was against this out of the fear of the shopping
assistants losing their jobs. Their proposition was to find new ways for the shopping assistants
in the store.
5. Participants all agreed on wanting recommendations based on their style. Generally, they were
open to a personalized experience and generated recommendations.
6. Another interesting idea was the change around the store logistics. Products could be moved
with smart elevators or tubes that transport the items from the stock to the customer.
7. We note here that none of the teams mentioned the SSK -in its current form in their future
vision, however, they described several functionalities that can be hosted by the SSK in the
upcoming years, such as a personalized experience with outfit recommendations, a size guide
based on a body scan, queuing for a fitting room.
Observations
• Participants had a hard time associating with the given personas and rather tried to find solutions
mainly for themselves.
• Understanding the concept of time on a larger scale is a difficult task.
• All the presented information, down to the smallest detail, must be intentional, otherwise, par-
ticipants will be distracted.
• Time allocation to activities must always be more generous than the actual time needed.
• The facilitator must be prepared for last-minute changes, technology malfunctions, activity
rescheduling, and participant cancellations.
• A competition and a prize is a positive motivators for enhancing participation.
• The connectivity between the SSK and the fitting room was tried out in the usability testing,
based on the results of the workshop.
6.3 User Study
The user study is one of the most important part of this research. In the following section, the di↵erent
types of results will be elucidated.

6.3. User Study 69
6.3.1 Introductory Question Analysis
After the demographic data questions, other introductory questions were asked about the user’s fa-
miliarity with Self-service technology and its use cases. 87.5% of the participants answered yes to the
question and gave examples of where they have seen and even interact with it before. Reoccurring
answers included fast-food chains, supermarkets, ticketing machines such as movie ticket machines
and public transport ticket machines, competitor brands, and finally the Adidas brand (only 12.5% of
all users had knowledge about the existence of the device to be tested).
Figure 6.1: Distribution of preferred shopping channels among participants
When asked about their preferred channel for shopping for Adidas products and similar products
respondents mentioned 3 diverse channels: in-store shopping, online shopping through the website,
and online shopping trough applications. Retailer brands and single-brand apps were both mentioned.
Online shopping is significantly preferred by the participants, as visible in figure 6.1 by 88.2% of the
participants. Some subjects gave conditional answers, choosing more than one option. ”When I don’t
know the brand I will go to the store because it makes me feel bad if I have to return clothes. When
I know the brand I will just shop online.” - said a participant who opted for both online and in-store
shopping. While others chose two online channels in their answer: ”I use both (the app and the web),
but I prefer the website when I’m home because the screen is bigger and I can open multiple windows
if I want to compare shoes.”

Figure 6.2: Emerging themes: motivation
The emerging themes were categorized by channel, resulting in a visual representation of what is the
driving motivation behind choosing one experience over the other. The two main reasons for choosing
online channels were comfort and the lack of pressure, while in-store shopping has the benefit of o↵ering
an experience where the consumer can interact with shopping assistants, can touch the materials, and
try on the liked items.
The biggest selling point of the web experience was its convenience and the o↵ered sense of security.
The web experience was always associated with a laptop device and never a phone device. The
navigation between the pages is easier and makes product comparison simple not just between same-
website products but also cross-website products. The payment process felt safer for participants on
the website. ”Whenever I buy something more expensive I buy it on the website.”
The convenience of the app was more diverse. Subjects found the phone experience more convenient ”I
can just browse whenever I want, on the bus or in a 5-minute break.” Users noted that they liked that
their credit card information was already linked (with apple pay for example) and made the buying
process much more seamless. One person remarked on how the app has additional features such as
the AR try-on and the find-by-image feature.
6.3.2 System Usability Scale
The first measured attribute was the overall usability of the system and its evolution through the lens
of comparing the two contrasting prototypes.

6.3. User Study 71
Interpretation Prototype A Prototype B

SUS 72,81 75,15
NPS Passive Passive
Acceptance Acceptable Acceptable
Adjective Good Good
Grade C B
Percentile 67 73
Table 6.1: SUS scores and their interpretation
In the process of conducting 16 interviews, participants answered the SUS questionnaire twice in each
interview - once for each prototype. These scores are evident in the table 6.1 that summarises the final
usability indexes of each prototype, where Prototype A is the first version that has been implemented
and Prototype B is a representation of the newer version of the SSK’s GUI.
Looking at Prototype A’s score of 72.81 we can conclude that it scores higher than average (68),
however, it does not reach a promoter NPS score, it can be considered good and acceptable, it can be
given the grade C and it scores higher than 67% of the system in the data set of the comparison.
Prototype B with the SUS score of 75.15 is higher than the score of Prototype A, it is also considered
passive on the scale of Net Promoter Scores. Like Prototype A, it scores the highest grade on the
acceptance scale and receives the adjective good. In terms of an associated grade, it scales higher than
the previous prototype. Prototype B is perceived as better than 73% percent of the systems in the
comparison data sets.
Looking at the table 6.2 we can see the Sum of the individual scores given by the 16 participants.
Points can vary between 16 and 80 as all participants can give an answer corresponding to a natural
number between 1 and 5. Odd number statements are positive in interpretation, while even number
statements are negative. So in comparison, a score is better when it is higher for the positive statements
and lower for the negative ones.
When inspecting the individual scores of statements for Prototype A we find that the highest ranking
positive statement is number 3: I thought the system was easy to use. with a score of 64 out of 80. The
lowest scoring negative statement was number 4: I think that I would need the support of a technical
person to be able to use this system. with a score of 27 out of 80. Points can vary between 16 and 80
as all 16 participants can give an answer between 1 and 5 points.

Nr Statement A B
1. I think that I would like to use this system 50 56
frequently.
2. I found the system unnecessarily complex. 32 36
3. I thought the system was easy to use. 64 62
4. I think that I would need the support of a 27 18
technical person to be able to use this system.
5. I found the various functions in this system 60 59
were well integrated.
6. I thought there was too much inconsistency 29 28
in this system.
7. I would imagine that most people would learn 57 52
to use this system very quickly.
8. I found the system very cumbersome to use. 31 27
9. I felt very confident using the system. 62 63
10. I needed to learn a lot of things before I could 28 22
get going with this system.
Table 6.2: SUS individual scores
Looking at individual scores of statements for Prototype B the highest ranking positive statement is
once again statement number 3, but with only 62 points. The lowest-scoring negative statement is
number 4, with a significant di↵erence from the previous prototype: only 18 points where the minimum
is 16.
It is also interesting to note that even tho overall, Prototype B has a higher SUS, Prototype A is
better received in four di↵erent statements. Prototype A is considered less complex (s2), easier to
use (s3), with functions slightly better integrated (s5), and participants believe that it has a higher
general learnability for most people (s7).
Prototype B has a higher probability to be used (s1), it is much easier for the users to operate alone
(s4), it is less inconsistent than its predecessor (s6), and also less cumbersome in usage (s7), users felt
more confident using it (s9) and finally, users thought they needed to learn fewer things before they
could get going with the system (s10).

6.3. User Study 73
6.3.3 User Experience Questionnaire
The summary of the results of the UEQ can be seen in the following graphs, where first we look at
the comparative data and later a more detailed breakdown of the prototypes.
All the calculations were done by the tool UEQ Data Analysis Tool, which is a free resource from The
UEQ Online web page.
Figure 6.3: UEQ comparison between prototype A and B
Figure 6.4: Table of values for UEQ comparison
In figure 6.3 the comparison between the two prototypes can be seen, where the blue columns represent
Prototype A and the red ones represent Prototype B. The graph was constructed based on the table
in figure 6.4. In this table the Mean is shown in the first column of numbers for each scale, where
STD is the Standard Deviation, N is the number of responses (and participants), and Confidence is
the 5% confidence for the scale means. The Confidence Interval is a measure of the precision of the
estimation of the scale mean. The smaller the Confidence Interval is, the higher the precision of the
estimation and the more you can trust your results - according to the data analysis tool. For example,
the Mean of perspicuity is the highest for prototype A with 1.44 and with a confidence of 0.58. And
the highest score for prototype B is its dependability mean with 1.39 and a confidence of 0.45.
It is evident that Prototype B is superior in attractiveness, efficiency, dependability and stimulation.

Meaning that the overall impression of the product is better in the new version. Users think they
performed the tasks with slightly less e↵ort and felt more in control interacting with this version, and
it was more exciting for them to use the product. The biggest di↵erence is notable in the novelty
aspect. This indicates that prototype B is perceived as more creative and innovative compared to
prototype A.
It is also important to note that perspicuity performs slightly worse than in its predecessor. The error
bars represent the 95% confidence intervals of the scale mean. These bars show an almost certain
interval where the values could fall in the case of repeating the usability testing and interviewing
process. It is meant to warn about the accuracy of the measurements. Since the error bars overlap
in all 6 of the scales, we can conclude that the di↵erence on the 5% level can be still significant.
The confidence interval is considered relatively large, due to the limited data sample and we treat
the information carefully and in combination with other methods such as open questions and the
observations.
Grouped quality Prototype A Prototype B

Attractiveness 1.02 1.24
Pragmatic Quality 1.26 1.33
Hedonic Quality 0.48 0.84
Table 6.3: Prototype A’s thematic means
Summarizing the distinct type of qualities we see that the pragmatic quality is assessed higher at 1.26
than the hedonic one at only 0.48.
In figure 6.5 the answer distribution is visible for prototype A, where the colors represent each answer
from 1 to 7 and the length of each colored bar represents the number of times each answer was given.
For example, for the first attribute pair (conservative vs innovative), shown on the last row of the
distribution table, 2 answers were given for the rating 3 which makes up 12.5% of the data, 5 answers
were given for a rating 4, making up 31.25% of the data, 4 answers were given for a rating 5 (25%)
and 5 for rating 6 (31.25%). No ratings of 1, 2 and 7 were given by the participants (0%).
For Prototype B we can see that in general there are fewer attributes on the negative-to-neutral side
than on the negative-to-the-positive side, as there are significantly more green bars on the distribution
than red ones. For example, looking at annoying/enjoyable there are 5 responses for answer 3, making
up 31.25% of the answers, there is 1 answer for 4 (6.25%), 3 answers for number 5 (18.75%), 6 for 6
6.3. User Study 75
Figure 6.5: Prototype A’s answer distribution
(37.5%) and 1 for 7 (6.25%).
Figure 6.6: Prototype B’s answer distribution
In conclusion, we can determine that prototype B achieved better results and the direction of the
design is adequate, however, we need to keep in mind to not overwrite existing cognitive patterns.
Additional suggestions are included in the design guidelines.
6.3.4 E↵ectiveness and Efficiency
E↵ectiveness was measured by how many participants were able to complete the task without help.
Ultimately all the participants completed the tasks, but in 3 cases participants needed guidance to
finalize task 1. This was due to not recognizing that the item was not in the store and they falsely
believed that they completed the task. This issue was also counted as an error later in the efficiency
measure.
Prototype Completion rate task 1 Completion rate task 2

A 87.5% 100%
B 93.75% 100%
Table 6.4: Prototype task completion rates
Figure 6.7: Task Efficiency for the two prototypes
The average time taken for task 1 was 3 minutes and 22 seconds for prototype A and 4 minutes and
29 seconds for prototype B. For task 2 the average time for prototype A was 2 minutes and 19 seconds
and for prototype B 2 minutes and 6 seconds as the graph in figure 6.7 illustrates. For the first task of
browsing, participants took significantly more time than with prototype B and then with prototype
6.3. User Study 77
A. However with the second task both scores decreased considerably. We take into account that the
participants did not aim to complete their task in the fastest way possible, as they were using the
thinking out loud method that slowed them down. The learnability of both prototypes is observable
in the di↵erence in time between tasks 1 and 2. The gap between prototypes A and B in task 1 can
be explained in the participants’ own words. Prototype A ”feels more like the website”. As most
participants were familiar with the website, they had no problem transposing that experience into
prototype A. Another important factor to consider here is that prototype B included an initial screen
at the beginning of the flow. This screen is a looped video that informs about the main services o↵ered
in the store.
Counting errors had the challenge of determining what is considered an error. The task descriptions
allowed the users to perform the task according to their mental model. Errors were defined as mis-
takes in understanding concepts and not as the number of clicks that a participant performed while
interacting with the prototype. This remark was made as there are many correct ways to perform a
filtering or searching process, but the prototype only covered the most probable paths, ad there were
7 primary filters leading to 5,040 combinations of filtering alone and covering the entire search process
through a prototype would have meant to anticipate all the key search terms that participants could
use.
In the first task with prototype A, 37.5% of users were unable to tell whether the shirt they were looking
for was in the store or not. Most participants realized their mistakes once they started exploring the
PDP, but 12.5% did not realize it on their own.
Regardless of prototype 31.25% of the participants were unfamiliar with the Bring it to me function,
but once they tapped on the option it became apparent what that functionality means. In some cases,
the concept of the shopping bag was not associated with the virtual shopping bag that a person had
on the web page, but participants believed that it will produce a change in their in-store environment
and they will find the shopping items at the cash register.
In general prototype B had more confusion and people got stuck in the flow for more time due to the
fact that the interaction barely resembled the web flow and did not match the subject’s mental model.
This is further discussed in the conclusion part of the chapter.

Prototype Related Errors and Remarks
As prototype A was made up of screenshots with the most possible user flows, the interaction was not
as smooth as it would have been with the developed system. 2 users remarked on the prototype being
”glitchy” or ”inconsistent”, even tho they were able to perform their tasks they defined it as a pain
point.
As prototype B does not have a developed version and the resources for the user test were limited.
Therefore only the most probable flows were designed with many non-functional elements. Users get
frustrated when the interactions they wanted to perform were not predicted by the facilitator and
therefore the buttons did not react to the touch.
6.3.5 Open Ended Questions
Each Participant was asked about their expectations regarding the SSK at two di↵erent times in the
interview. The first time was after the introduction questions when they already saw the device but
have not interacted with it and the second time was at the very end when they already completed all
of their tasks.
During the first assessment, participants remarked on 3 separate types of information, as can be seen
in figure 6.8. Expectations about the product information were outlined. The most reoccurring aspect
was product availability: every participant expected to get information about product availability in
the store. Subjects’ expected size guidance, a display of the reviews, and product details such as
materials and technology used in the manufacturing process and in the product innovation. Store
information was also expected: some users expected to see a map of the store so they can recog-
nize where certain types of products are located. The Bring it to me and the Call store associate
function were also expected even without knowing about the name and process of the functionalities.
Participants expected an immersive experience. ”I expect to see myself with some AR clothes on” -
according to a participant. If not the real reflection of the participant, an avatar representation was
also expected in some cases.
After the user tasks were performed participants were asked again to define their future expectations in
the next five years. This time participants focused more on information that they were lacking during
their interaction. ”The screen should adjust to me in size”, ”I want recommendation based on what I
6.3. User Study 79
Figure 6.8: Grouped expectations from the SSK before and after interaction
like”, ”I want a personalized experience. When I walk up to the screen I want to be greeted by name.”
- subjects said. Users proposed additional functionalities too. The idea of AR try-on resurfaced.
”It would be cool to see a popular outfit on myself as I walk by the mirror.” - the participant was
referring to the SSK and associated it with a mirror. The functionality of a body scan that was
discussed in the workshop came up from distinct participants. The idea was that the SSK would
capture the user’s height and the sizes of the clothes the user is wearing and the calculations would be
sent to the user’s profile in the app. Users also wanted to perform a self-checkout with the machine
as well as get a number on the waiting list for a fitting room entry. Users wanted to get inspiration
and recommendations on the screen and even mix and match outfits virtually through the UI. One
participant mentioned voice interaction. Participants liked that the experience can be transferred onto
their phones, but they also wanted to transfer their experience from the phone to the SSK. Changes
to the already existing functions were defined too: users wanted a clear di↵erentiation between what
is online and what can be found in the store, as they mostly were only interested in the latter. Some
users wanted to see items on sale through a shortcut, in a similar way to how the website has it. Many
participants remarked that they would much rather prefer an interaction on a smaller screen.
When asked about which prototype did the participants prefer, prototype B won with a 10:6 ratio.
Many participants added that they would like a combination of the 2 versions and began to highlight
the parts that they would keep from each version.
6.4 Summary of Findings
As a result of the interview analysis together with the emerging themes of the workshop and the
context analysis, the findings were structured and will be now presented.
6.4.1 Shopping Behaviours
Shopping behaviors and expectations cannot be held under a singular customer archetype. As the
context analysis shows, consumers have varied levels of preparation when it comes to researching the
products and making purchasing decisions. They also need a dissimilar amount of advice and human
interaction. The author adds a new dimension of categorization by examining the relationship between
the consumer and the shopping items pre-purchase. From the user, research knowledge was acquired
on how users group the try-on process. As part of the task, users could request a chosen shirt to be
brought to the fitting room. While some users wanted to first inspect the shirt and only after deciding
to try it on, other users had no desire to do so, they just wanted to shop further and request multiple
items at the same time to be brought to them. There were users who did not want to try on the shirt
in the store, they just wanted to buy it in the quickest way possible. Users also remarked on how
there is a di↵erence between shoes and garments that were handled di↵erently in the store. While
shoes were brought to the customer in the store in whichever area the user was located, apparel was
o↵ered to be brought to the fitting room. This decision of distinguishing the products and where they
are transported in the store split the opinion of the test participants. While some people preferred the
efficiency of the shoes being brought to them in the store, thus saving time not waiting for a fitting
room and not having to change location, other attendees opted for putting together entire outfits in
6.4. Summary of Findings 81
the fitting room including shoes. A secondary motivational factor appeared in a few cases: the privacy
of the trial process. These users felt more comfortable trying on shoes in a fitting room.
6.4.2 Sizing
Based on the facilitator’s observation and the participants’ validation users had issues with the size
of the screen. Most users had to take a step back in the physical space to be able to grasp the entire
screen since the size of the screen was too big for the participants’ liking. ”Standing next to this [the
SSK] is not comfortable for long.” - a participant remarked. The bottom of the screen that holds
information was noticed very late in the journey. Some participants only noticed it while performing
the second task or when answering questions. Even tho the main functionalities are placed at eye
level, secondary information is visible at the bottom of the screen.
The opposite problem with sizing also appeared with some specific components. The Navigation bar
was very hard to touch for participants. Most trials of pushing a button happened in this faze. As
the buttons in the navigation bar were too small and on the right side of the screen, participants had
the tendency to push the button too much on the left side. This issue appeared because the user’s
perspective standing in front of the middle of the screen was slightly distorted towards the side of
the screen. Users were more successful pressing the button once they stepped closer to the right side
of the screen. This occurrence was discouraging for the users, as they were doubting if this issue is
caused by the prototype or by them.
6.4.3 The Store Context
First of all, users were confused by the concept of the shopping bag. The system intended the shopping
bag to represent the same concept as the web store’s shopping bag: holding products that the user
is ready to buy online. However, because of the context change, most participants were confused by
what the shopping bag represents on the SSK. When asked about it, participants thought that they
are storing items in their virtual shopping bags that will later be available in the store for them.
Participants wanted this shopping bag to either be brought to them or to find it at the cash register.
Some participants wanted to have it in the fitting room. Few participants interpreted the shopping
bag as it was intended and saw the value of the functionality. ”Even if the shirt is not in the store, I
can just order it online and it will come to my house.” - one participant said.
Looking at the holistic picture, we highlight that all participants expected the SSK to act as an
extension of the physical store. They automatically assumed that the catalogue of products that they
are seeing is only showing readily available in-store products, not the entire catalogue that includes
online products too. This expectation was confirmed by the shopping assistants during the context
analysis data collection process. Observed in-store users also expressed that they were expecting to
see information about where the product is placed inside the store. Shopping assistants said that this
is a common question they get from customers. The store assistants also remarked on the fact that
users do not understand the store layout and do not distinguish between sections such as ”Originals”,
”Sportswear”, ”newest collection”, and ”special collaboration collection”.
6.4.4 Functionalities
The first evaluated function is the filtering that allows for catalogue filtering on the PLP. In prototype
A all the filters on the PLP are visible separately in a horizontal list, where each filter’s options can
be opened in a drop-down multi-select list. However, in prototype B all the filters are hidden under a
filtering icon, and opening the overlay component contains all the filters in a dashboard-like manner.
Figure 6.9: Filtering panel of prototype B
Participants had split opinions about the filtering process, as some preferred the efficiency of prototype
B while others preferred the simplicity of prototype A. Everyone agreed that the filter component of
6.4. Summary of Findings 83
prototype A not closing automatically was a pain point. Participants expected to close the drop-down
by clicking away instead of the small arrow located next to the filter’s header. Prototype B was too
cluttered for participants. They highlighted the important filters that they would use and wanted the
others to be visible only additionally.
The search function is the second evaluated feature. Previous research conducted by the digital retail
team shows that users find it inconvenient to type on the keyboard provided on the touchscreen. First
of all, it requires too many actions and it is not a seamless experience as the results refresh after each
added or deleted letter. Second, the keyboard appears on the top of the result area, which annoys the
users even with the capability of dragging it across the screen.
6.4.5 Connecting to the Store Ecosystem
Creating a seamless shopping experience was attempted by connecting the SSK’s purchasing flow with
the user’s information through QR code scanning with a smartphone. As all users had the flagship app
installed on their devices the experience was uninterrupted and enjoyable for the user test participants.
They favored having their personal information displayed on their devices rather than on the public
screen of the SSK. This function was convenient as it allowed for the user’s mobility in the store.
”Because I am on my phone now, I won’t hold up the line.” - expressed a participant worried about
creating a queue for the kiosk.
The bring it to me function also allowed for flexible waiting. The estimated time for shoes to be
brought out to users is between 3 and 5 minutes. At this time users appreciated that they were free
to browse the store.
As it was mentioned earlier, users expected to connect their products to the cash register based on the
copy (the written content of the prototype) displayed, however, these needs were not met, as adding
products to the shopping bag resulted in online shopping and home delivery instead of immediate
purchases at the cash register in the store.
The option of finding the products in the fitting room surprised some users while others were expecting
it. Users’ level of trust in the brand influenced their attitude toward the functionality. When the users
were familiar with the brand: they have purchased the brand’s products before, they knew their size
and recognized the used material they were very easygoing and did not need advice or reassurance of
the quality or fit. Novice users preferred to inspect the products before even deciding to try them on.
Figure 6.10: Browse while you wait option in BITM function

Chapter 7
Design guidelines
The literature review served as the basis for developing the design guidelines for the self-service kiosk
in the digital retail ecosystem by laying the foundations inspired by the digital signage technology and
self-service technology design guidelines. These guidelines were further informed by the insights gained
through field research. The resulting guidelines were formulated by integrating the recommendations
from both sources. Many of these guidelines have shared objectives and complementary e↵ects in
enhancing the user experience. Hence, they are often interrelated and reinforce each other in ensuring
optimal usability and accessibility.
7.1 Encouragement of use
The literature defines the guideline to encourage usage by suggesting high contrast design, easily
readable fonts, and location visibility (Maguire, 1999). A more recent study found that videos can
be an e↵ective way to engage customers and increase their likelihood of using new technologies by
educating customers on the features of a new SSK in a restaurant (Smith and Lee, 2022). Based on
the results from the user tests the following guideline was defined.
When the SSK is not used by the consumers a demonstrative video should be playing, that showcases
the main functions o↵ered by the kiosk. In general, these functions should indicate the added value to
the user and inform users about context-specific advantages, if there are multiple ways of performing
them. In the retail context, in-store services should be highlighted so that consumers are aware of
all the benefits of shopping in person. These services are meant to elevate the user’s experience on a
85
86 Chapter 7. Design guidelines
relationship level 4.1. The video should utilize cohesive iconography and font to increase memorability.
The content of the video gives an opportunity to inform the user about omnichannel connectivity.
Figure 7.1: Image from the starting video
This feature was tested in the user testing phase. The demonstrative video was included in the
beta version only (in the newer prototype). Some participants intuitively skipped over the video by
tapping the screen, even when they were presented with prototype B first. Some participants waited
and read the short informative texts. ”Oh you can have things brought to you.” Some participants
later recognized the icon used in the video, as it is the same as the icon in the navigation bar.
7.2. Language 87
7.2 Language
Another important design guideline for a self-service kiosk is to o↵er language selection to accommo-
date diverse customers. This means that the kiosk should provide users with the ability to choose their
preferred language for the user interface and any other information presented on the kiosk screen.
The currently released version of the SSK does not include a language selector. In the Dutch market,
the only available language is Dutch. Shopping assistants noticed how the majority of the consumers
asking for help are non-dutch-speaking customers. This finding primarily aroused from the context
analysis when the in-store SSK was tested by the author.
Figure 7.2: Image from the starting video
Providing language selection can enhance the usability and accessibility of the kiosk for customers who
are more comfortable using a language other than the default language of the kiosk. This is extremely
important in the international context of the SSK, as an important user group is made up of tourists.’s
officially spoken language. To implement this guideline, the SSK should include a language selection
option in a prominent location on the screen, and o↵er a range of language options to choose from.
Additionally, the language options should be clear and easy to understand, and any text or graphics
on the kiosk screen should be available in the selected language. In each European market, at least
the English language should be added next to the country.

The used language for the user tests was English (as was the entry conducted research). This guideline
was briefly tested with prototype B as an additional button was added to the navigation bar. The
button utilized the flag metaphor. Participants were pleasantly surprised to see that they have the
option to change language because the website and the application do not contain this function.
The literature supports the findings of my research, as both indicate the importance of considering
user preferences and cultural di↵erences when designing for internationalized user interfaces (Li and
Takagi, 2010).
7.3 Flexibility
When it comes to the system’s flexibility, the kiosk should be able to adapt to the changing needs
and preferences of all customers. It should be able to support new technologies and features as they
become available.
As a finding of this research, users think about the products displayed on the SSK as the digital
representation of the extant inventory. Users also define the shopping bag as a virtual container
that holds all the items that are readily available for purchase. It is recommended that the default
product catalogue should only display products that are available in the physical store. To convey
this information, an initial filter that is prominently displayed on the first layer of the PLP could be
used. Additionally, users should have the option to extend this filter to include or switch to products
available in the online store, if they so choose.
Flexibility should also consider accommodating emerging technologies. A modular design would assure
the relevance and e↵ectiveness of the system over time while providing a more future-proof investment
for retailers. As mentioned in the future casting workshop and the user tests, the system could include
sensors for augmented reality (AR) try-on or product recognition, which could be added or replaced
as new technologies emerge.
The direction of my findings is in line with what has been reported in the literature, which suggests
that flexibility is a crucial aspect of self-service technology design, as it enables the system to adapt
to changing user needs and preferences, as well as emerging technologies (Cheraghi et al., 2016).
The literature goes on to specify that agile development methods are advised to be utilized to enable
rapid prototyping and testing of new features (Kasinath et al., 2018).

7.4. Informative Progress Display 89
7.4 Informative Progress Display
Figure 7.3: Concept design of progress bar in fly-out component
Based on Nielsen’s first rule of thumb (Nielsen, 1994), the system should always inform the user about
the progression of the requested service. With respect to this rule and in line with the feedback
gathered from the usability test, we remark that the time-consuming functionalities - such as the help
request through calling a store associate or the BITM function, - a status bar should be provided
with descriptive labels visible throughout the duration of the request. The component should contain
information about the request summary, the estimated time needed to complete the request, and a
timeline progression showing the already completer steps and the following steps. The copy should be
short and easily interpretable. Addressing the user directly strengthens the user-friendly environment.
It is suggested to apply animations to the progress bar as well as a countdown of the time estimation.
This guideline is illustrated in the example component in figure 7.3. The example shows the BITM
function with a status bar composed of time estimation and a 3-step progress bar with labels and
short explanations of the current step. Currently, the second step was completed and the request will
be completed in a maximum of 3 minutes.

7.5 Personalized Experience
The kiosk should be able to provide personalized recommendations and suggestions based on the
customer’s profile preferences regarding fit, size, and collection, as well as past purchases and wishlists.
Personalization can appear in many forms, such as greeting the users by name, recommending products
based on the wishlist, or suggesting products to ”complete the look”. All of these were reoccurring
themes in the user research. ”I like to see outfits put together for inspiration, I always look at the
mannequin.” -said a participant. Transferring and scaling this idea could be realized by adding so-
called lookbook pictures that can present the outfits together, but users can also choose just one
product shown in the pictures.
Based on the literature we can add that personalization should be done intentionally and carefully
to achieve a positive impact. This should be transparent and clearly communicated in order to build
trust with the users (Verhoef et al., 2007).
This guideline is being tackled in the beta version already, but it was not tested due to the compar-
ative nature of the usability test, as the current version does not include personalized content. The
membership concept is being translated from the mobile app to the SSK, allowing the users to become
members or utilize their preferences in the recommendations.
7.6 Providing Choices to Users
Similarly to the previous guidelines designers should opt to personalize the experience of the user to
enhance it. This can be done through the provision of choices to the users so that they can pick
their favored option. Designs should present choices in a hierarchically structured way, which is not
overwhelming for the user and allows for easy navigation and selection of options. By providing users
with choices, designers can give users a sense of control and ownership over their experience, this is
essential for user satisfaction (Norman, 2013).
As it was mentioned earlier, the users should be able to choose if they want to see in-store products,
online products, or both. Furthermore, users should be able to choose their method to try-on products.
Here we currently di↵erentiate between shoes and clothing, as the BITM function is only available
for shoes at the moment. As users were worried about creating queues, the option of browse while
7.6. Providing Choices to Users 91
Figure 7.4: Browse while you wait or stay close to the kiosk choice
you wait was created, which transfers the request from the kiosk to the phone and lets the user move
around the store freely. This choice can be seen in figure 7.4. Primary or recommended functions
should be designed with a visual distinction and the choices should be limited.
Other similar choices can be implemented for the trial of the clothes. One choice could be to find
the product with RFID technology and the help of the flagship app’s in-store mode which could help
navigate the user to the product. Another option could be the BITM for clothing items. Users should
be able to decide to try it on once the product is bought by them, if this step is not necessary, they
cloud choose to request the liked items to be brought to the fitting room. For a seamless experience
users could also have the option to find their items directly at the cash register.
Further feasibility limitations must be measured to understand if these functions can be achievable,
as some of them require human interactions.
This falls in line with the findings in the literature mentioning that decision-making is the cognitive
process of choosing a course of action from among several alternatives. In interface design, understand-
ing how people make decisions can help designers create interfaces that guide users toward making
the best decisions (Johnson, 2020). Designers can appeal to users’ emotions through visual design
elements, such as color and imagery, to influence their decisions.

7.7 Accessibility and Inclusion
The self-service kiosk should be designed with all users in mind including those with disabilities. Every
user should be able to use the kiosk independently and easily. As the kiosk is a touchpoint designed
for public use, this is a crucial aspect.
A large legible font should be used consistently in the UI design. High-contrast elements should be
defined with only a few color variations that pass the applicable WCAG AAA success criteria. A
clear hierarchy should be defined for prioritizing primarily functions. Furthermore, the UI should
be designed with the screen size in mind. The bottom area of the screen should not contain crucial
information, as the users will not register it immediately and will be forced to distance themselves
from the screen to be able to comprehend the information at the bottom. From the user research we
have seen that with the current version, users are banding down to see the visualized information.
The text should be included under or above the visual content such as images and video.
It is advised to create an accessibility mode for the kiosk that allows users should to pause the promo
video if they choose to. In this mode, the interface should be navigated through assistive technologies
such as screen readers and speech recognition. Content should also be resizable.
There should be no objects blocking access to the kiosk in the physical space in at least a 1.5-meter
radius from the SSK. The screen size should adapt to the user’s height and interaction area either
by a change in the device’s height or through the changed height of the UI, placing it on the screen
according to the user’s convenience.
To ensure the e↵ectiveness of the accessibility features, it is recommended to test them with users who
have disabilities, as this practice can help identify and address any issues in the design process.
Adidas has its custom font that is part of the Sans-serif font family, making for easy legibility. As
the design uses mainly white (color code: FFFFFF) as the background color and black (color code:
000000) as the font color, with the custom font, this passes the accessibility guidelines proposed for
web pages. Additional testing processes should be introduced for the digital retail team to use.
7.8. Ecosystem Connectivity 93
7.8 Ecosystem Connectivity
Staying ahead of trends is a crucial objective for many brands, as it helps to maintain relevance
in a constantly evolving market. This objective extends to the physical retail spaces of a brand as
well, where it is important to keep up with changing trends in retail design and customer experience.
Brands should also be open to experimentation and innovation in their physical retail spaces, and be
willing to take risks in order to stay ahead of the curve. They should also take into account the needs
and wants of their target audience.
As the literature suggests, the omnichannel ecosystem is an emerging trend and o↵ers a big opportunity
for Adidas to enhance the retail experience of their customer.
The SSK should be connected to various touchpoints in the store such as the customer’s cellphone, the
fitting room, and the shopping assistant’s device. This can help provide customers with a consistent
and cohesive brand experience (Fuentes-Blasco et al., 2017; Lindgreen et al., 2021).
When it comes to the connection between the customer’s smartphone and the kiosk, the flagship app
should be used by scanning the QR codes provided by the SSK. This opportunity could be provided in
the browsing and the shopping journey. Users should be able to add items to their wishlist connected
to their profile from the SSK directly from the PDP or from the shopping bag used on the kiosk.
Users could choose between purchasing products on the spot or ordering them online (also depending
on availability) with the help of their phone.
In order to establish a connection between the cellphone and the kiosk, a code can be generated
through the app which the users can then enter into an input field that is provided in the kiosk’s user
interface.
The shopping assistant’s tablet should be connected to the SSK so that incoming requests for services
can be successfully and efficiently handled. Looking at the logistics behind the service, each kiosk
device should have a unique identifier that the shopping assistant can recognize. The needed time to
complete the request should be editable by the shopping assistant involved in the process. User test
participants decided on the usefulness of the BITM function based on the amount of time it requires
to completer the request. ”If it takes so much time, I would just look for an associate myself.” - said
one participant when asked about the convenience of using the BITM function. Further research is
needed to define the feasibility of this feature.

The connectivity to the fitting room should consist of the reservation of it when the products are
ready for trial. For this connection, a similar request to the previously discussed one needs to be sent
to the shopping assistant who will look for the products and bring them to the right room before the
consumer enters. This journey could also be connected with the mobile phone to handle the fitting
room queue and to provide mobility to the users in their waiting time.
Products should also directly link to the SSK. Users expect to find in-store navigation that helps
them find a product in the store. This could be done with the help of the RFID technology that is
attached to every product in the store. A map should be presented to the users so that they can orient
themselves in the store.
These features could have a long-term e↵ect on how the store is organized. Additional research could
be done on how users would adapt to such changes.
The literature has also confirmed the results of the research about how customers are more likely to
use self-service kiosks when they are integrated with other technologies, such as mobile devices or
augmented reality (Wang et al., 2019).
7.9 Reducing the Cognitive Load
A user’s attention span and working memory are limited, so it is important to minimize the amount
of information and interactions required to complete a task. This can be done by reducing stimulating
components and providing a clean design.
Users expressed that they are overwhelmed on multiple occasions during the usability tests. ”I don’t
know what I am looking at.” expressed some participants when looking at the landing page and the
filtering dashboard of prototype B.
As we know from previous eye-tracking studies, the eyes of the user tend to follow an ”F-shape” when
scanning a user interface. This pattern is characterized by horizontal eye movements across the top of
the page, followed by a vertical movement down the left side of the page, and then further horizontal
movements across the middle of the page. Changing the direction of the movement should be avoided.
On the PLP, filtering should be prioritized and an accordion component should be constructed to
show primarily filtering categories and hide secondary ones. On the PDP the same guideline should
7.10. Size Considerations 95
be used to provide information about a product. Instead of having all the information displayed, some
sections should only have the headers visible and collapsed by default.
Another example is the display of the filtering process, the UI must show the filtering path that the
user has so far selected instead of forcing the user to remember it. This leads to easy error recovery
and a better experience. Similarly, in the BITM function, a summary of information should be visible
so that the user can have transparency about what they requested.
Designers can help users remember information by using clear and concise language, chunking infor-
mation into smaller pieces, and using visual aids to reinforce information (Johnson, 2020).
7.10 Size Considerations
In UX design guidelines, the size of the screen is an important consideration as it can impact the
layout, visual hierarchy, content, and usability of the interface. In addition to the physical screen size,
designers also need to consider the interaction area or touch target size, which refers to the area on
the screen that a user can interact with using touch gestures.
With a screen such as a kiosk, the interaction areas must be limited to the middle of the screen which
should be the eye level of the user. It was already mentioned that the screen should adjust to the user.
Findings show that users struggle with tapping the sides of the screen to access the navigation bar.
Therefore, it is recommended to increase the touch-responsive area of the buttons in the navigation
bar.
Considering the extended amount of content that needs to be accessed through the SSK, the product
grid on the PLP and the UGC on the PDP should be visualized in such a way that allows bottom-row
items to be visible on the eye level. For example, this could be achieved by adding white space on the
bottom row for the PLP. Another way to design for this issue is to loop the user-generated images so
that when the user scrolls through them vertically the last picture is followed by the first one.
Prototype B was designed with a re-sized PDP, where the top image was enlarged in hopes of encour-
aging use. User test participants did not like this change as it resulted in seeing the most important
buttons in a lower position on the screen. Participants felt uncomfortable with the new size of the
images and expressed that the experience ”feels too public”. An addition to the size-related design
guideline is therefore added with the suggestion of defining a maximum size for all screen elements.
Interactive components should not be placed over the height of the user.
Chapter 8
Conclusion
This chapter presents the conclusions of this thesis work along with a short personal reflection, the
limitations of the study, and the possibilities of future work.
8.1 Summary of Thesis Achievements
This thesis work aimed to define unified guidelines for the self-service kiosk to answer the research
question How can the SSK increase customer engagement in an omnichannel retail shopping envi-
ronment in the Adidas Flagship store? And how can we design to achieve this engagement?. The
literature review revealed a gap in defined guidelines for self-service technology in a retail omnichan-
nel ecosystem. The literature also revealed that there is no convention when it comes to defining
guidelines, only popular approaches dependent on the case study. When it comes to methodology, an
interpretivist stance was taken with qualitative data collection methods.
The research process was composed of the context analysis realized through a heuristic evaluation and
an observation method, the future casting workshop with 7 participants, and a user testing process
with 16 participants. 10 design guidelines were defined some of which were based on already existing
guidelines with specifications for the retail context. The relation between the methods can be seen in
figure 8.1.
This work sets up a baseline for further research, evaluation, and design process that follows the
iterations of the SSK design. By laying this groundwork, the hope is to create a strong framework for
the continued development and improvement of the SSK and the wider digital retail ecosystem. In
97
98 Chapter 8. Conclusion
Figure 8.1: The revised process of the research
addition, this research also identified emerging trends that could be leveraged to further enhance the
SSK and the overall digital retail experience. Specifically, there is a focus on understanding the future
expectations of younger users within the target group toward the store’s touchpoints. By identifying
these trends and taking them into account, it is possible to create more user-centered and impactful
design guidelines for the SSK and other digital retail touchpoints.
This work was realized in collaboration with the internship host company, with the help of the digital
retail team who provided their previous research for analysis. The findings were presented to the
team at the end of the project in form of design suggestions, some of which are included in the
design guidelines. The results of the field research filled the gap in the user expectations towards the
omnichannel ecosystem and reassured the team about the proposed and tested design guidelines as
they were overall well received by the research participants.
As a personal reflection, working in a corporate environment on retail topics has shown me how
academic work can be translated into practical applications. It has also highlighted the di↵erences
between academic and corporate research methods. While academic research is often structured and
follows a linear process, the corporate world can be more non-linear and exploratory, with limited time
for research activities. Nevertheless, I found it fascinating to see how the methods and frameworks
learned at university can be adapted and applied in a practical setting, and how they can be used to
8.2. Limitations 99
inform and improve business decisions.
During my internship, I observed the ways of working of the digital retail team. In the design pro-
cess, decision-making relies on the designer’s intuition and past experience rather than testing each
hypothesis assumed during the procedure. In my opinion, utilizing qualitative methods in my research
is reflective of the real-life scenario of how the team operates. many design forms rely on user-centered
data, thereby emphasizing the importance of incorporating such methodologies in research practices.
8.2 Limitations
The findings of this study should be interpreted with caution due to multiple factors. This thesis work
is limited in its degree of accuracy, applicability, and objectiveness. Since the research work is deeply
rooted in interpretivism and qualitative methodology, there is a probability that biases have been
introduced into the study. As the interpretation bias implies, the data in this work was interpreted
subjectively by the researcher, who can have their own preconceived notions that influence the way
the data is interpreted. The small sample size used along with the non-probabilistic sampling method
introduces the sampling bias and therefore the generalizability of the findings is limited (Silverman,
2019). Moreover, the observation bias must be mentioned. Since the methodology involves observation
and interaction with participants, the researcher might unconsciously influence the behavior of the
participant or interpret their behavior in a way that confirms their expectations. As participants
were aware that they were being observed, this might have also influenced their behavior. Lastly, the
open-ended nature of the interviews with shopping assistants might have influenced the results based
on the improvised character of the follow-up questions, as the researcher could not anticipate the flow
of the conversation.
8.3 Future Work
To expand on this research, a thorough evaluation of the proposed design guidelines should be con-
ducted. Further design- and test iterations could be realized in order to test the limitations of the
guidelines. Regarding the user test subjects, a di↵erent larger user group could be recruited with
probabilistic sampling, ensuring that the targeted user group is fairly represented. In this case, the
automation of the usability tests could be done using testing tools such as UserTesing or UserZoom.
100 Chapter 8. Conclusion
In the next iteration, the beta version (represented briefly by prototype B) could be examined and
iterated. Additional sensors could be added such as an RFID reader, a camera, or a scanner. This
would enable the addition of new functionalities such as self-checkout, return handling of products,
and AR try-on. These functions could be evaluated with the help of the research question ”How can
additional functions increase the perceived value of the self-service kiosk and enhance the customer’s
shopping experience?” Further connectivity options could be examined in the retail ecosystem using
the SSK in relation to other touchpoints such as a secondary SSK or the digital footwear wall.
Before implementing these features, it is important to conduct user testing to determine whether
it meets the needs and desires of users. This can help to identify any potential issues or areas for
improvement, ensuring that the feature is as e↵ective and user-friendly as possible. In addition to
user testing, it is also essential to conduct research to assess the impact of this feature on key business
metrics such as sales, revenue, and customer loyalty. This can help to ensure that the implementation
of the feature will result in a positive return on investment and contribute to the overall success of
the business.
The study could be conducted in a di↵erent market (preferably outside of the EU) evaluating the
proposed guidelines to examine if they are applicable in di↵erent cultural contexts.
A similar study to this thesis work could be done based on the same methodology but with the fitting
room’s smart mirror as its main technology. The smart mirror is also a crucial part of the omnichannel
ecosystem in the retail industry. The findings point toward the potential of the fitting room and how
there is a possibility for it to become a central touchpoint in the shopping journey. Users found the
privacy of the fitting room to be suitable for a body scan, product trials, and even checkout, as seen
in the workshop findings.
Another future work area could extend the design guidelines into other industries that might not use
self-service technology at the moment. For example, a similar study could be conducted in the field
of government (Chen et al., 2021) or education, as the SST is not yet popular in these fields.
Bibliography
Adidas, 2018, 2.
J. C. Soares, R. Limongi, J. H. De Sousa Júnior, W. S. Santos, M. Raasch and L. Hoeckesfeld, Journal
of Marketing Analytics, 2022, 1–13.
A. D’Innocenzio, 2021.
F. E. Saputra and S. A. Praningrum, Planning, 2022, 17, 1937–1942.
P. A. Dabholkar and R. P. Bagozzi, Journal of the academy of marketing science, 2002, 30, 184–201.
M. L. Meuter, A. L. Ostrom, R. I. Roundtree and M. J. Bitner, Journal of marketing, 2000, 64, 50–64.
J. Lazar, J. B. Jordan and G. Vanderheiden, Interactions, 2019, 26, 74–77.
H. R. Yen, The Service Industries Journal, 2005, 25, 641–659.
A. Lindgreen, C. A. Di Benedetto and C. Kanitz, Omni-channel retailing, Springer, 2021, pp. 1–8.
J. Randolph, Practical assessment, research, and evaluation, 2009, 14, 13.
B. Glaser and A. Strauss, The Discovery of Grounded Theory: Strategies for Qualitative Research,
Aldine Transaction, 1967.
I. Coyne and S. Cowley, Journal of Research in Nursing, 2006, 11, 501–515.
J. W. Creswell, Qualitative inquiry and research design: Choosing among five traditions, SAGE Pub-
lications, Incorporated, 1998.
I. McMenamin, PS: Political Science & Politics, 2006, 39, 133–135.
B. G. Glaser, Theoretical sensitivity, University of California,, 1978.
J. Corbin and A. Strauss, Sage: Thousand Oaks, CA, USA, 2015.
101
102 BIBLIOGRAPHY
K. Charmaz, Constructing grounded theory, sage, 2014.
R. Thornberg and C. Dunne, The Sage handbook of current developments in grounded theory, 2019,
206–221.
J. F. Wolfswinkel, E. Furtmueller and C. P. Wilderom, European journal of information systems, 2013,
22, 45–55.
T. MFG, A brief history of the Kiosk, 2019, https://www.technikmfg.com/
a-brief-history-of-the-kiosk/.
M. L. Kasavana, Hospitality Upgrade, 2008, 122–128.
H. Kaur and S. Malhotra, 2018 5th International Symposium on Emerging Trends and Technologies
in Libraries and Information Services (ETTLIS), 2018, pp. 269–271.
M. Hampshire and C. E. Sanford, New library world, 2009.
S. Park, X. Lehto and M. Lehto, International Journal of Hospitality Management, 2021, 92, 102757.
R. Law, D. Leung and I. C. C. Chan, International Journal of Contemporary Hospitality Management,
2019, 32, 511–534.
Y. Lyu, C. J. Vincent, Y. Chen, Y. Shi, Y. Tang, W. Wang, W. Liu, S. Zhang, K. Fang and J. Ding,
Applied ergonomics, 2015, 47, 157–169.
E. Gökgür, M.Sc. thesis, Fen Bilimleri Enstitüsü, 2020.
Y. B. Sirotin, J. A. Hasselgren and A. Vemury, Proceedings of the Human Factors and Ergonomics
Society Annual Meeting, 2016, pp. 2019–2023.
S.-M. Wang and C. Han, International Conference on Human-Computer Interaction, 2022, pp. 337–
349.
S. G. Abdelaziz, A. A. Hegazy and A. Elabbassy, International Journal of Computer Science Issues
(IJCSI), 2010, 7, 30.
H.-L. Chang and C.-H. Yang, Tourism Management, 2008, 29, 980–993.
M. A. Aziz, N. S. Lop, I. F. M. Kamar, L. Abdullah and N. M. Akhir, Int. J. Acad. Res. Bus. Soc.
Sci., 2020, 10,.

BIBLIOGRAPHY 103
M. B. Garcia, 2019 IEEE 11th International Conference on Humanoid, Nanotechnology, Information
Technology, Communication and Control, Environment, and Management (HNICEM), 2019, pp.
1–6.
J. M. Curran and M. L. Meuter, Journal of services marketing, 2005.
R. N. Bolton, J. R. McColl-Kennedy, L. Cheung, A. Gallan, C. Orsingher, L. Witell and M. Zaki,
Journal of service management, 2018, 29, 776–808.
P. Kral, K. Janoskova and A.-M. Potcovaru, Linguistic and Philosophical Investigations, 2022, 21,
252–267.
H.-J. Lee, A. Fairhurst and H. J. Cho, The Service Industries Journal, 2013, 33, 248–265.
A. Günay, Ç. Erbuğ, P. Hekkert and N. R. Herrera, in Human-Computer Interfaces and Interactivity:
Emergent Research and Applications, IGI Global, 2014, pp. 14–32.
J. Gill, ISBN, 1997.
A. Veijalainen, 2017.
E. E. Mattheiss, J. Schrammel and M. Tscheligi, Symposium of the Austrian HCI and Usability
Engineering Group, 2011, pp. 595–606.
H. Chung and W. Park, arXiv preprint arXiv:2105.07445, 2021.
E. C. Hayashi, E. M. Menezes, R. E. Duarte, F. de Souza, C. M. Ogushi and S. H. M. Duarte,
International Conference on Information Society (i-Society 2014), 2014, pp. 1–6.
M. C. Maguire, International journal of human-computer studies, 1999, 50, 263–286.
P. A. Dabholkar, International Journal of research in Marketing, 1996, 13, 29–51.
H.-J. Lee, A. E. Fairhurst and M.-Y. Lee, Managing Service Quality: An International Journal, 2009.
D. X. Ding, P. J.-H. Hu and O. R. L. Sheng, Journal of Business Research, 2011, 64, 508–515.
J.-S. C. Lin and P.-L. Hsieh, Journal of retailing, 2011, 87, 194–206.
J. J. Cronin Jr and S. A. Taylor, Journal of marketing, 1992, 56, 55–68.
A. Parasuraman, V. A. Zeithaml and L. Berry, 1988, 1988, 64, 12–40.

104 BIBLIOGRAPHY
P. Shamdasani, A. Mukherjee and N. Malhotra, The Service Industries Journal, 2008, 28, 117–138.
M. L. Meuter, A. L. Ostrom, M. J. Bitner and R. Roundtree, Journal of Business Research, 2003, 56,
899–906.
A. Singh, Ph.D. thesis, Arizona State University, 2018.
A. May, Master of Science thesis in Ergonomics, Faculty of Science, University College, London, 1993.
M. Maguire, Applied Ergonomics, 2014, 45, 162–170.
M. Maguire, SIGCAPH Comput. Phys. Handicap., 2001, 4–7.
K. Siebenhandl, G. Schreder, M. Smuc, E. Mayr and M. Nagl, IEEE Transactions on Professional
Communication, 2013, 56, 138–159.
I. Tala, M.Sc. thesis, 2016.
J. Nielsen, Conference companion on Human factors in computing systems, 1994, pp. 413–414.
J. Nielsen, Ten usability heuristics, 2005.
A. Asmare and S. Zewdie, The International Review of Retail, Distribution and Consumer Research,
2022, 32, 59–79.
M. Garaus and U. Wagner, Journal of Retailing and Consumer Services, 2019, 47, 331–338.
M. Frasquet-Deltoro, A. Molla-Descals and M.-J. Miquel-Romero, Journal of Brand Management,
2021, 28, 388–401.
B. Alexander and M. B. Cano, Exploring omnichannel retailing: Common expectations and diverse
realities, 2019, 197–223.
B. Alexander and M. B. Cano, Journal of Retailing and Consumer Services, 2020, 55, 101913.
J. Johnson, Designing with the mind in mind: simple guide to understanding user interface design
guidelines, Morgan Kaufmann, 2020.
T. Chen, W. Guo, X. Gao and Z. Liang, Government Information Quarterly, 2021, 38, 101520.
V. Kaptelinin, A. Rizzo, P. Robertson and S. Rosenbaum, in Proceedings of the 2014 companion
publication on Designing interactive systems, 2014, pp. 199–202.

BIBLIOGRAPHY 105
H. Demirkan and J. Spohrer, Journal of Retailing and Consumer Services, 2014, 21, 860–868.
P. Kelly, J. Lawlor and M. Mulvey, in Robots, artificial intelligence, and service automation in travel,
tourism and hospitality, Emerald Publishing Limited, 2019.
Y. Vakulenko, D. Hellström and P. Oghazi, International Journal of Retail & Distribution Manage-
ment, 2018, 46, 507–527.
A.-L. Syrjänen, V. Sihvola, K. Kuutti and R. Vilmunen, Proceedings of the 7th Nordic Conference on
Human-Computer Interaction: Making Sense Through Design, 2012, pp. 288–297.
K. McKenzie, N. Lowres, J. Orchard, C. Hespe, B. Freedman and K. Giskes, Cardiovascular Digital
Health Journal, 2022, 3, 212–219.
S. Galdolage, 2020.
T. Cinto, in Handbook of research on human-computer interfaces, developments, and applications, IGI
Global, 2016, pp. 525–544.
P. Jones, D. Comfort, C. Clarke-Hill and D. Hillier, Marketing Intelligence & Planning, 2010, 28,
241–248.
E. Nica, M. Poliak, G. H. Popescu and I.-A. Pârvu, Linguistic and Philosophical Investigations, 2022,
21, 137–153.
U.S. Food and Drug Administration, Radio Frequency Identification (RFID), https:
//www.fda.gov/radiation-emitting-products/electromagnetic-compatibility-emc/
radio-frequency-identification-rfid.
J. Lazar, J. H. Feng and H. Hochheiser, Research methods in human-computer interaction, Morgan
Kaufmann, 2017.
S. Card, T. Moran and A. Newell, Applied Information-Processing Psychology: The, 1983.
K. Nygaard and O. T. Bergo, Personnel review, 1975.
G. Paré, M.-C. Trudel, M. Jaana and S. Kitsiou, Information & Management, 2015, 52, 183–199.
R. Weber, MIS quarterly, 2004, iii–xii.
S. Harrison, D. Tatar and P. Sengers, Alt. Chi. Session at the SIGCHI Conference on human factors
in computing systems San Jose, California, USA, 2007, pp. 1–18.

106 BIBLIOGRAPHY
N. McDonald, S. Schoenebeck and A. Forte, Proceedings of the ACM on human-computer interaction,
2019, 3, 1–23.
P. Dourish, Where the action is: the foundations of embodied interaction, MIT press, 2004.
J. Nielsen, Usability engineering, Morgan Kaufmann, 1994.
J. Lazar and A. Norcio, in Handbook of human-computer interaction, Springer, 2006, pp. 917–941.
J. Kirakowski, International Journal of Human-Computer Studies, 2002, 57, 295–313.
Y. Lincoln and E. Guba, Naturalistic inquiry, Sage publications, 1985.
C. Thomas and N. Bevan, 1996.
J. Pruitt and T. Adlin, The persona lifecycle: keeping people in mind throughout product design,
Elsevier, 2010.
L. de Greef, A. Goguey, A. Boucher and P. Leproux, Journal of Usability Studies, 2016, 11, 1–17.
A. Tang and W. Liu, Interactions, 2017, 24, 44–49.
C. Morrison and G. Bell, Healthcare Quarterly, 2010, 14, 64–70.
T. B. Futurist, Futurecasting, https://www.baldfuturist.com/futurecasting/, 2015.
R. A. Virzi, Human factors, 1992, 34, 457–468.
J. Nielsen, Usability engineering, Morgan Kaufmann, 1994.
J. S. Dumas, J. S. Dumas and J. Redish, A practical guide to usability testing, Intellect books, 1999.
J. Brooke et al., Usability evaluation in industry, 1996, 189, 4–7.
D. Bertram, Retrieved November, 2007, 2, 1–10.
MeasuringU, How to Interpret the System Usability Scale (SUS) Score, 2016, https://measuringu.
com/interpret-sus-score/.
A. Bangor, P. T. Kortum and J. T. Miller, Intl. Journal of Human–Computer Interaction, 2008, 24,
574–594.
BIBLIOGRAPHY 107
B. Laugwitz, T. Held and M. Schrepp, HCI and Usability for Education and Work: 4th Symposium of
the Workgroup Human-Computer Interaction and Usability Engineering of the Austrian Computer
Society, USAB 2008, Graz, Austria, November 20-21, 2008. Proceedings 4, 2008, pp. 63–76.
M. Schrepp, A. Hinderks and J. Thomaschewski, Design, User Experience, and Usability. Theories,
Methods, and Tools for Designing the User Experience: Third International Conference, DUXU
2014, Held as Part of HCI International 2014, Heraklion, Crete, Greece, June 22-27, 2014, Proceed-
ings, Part I 3, 2014, pp. 383–392.
G. Guest, K. M. MacQueen and E. E. Namey, Applied thematic analysis, sage publications, 2011.
J. Martı́nez-Mesa, D. A. González-Chica, R. P. Duquia, R. R. Bonamigo and J. L. Bastos, Anais
brasileiros de dermatologia, 2016, 91, 326–330.
B. D. Johnson, Synthesis Lectures on Computer Science, 2011, 3, 1–190.
Google LLC, Tilt Brush, 2021, https://www.tiltbrush.com/.
LamasaTech, Lamasatech Blog, 2021.
M. Rouse, TechTarget, 2021.
OpenAI, DALL·E: Creating Images from Text, https://openai.com/research/dall-e/, 2021.
I. M. Hub, Social Commerce Trends: The Ultimate Guide, https://influencermarketinghub.com/
social-commerce-trends/, 2021.
E. H. Glaessgen and D. S. Stargel, Nature, 2012, 481, 265–268.
J. Grudin, Proceedings of the SIGCHI conference on Human factors in computing systems, 1989, pp.
397–404.
J. Smith and J. Lee, Journal of Retailing, 2022, 34, 45–58.
L. Li and H. Takagi, International Journal of Human-Computer Studies, 2010, 68, 277–293.
S. H. Cheraghi, L. Alem and A. Kazemi, Journal of Theoretical and Applied Electronic Commerce
Research, 2016, 11, 32–49.
V. Kasinath, P. A. Dabholkar, M. Bengtsson and B. Edvardsson, Journal of Retailing and Consumer
Services, 2018, 40, 206–218.

108 BIBLIOGRAPHY
P. C. Verhoef, S. A. Neslin and B. Vroomen, International journal of research in marketing, 2007, 24,
129–148.
D. A. Norman, The design of everyday things: Revised and expanded edition, Basic Books, Revised
and expanded edition edn., 2013.
M. Fuentes-Blasco, B. Moliner-Velázquez, D. Servera-Francés and I. Gil-Saura, Journal of product &
brand management, 2017.
T. Wang, X. Li, X. Li and Q. Liu, Journal of Retailing and Consumer Services, 2019, 51, 340–346.
D. Silverman, Interpreting Qualitative Data, 2019, 1–568.

Appendix A
Literature Review Process
109
110 Appendix A. Literature Review Process
Used citations: 1 - 5, 7, 11, 13 - 15, 17, 18, 22 - 25, 27 - 37, 39, 41, 47 - 49, 51 - 53, 55, 56, 58, 65, 68
- 70, 72, 74, 76, 78 - 80.

111
112 Appendix A. Literature Review Process
Appendix B
Context Analysis Template
1. Observation context - location and time: when and where was the observation performed?
By whom?
2. Product
(a) Name
(b) Description and purpose
(c) Field of application
(d) Main functionalities
3. Specification
(a) Hardware - desk research
(b) Software specifications
4. Maguire, 1999 criteria
(a) Location
(b) Encouraging use
(c) Physical access
(d) Introduction and instruction
(e) Language selection
(f) Privacy
113
114 Appendix B. Context Analysis Template
5. Ergonomics
(a) Positioning
(b) Tilt
(c) Interaction
6. User types
7. Observing tasks performance
(a) How do people use the machine?
(b) How do they react?
8. Heuristics - self assessment
(a) Visibility of system status
(b) Match between system and real world
(c) User control and freedom
(d) Consistency and standards
(e) Error prevention
(f) Recognition rather than recall
(g) Flexibility and efficiency to use
(h) Aesthetic and minimalist design
(i) Recovery from errors
(j) Help and documentation
9. Organizational environment
Nielsen’s 10 rules of thumb - a heuristic evaluation
1. Visibility of system status
The system should always keep users informed about what is going on, through appropriate feedback
within reasonable time.
2. Match between system and the real world

115
The system should speak the users’ language, with words, phrases and concepts familiar to the user,
rather than system-oriented terms. Follow real-world conventions, making information appear in a
natural and logical order.
3. User control and freedom
Users often choose system functions by mistake and will need a clearly marked ”emergency exit” to
leave the unwanted state without having to go through an extended dialogue. Support undo and redo.
4. Consistency and standards
Users should not have to wonder whether di↵erent words, situations, or actions mean the same thing.
Follow platform conventions.
5. Error prevention
Even better than good error messages is a careful design which prevents a problem from occurring
in the first place. Either eliminate error-prone conditions or check for them and present users with a
confirmation option before they commit to the action.
6. Recognition rather than recall
Minimize the user’s memory load by making objects, actions, and options visible. The user should
not have to remember information from one part of the dialogue to another. Instructions for use of
the system should be visible or easily retrievable whenever appropriate.
7. Flexibility and efficiency of use
Accelerators – unseen by the novice user – may often speed up the interaction for the expert user such
that the system can cater to both inexperienced and experienced users. Allow users to tailor frequent
actions.
8. Aesthetic and minimalist design
Dialogues should not contain information which is irrelevant or rarely needed. Every extra unit of
information in a dialogue competes with the relevant units of information and diminishes their relative
visibility.
9. Help users recognize, diagnose, and recover from errors

116 Appendix B. Context Analysis Template
Error messages should be expressed in plain language (no codes), precisely indicate the problem, and
constructively suggest a solution.
10. Help and documentation
Even though it is better if the system can be used without documentation, it may be necessary to
provide help and documentation. Any such information should be easy to search, focused on the user’s
task, list concrete steps to be carried out, and not be too large
Appendix C
Workshop Script
Itinerary
• 10:20 - 10:30 Prepare screen and snacks + glasses
• 10:31 - 10:36 Getting people from the elevator
• 10:37 Title slide + welcome
• 10:38 - 10:39 Agenda
• 10:40 - 10:45 Icebreaker
• 10:46 - 10:58 Present the topic and the trends + discuss
• 10:59 - 11:00 Divide in groups + give them the personas
• 11:01 - 11:03 Read and understand the persona
• 11:04 - 11:15 Future casting
• 11:11 - 11:15 Break - game “ha” if we have the time
• 11:15 - 11:25 SF prototyping [5 minutes/ 2 minutes/ 3 minutes]
• 11:26 - 11:39 Back casting
• 11:40 - 12:00 Prepare for presentation and present
117
118 Appendix C. Workshop Script
Script
1. Intro + protocol: thanks, and how this info is going to be used + schedule
2. Icebreaker: favourite show with futuristic technology in it For our icebreaker, take a second and
think about your favourite TV show/ film that has futuristic technology in it.
3. Future casting flow:
(a) Expert Testimony
i. A review of social trends (the current and emerging habits, practices, needs and wants
of people), all major technological trends relevant to the domain, current business
trends, and any relevant ecosystem trends.
ii. Social trends - emerging habits, practices, needs and wants
iii. Technological Trends - business trends
iv. Ecosystem trends - regulation by government
(b) Synthesis How do you think this a↵ects the company, your department, your team?/ What
would be a good response action? Secondly, how does this a↵ect you as a user?
(c) Personas
(d) Rapid future casting Imagine that you are from 2047 (25 years from now) and you have to
explain to me (who is stuck in the present) how the in-store shopping experience looks like
in your time in a store.
i. What kind of digital signage is in the store?
ii. Are there new ones?
iii. Did the ones existing in 2022 evolve? How?
iv. Did the store’s layout change? Or the logistics?
v. What kind of interactions are happening?
vi. Briefly describe the main device from your vision with the most important feature and
interaction.
vii. COME UP WITH 8 IDEAS AS A GROUP, YOU HAVE 10 minutes
(e) Science fiction prototyping It is now 5 years later, 2052. Your favorite idea has been
developed and it is live in the store. Describe in a couple sentences how does this experience
look like. Feel free to draw.

119
i. What is your chosen digital signage? How does it look like?
ii. Where is it situated?
iii. Where is the customer situated?
iv. What kind of interaction is happening?
v. What problem is being solved?
(f) Back casting In a timeline write down how do you imagine we get to your FINAL IDEA?
i. What needs to happen in 20 years?
ii. What needs to happen in 10 years?
iii. What needs to happen in 5 years?
iv. What needs to happen tomorrow? What do we need to start working on?
(g) Presentations
(h) Thanks
Appendix D
Personas
120
121
122 Appendix D. Personas
Appendix E
Completed Artifacts
123
124 Appendix E. Completed Artifacts
125
126 Appendix E. Completed Artifacts
Appendix F
User Testing Script
1. Protocol Thank you for agreeing to help me out with this user testing. It is very valuable for
me for my thesis and the digital retail team’s research. We are going to take a look at the
SSK today, which is a touch point located in the store. I am going to ask you questions and
present you with 2 tasks that you will perform with 2 di↵erent versions of prototypes. This
means that not every function is going to be working and some interactions will not align with
your expectations.
I want to emphasise that we are testing the prototype and not you in this process, there are no
wrong answers. I am looking for your honest opinion. During the task performance, I would like
to ask you for your permission to record your voice.
2. Intro Questions
(a) What is your name?
(b) Could you tell me your age, if you are comfortable sharing this information?
(c) In what capacity do you work at the company?
(d) Have you ever interacted with Self -service Technology before? (If yes to previous question)
Where did you encounter it?
(e) When it comes to buying adidas products (or similar products) what is your preferred
purchasing channel?
(f) What makes you choose this method of shopping?
(g) Why?
127
128 Appendix F. User Testing Script
(h) Just by seeing the SSK here, what are your expectations towards the test?
3. Presenting Scenarios Now I am going to present you with a task and I will ask you to perform
it using this screen. I will ask you to think out loud, and tell me your tough process and your
expectations. Ideally I will not help you perform the task, but I am here to answer your questions
if necessary.
Imagine you are out in the city shopping and you just entered the adidas store. You already
know what kind of items you are looking for. Imagine that the environment that you are in is
like the store: there are many di↵erent types of products on display, shopping assistants greeting
you, music playing, and other shoppers are around.
4. Steps for both prototypes
(a) Task 1 : Clothing item browse You saw someone wearing a dark red T-shirt, with white
stripes on the side of the sleeve, and with a big Originals logo in the middle, that you really
liked. You want to see if it is available in the store, and buy it. if yes you want to try it
on, if not you want to add it to your online shopping list In this scenario you are looking
for a women’s shirt, size 12.
(b) Task 2 : Shoe search You are looking for the shoes that are called vegan superstar shoes
that you want to try on. In this scenario you are looking for a women’s shoe, size 8.
(c) SUS I will read out 10 statements now, and I want you to tell me how much you agree with
each statement regarding your experience with the SSK on a scale from 1 to 5, where 1 is
I completely disagree and 5 is I completely agree. 3 is the neutral answer (see template
below).
(d) UEQ Now you see a table on the page where each row has 2 attributes that are the opposite
of each other. Think about your experience from before and rate your experience from 1
to 7, where 1 is closest to the attribute on the left and 7 is closest to the attribute on the
right. 4 is the neutral answer (see template below).
i. annoying/ enjoyable
ii. not understandable/ understandable
iii. creative/ dull
iv. easy to learn/ difficult to learn
v. valuable/ inferior
129
vi. boring/ exciting
vii. not interesting/ interesting
viii. unpredictable/ predictable
ix. fast/ slow
x. inventive/ conventional
xi. obstructive/ supportive
xii. good/ bad
xiii. complicated/ easy
xiv. unlikable/ pleasing
xv. usual/ leading edge
xvi. unpleasant/ pleasant
xvii. secure/ not secure
xviii. motivating/ demotivating
xix. meets expectations/ doesn’t
xx. inefficient/ efficient
xxi. clear/ confusing
xxii. impractical/ practical
xxiii. organised/ cluttered
xxiv. attractive/ unattractive
xxv. friendly/ unfriendly
xxvi. conservative/ innovative
(e) Open ended questions
i. How would you describe your experience with this prototype?
ii. What was the most enjoyable part of this interaction?
iii. What was the least enjoyable part of this interaction?
iv. What would you change about this prototype?
5. Comparison questions
(a) Which experience was better and why?
(b) Is there anything else you would like to say/ suggest/ remark?
130 Appendix F. User Testing Script
(c) How did your expectations change after interacting with the prototypes?
6. Thanks
Figure F.1: The SUS questionnaire

131
Figure F.2: The UEQ template

TFM Timea Udvarhelyi-Unlocked

Uploaded by

Copyright:

Available Formats

TFM Timea Udvarhelyi-Unlocked

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

TFM Timea Udvarhelyi-Unlocked

Uploaded by

Copyright:

Available Formats

Universidad Politécnica

Master in Human Computer

Self-Service Kiosk Design Guidelines for a

Author: Timea Udvarhelyi

Trento, March, 2023

Title: Self-Service Kiosk Design Guidelines for a Digital Retail Ecosystem

Author: Timea Udvarhelyi

Luca Turchet, UniTrento Ricardo Imbert, UPM

BITM Bring It To Me (function)

Table 1: List of abbreviations used in the research work

1.1 The Host Company . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 The Problem Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 The Self-Service Kiosk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.4 Purpose and Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.5 The Research Question . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.6 The Proposed Research Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.7 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.3.2 Search, Select, and Analyse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.4.1 A Brief History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.4.2 Fields of Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.4.4 Accessibility and Inclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.4.5 Quality Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.4.6 Self Service Technology Design Guidelines . . . . . . . . . . . . . . . . . . . . . 18

2.4.7 Digital Retail Ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3 Design and Redesign 24

3.1 Released Version . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

3.2 Current Retail Ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

3.3 Beta Version . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.1 Methodological Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.2.1 Context Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.2.3 Usability Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

4.2.4 System Usability Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.2.5 User Experience Questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4.2.6 E↵ectiveness and Efficiency Measurement . . . . . . . . . . . . . . . . . . . . . 49

4.2.7 Thematic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

5.1 Context Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

5.1.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

5.1.2 Data Collection Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

5.2.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5.2.2 Data Collection Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

5.3 User Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

5.3.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

5.3.2 Data Collection Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

6.1 Context Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

6.2.1 Idea Presentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

6.2.2 Emerging Themes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

6.3 User Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

6.3.1 Introductory Question Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

6.3.2 System Usability Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

6.3.3 User Experience Questionnaire . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

6.3.4 E↵ectiveness and Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

6.3.5 Open Ended Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

6.4 Summary of Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

6.4.1 Shopping Behaviours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

6.4.3 The Store Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

6.4.5 Connecting to the Store Ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . 83

7.1 Encouragement of use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

7.4 Informative Progress Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

7.5 Personalized Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90