The HORM Diagramming Tool: A Domain-Specific Modelling Tool for
SME Cybersecurity Awareness
Costas Boletsis
1 a
, Sefat Noor Orni
2
and Ragnhild Halvorsrud
1 b
1
SINTEF Digital, Oslo, Norway
2
Department of Informatics, University of Oslo, Oslo, Norway
Keywords:
CJML, Cybersecurity, Modelling, User Journey, Visualisation.
Abstract:
Improving security posture while addressing human errors made by employees are among the most chal-
lenging tasks for SMEs concerning cybersecurity risk management. To facilitate these measures, a domain-
specific modelling tool for visualising cybersecurity-related user journeys, called the HORM Diagramming
Tool (HORM-DT), is introduced. By visualising SMEs’ cybersecurity practices, HORM-DT aims to raise
their cybersecurity awareness by highlighting the related gaps, thereby ultimately informing new or updated
cyber-risk strategies. HORM-DT’s target group consists of SMEs’ employees with various areas of technical
expertise and different backgrounds. The tool was developed as part of the Human and Organisational Risk
Modelling (HORM) framework, and the underlying formalism is based on the Customer Journey Modelling
Language (CJML) as extended by elements of the CORAS language to cover cybersecurity-related user jour-
neys. HORM-DT is a fork of the open-source Diagrams.net software, which was modified to facilitate the
creation of cybersecurity-related diagrams. To evaluate the tool, a usability study following a within-subject
design was conducted with 29 participants. HORM-DT achieved a satisfactory system usability scale score of
80.69, and no statistically significant differences were found between participants with diverse diagramming
tool experience. The tool’s usability was also praised by participants, although there were negative comments
regarding its functionality of connecting elements with lines.
1 INTRODUCTION
Small and medium enterprises (SMEs) are increas-
ingly being exposed to cyber-risks. In 2017 alone,
61% of SMEs were exposed to malware cyberattacks
(Khan et al., 2020). Despite their increasing expo-
sure to cyberattacks, SMEs rarely conduct thorough
cyber-risk assessments, and more than half of them
provide either an outdated or no cyber-risk strategy at
all (Benz and Chatterjee, 2020; Paulsen, 2016; The
National Center for the Middle Market, 2016; Bo-
letsis et al., 2021). This may be due to internal is-
sues that arise when attempting to set up cyber-risk
strategies, such as ‘having small IT teams, inade-
quate security budgets, and disagreements between
IT and business leadership teams regarding cyberse-
curity risk management’ (Boletsis et al., 2021). It has
been reported that the most challenging tasks for cy-
bersecurity risk management in SMEs are i) defining
a
https://orcid.org/0000-0003-2741-8127
b
https://orcid.org/0000-0002-3774-4287
and taking the first step towards improving their secu-
rity posture and ii) addressing human errors made by
employees, which is the ‘human element’ sometimes
referred to as the biggest internal threat for SMEs
(Arctic Wolf, 2017; Meshkat et al., 2020; Symantec,
2019; Boletsis et al., 2021).
To address the first challenge, it has been sug-
gested that mapping SMEs’ current practices and the
potential threats they face would be a useful ini-
tial move in cybersecurity risk management (Benz
and Chatterjee, 2020; Paulsen, 2016; Meszaros and
Buchalcevova, 2017). This mapping should also ad-
dress the second big challenge by modelling the hu-
man element in cybersecurity-related scenarios, es-
sentially through the documentation of user journeys
related to cybersecurity, which are the visual paths
that users may take when performing an action or us-
ing a service (Stickdorn et al., 2018; Boletsis et al.,
2021). At the same time, end-users/employees should
be involved in the modelling process, always taking
into consideration their diverse technical backgrounds
(Paulsen, 2016; Kullman et al., 2020; Bellamy et al.,
Boletsis, C., Orni, S. and Halvorsrud, R.
The HORM Diagramming Tool: A Domain-Specific Modelling Tool for SME Cybersecurity Awareness.
DOI: 10.5220/0011786600003417
In Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 3: IVAPP, pages
203-213
ISBN: 978-989-758-634-7; ISSN: 2184-4321
Copyright
c
2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
203
2007; Boletsis et al., 2021).
This paper presents, the HORM Diagramming
Tool (HORM-DT)
1
, a domain-specific modelling tool
for visualising cybersecurity-related user journeys,
and evaluates its usability. The tool’s main target
group consists of SME employees with various areas
of technical expertise and different backgrounds, and
so usability is of the essence. By using the tool and
the models produced by it, the aim is to raise cyber-
security awareness so that SMEs can visualise their
practices and the potential cyber threats they face and
identify the gaps. In this way, they can develop up-to-
date or new cyber-risk strategies in an informed way.
The tool was developed as part of the Human and
Organizational Risk Modelling (HORM) framework
2
to be a comprehensible and easy-to-use cybersecurity-
related framework for capturing the risks that ordi-
nary people may be exposed to (Fair et al., 2022).
HORM is based on the Customer Journey Mod-
elling Language (CJML) (Halvorsrud et al., 2021) and
extended by formalism from the CORAS language
(Lund et al., 2011; Vraalsen et al., 2005) so that it
can address cybersecurity-related user journeys.
2 RELATED WORK
When conducting cybersecurity assessments, sim-
ply providing accurate risk information may not be
enough to ensure that individuals will be able to com-
prehend a risk message and act on it (Nurse et al.,
2011; Slovic, 1999; Skubisz et al., 2009; Boletsis
et al., 2021). However, visualisations can be an im-
portant factor in communicating cyber risks (Bolet-
sis et al., 2021). In the field of SME cybersecurity
awareness, a limited number of works have evaluated
SMEs’ cybersecurity practices by utilising visual ele-
ments (Boletsis et al., 2021), including the following:
CYSEC, a do-it-yourself cybersecurity assess-
ment method for SMEs that automates elements
of a counselling dialogue between a security ex-
pert and employees in the SME to counter cyber
threats (Shojaifar, 2019).
The SME Cybersecurity Evaluation Tool, which
consists of a 35-question online survey to be com-
pleted by IT leaders to self-rate their maturity
based on the National Institute of Standards and
Technology cybersecurity framework (Benz and
Chatterjee, 2020).
1
HORM-DT can be accessed at: https://cjml.no/horm2/
and its code is available from the Github repository at:
https://github.com/CostasBoletsis/HORM-DT
2
HORM website: https://cjml.no/horm/
The System Security Modeller (SSM), which is an
asset-based risk-analysis tool that provides an in-
formation security perspective to the interactions
between assets across an entire system and can be
also applied in SME settings (Boletsis et al., 2021;
Surridge et al., 2019).
A gamified approach for raising awareness sur-
rounding SMEs’ level of cybersecurity and re-
silience (Ponsard and Grandclaudon, 2019), for
which SME employees answer a cybersecurity
quiz and a self-assessment questionnaire.
It can be said that these approaches for mapping
SMEs’ cybersecurity practices focus more on repre-
sentations of infrastructure than on human actors and
their behaviours. At the same time, ‘there is no holis-
tic visualisation approach that could facilitate infor-
mation distribution between employees of different
levels of expertise’ (Boletsis et al., 2021).
Each SME process essentially represents a path-
way through a sequence of events. The modelling
and visualisation of these processes could be covered
by user journey modelling languages (Boletsis et al.,
2021). To that end, the HORM framework was devel-
oped (Fair et al., 2022).
The HORM framework focuses on the ‘human el-
ement’ and its modelling, and it consists of a mod-
elling language, based on a version of the CJML
(Haugstveit et al., 2016; Halvorsrud et al., 2016a;
Halvorsrud et al., 2014; Halvorsrud et al., 2016b) ex-
tended through contributions from the CORAS lan-
guage to fulfil cybersecurity-related purposes, and a
set of tools for modelling.
CJML is a visual language for the modelling and
visualisation of service and work processes in terms
of customer or user journeys (Halvorsrud et al., 2021;
Boletsis et al., 2021). Being centred on humans and
their activities, CJML appeals to a broad user group
through its simple and intuitive form (Halvorsrud
et al., 2016a), a feature that is attributed to the user-
centred design methodology that lead its design. The
basic units of CJML are observable touchpoints that
can take the form of a communication event or a non-
communicative activity or action. In CJML, the se-
quence of touchpoints that a user follows to achieve
a specific goal constitutes a user journey. There are
two types of diagram available in CJML that serve
different purposes (Figure 1). The simple journey di-
agram is suitable for journeys with few actors and
emphasises any deviation from the planned journey.
The swimlane diagram is useful for journeys involv-
ing several actors and emphasises both the initiator
and the recipient of a touchpoint (Halvorsrud et al.,
2021; Halvorsrud et al., 2016a).
CJML has been extended to model cybersecurity-
IVAPP 2023 - 14th International Conference on Information Visualization Theory and Applications
204
Figure 1: CJMLs communication model with a sender
transmitting a message to a receiver through a communi-
cation channel (upper part). The visual representation of
a touchpoint in the case of a journey diagram (left) and a
swimlane diagram (right) (Boletsis et al., 2021).
related user journeys using elements from the
CORAS language. CORAS is ‘a model-driven ap-
proach to risk analysis that consists of a method, a
language and a tool to support the risk analysis pro-
cess’ (Lund et al., 2011). The CORAS graphical lan-
guage, an extension of the Unified Modelling Lan-
guage (UML) 2.0 specification language, allows for
‘documentation of undesirable behaviour in the form
of threat scenarios’ (Vraalsen et al., 2005). In the
CORAS language, a threat is described as ‘using a
threat agent, e.g., a disloyal employee or a computer
virus. The threat agent initiates a threat scenario,
which is a sequence of events or activities leading
to an unwanted incident, i.e., an event resulting in a
reduction in the value of the target asset’ (Vraalsen
et al., 2005). These two new elements of threats and
unwanted incidents have been added in both CJML
formalism and its concrete syntax as icons (Figure
2). Moreover, the CJML users and ‘deviations’ were
redefined to fit the cybersecurity context, such that
the users are SME employees and external users that
make use of SME infrastructure under business-to-
business (B2B) offerings, and the ‘deviations’ are the
cyber threats (Boletsis et al., 2021).
Within the HORM cybersecurity-related context,
CJML has been applied using MS PowerPoint tem-
Figure 2: The icons of threat and unwanted incident that
were integrated from CORAS to CJML to extend CJML
and address cybersecurity-related user journeys.
plates as its tool
3
. Therefore, rather than relying on
third-party software, there is room and a need for the
development of a standalone modelling and diagram-
ming tool to act as a user interface (UI) for the ex-
tended CJML (i.e., HORM) modelling language (Fig-
ure 3). At the same time, UI usability is important,
and having a standalone, independent modelling tool
would enable developers to tweak the tool and its
source code until they reach satisfactory levels of UI
usability.
3 HORM DIAGRAMMING TOOL
3.1 Development
HORM-DT was developed using the open-source Di-
agrams.net (previously Draw.io) graph drawing soft-
ware by JGraph
4
(Figure 3), developed in HTML5
and JavaScript. The interface of the Diagrams.net
open-source software enables the user to easily create
and manage diagrams such as flowcharts, wireframes,
organisational charts and network diagrams
5
. Its main
features include the creation of a wide range of di-
agrams, the availability of templates for diagrams,
display relationships between objects using various
types of graph element (such as shapes, arrows, text,
images and icons), the import of images/icons from
external sources and the storage and export of cre-
ated diagrams in several image formats and other for-
mats, such as XML, PDF, HTML and more (JGraph,
2022). The Diagrams.net open-source software was
Figure 3: The UI and modelling formalism components that
were utilised for the creation of HORM-DT.
3
MS PowerPoint templates for the extended CJML:
https://cjml.no/horm/CJML Diagram generator v1.pptx
4
Code repository of the open-source Diagrams.net soft-
ware: https://github.com/jgraph/drawio
5
Diagrams.net website: https://diagrams.net/
The HORM Diagramming Tool: A Domain-Specific Modelling Tool for SME Cybersecurity Awareness
205
Figure 4: A screenshot of the HORM-DT UI with the tweaked interface areas, based on a heuristic evaluation, highlighted in
red. HORM-DT can be accessed at https://cjml.no/horm2/
considered and used because i) it is widely pop-
ular among software engineers (Rauer, 2019; Von
Der Assen et al., 2022) and ii) it was presented, anal-
ysed, and/or used in several studies regarding mod-
elling languages and tools (Von Der Assen et al.,
2022; Jaimez-Gonz
´
alez and Mart
´
ınez-Samora, 2020;
Ag
´
ardi, 2022; Vakaliuk et al., 2021; Ch
´
avez-Feria
et al., 2022; Ch
´
avez-Feria et al., 2021).
3.2 User Interface Design
Although the Diagrams.net software comes with an
already designed UI, a heuristic evaluation of it was of
the essence because HORM-DT should address users
of all technical backgrounds (Figure 3). The heuristic
evaluation was conducted in an informal manner by
the authors, who utilised the ten usability heuristics
for UI design (Nielsen, 2020). The UI design inher-
ited from the Diagrams.net software fulfilled almost
all heuristics; however, the following few tweaks were
made to further address the specific nature and do-
main of our tool:
Templates: Based on the CJML formalism, a user
journey modelling process should start by using a
template that contains the basic elements of the
concrete syntax. To that end, templates for swim-
lane and journey diagrams, as well as examples
of models, were added and made accessible right
from the landing page. At the same time, a dis-
tinctive ‘Templates’ menu button was added to the
left-hand side icon library (Figure 4) so that users
can access the templates at any point, thereby ad-
dressing the heuristic of providing ‘user control
and freedom’ (Nielsen, 2020). A text message at
the top of the UI was inserted (Figure 4) to remind
users about always using a template when start-
ing to design a cybersecurity-related diagram (ad-
dressing the ‘error prevention’ heuristic; Nielsen,
2020).
Icon Library: The icon library at the left-hand
side of the UI (Figure 4) was redesigned and
simplified in relation to the one ‘inherited’ from
the Diagrams.net software. Icon groups of vari-
ous other charts included in Diagrams.net, such
as network charts, flowcharts, tables, wireframes
and business charts, were removed, since they did
not comply with the tool’s cybersecurity-related
intended use. That way, a more focused, min-
imalist design approach was followed (address-
ing the ‘aesthetic and minimalist design’ heuris-
tic; Nielsen, 2020) with the aim of minimising
users’ cognitive load by enabling them to focus
only on icons defined by the CJML concrete syn-
tax under the icon categories of Actors, Commu-
nication channels and Special symbols (address-
ing the ‘recognition rather than recall’ heuristic;
Nielsen, 2020). Naturally, users are still able to
create simple shapes, such as squares and circles,
by using the search function or the ‘+ symbol on
the horizontal toolbar.
Saving and Exporting Diagrams: Users can
save their diagrams in the native Draw.io, XML-
based format and can also export them in var-
ious image and document formats. The saving
and export functionalities are located under the
‘File’ drop-down menu (Figure 4). Diagrams.net
software, by default, enables saving diagrams lo-
IVAPP 2023 - 14th International Conference on Information Visualization Theory and Applications
206
Figure 5: Four screenshots presenting the UI of HORM-DT: a) the landing page, b) the ‘Templates menu’, c) the canvas and
d) the ‘Save as’ box.
cally and on connected, commercial cloud stor-
age providers, such as Google Drive and MS
OneDrive. Since these storage options belong
to commercial platforms (Von Der Assen et al.,
2022), which users may or may not use, their in-
tegration was excluded, and only saving/exporting
to the local drive was provided as an option; that
way, confusing users by suddenly introducing var-
ious storage options when saving/exporting could
be avoided (addressing the ‘consistency’-related
heuristic; Nielsen, 2020).
Help and Guidance: The ‘Help’ dropdown menu
(Figure 4) was edited so that it provides users
guidance on the use of the tool and information
on the HORM framework. This change addressed
the ‘help and documentation’ heuristic (Nielsen,
2020).
3.3 Use
The basic functionalities of HORM-DT are presented
as follows:
On the landing page, users can choose whether to
load a previously saved diagram or to create a new
one (Figure 5a).
Upon making the choice of creating a new dia-
gram, users are asked to choose a template (Fig-
ure 5b). Templates can contain just the elements
of the HORM formalism or come with guidance
about what each symbol represents and how it is
used. Upon choosing a template, a canvas appears
on which users can work and create models using
icons, shapes, text and arrows (Figure 5c).
When users want to save the model they have cre-
ated, there are two options. They can save the
model locally as a diagram file in the software’s
native XML-based format (Figure 5d), or they can
export it in the format of their choice (image or
document).
4 EVALUATION
4.1 Methodology
For the evaluation of HORM-DT, the focus was on the
tool’s perceived usability. Since the CJML formal-
ism and the CORAS framework have been validated
in a number of past user studies (Halvorsrud et al.,
2016a; Halvorsrud et al., 2021; Stølen, 2001; Stølen
et al., 2002; Dimitrakos et al., 2002; Raptis et al.,
2002), the main purpose of this work was to evalu-
ate the new UI elements and investigate the extent to
which HORM-DT (i.e., a diagramming tool based on
the Diagrams.net software) can facilitate user interac-
tion with the concrete syntax and the model creation
under the HORM-related formalism. A common way
of conducting the evaluation of modelling languages
and tools is to provide participants with textual de-
scriptions of scenarios/cases and ask of them to turn
them into models using the respective languages and
tools (Silva et al., 2018; Miranda et al., 2018). In this
work, and since the focus was exclusively on the tool,
the participants were provided with the visual rep-
resentations of cybersecurity-related user journeys,
The HORM Diagramming Tool: A Domain-Specific Modelling Tool for SME Cybersecurity Awareness
207
Figure 6: a) A HORM model of an online-service user journey with a problematic point/deviation, which had to be reproduced
by the study participants as the first, introductory task. b) A HORM model of a phishing attack against an SME employee that
had to be reproduced by the study participants as the second, main task.
namely ready-made models, and asked to reproduce
them using HORM-DT. Therefore, the participants
were asked to use and evaluate the UI of HORM-DT
without needing to ‘speak’ HORM’s modelling lan-
guage.
The evaluation study took place between March
and June 2022 and followed a within-subject design.
Moreover, a within-subject comparison took place be-
tween the participants with prior experience with di-
agramming tools (hereafter called ‘experienced’) and
those without (‘inexperienced’).
4.2 Participants
The participants were recruited from the authors’ in-
stitutions, and the majority were students from the
University of Oslo. Prior knowledge of cybersecu-
rity and experience using diagramming tools were not
required. All the participants provided informed con-
sent to participate in the study.
4.3 Measures
The study deployed a mixed-methods research design
utilising qualitative and quantitative measures.
Demographic data were collected in the initial
stage of the study. A demographic questionnaire
collected information on age, gender, whether par-
ticipants had previous experience with diagramming
tools and, if so, which tool(s) they were.
To measure usability, the 10-item System Usabil-
ity Scale (SUS) questionnaire (Brooke, 2013) was
used. SUS is an instrument that allows usability prac-
titioners and researchers to measure the subjective
usability of products and services. SUS has been
utilised in several studies for measuring the perceived
usability of modelling languages and their compo-
nents (Eterovic et al., 2015; Miranda et al., 2018;
Nair et al., 2021; Silva et al., 2018; Aboussoror et al.,
2013). The 10-item questionnaire can be adminis-
tered quickly and easily, and it returns scores rang-
ing from 0 to 100. SUS scores can also be trans-
lated to adjective ratings, such as ‘worst imaginable’,
‘poor’, ‘OK’, ‘good’, ‘excellent’, ‘best imaginable’
and grade scales ranging from A to F (Bangor et al.,
2009). SUS has been shown both to be a reliable and
valid instrument that is robust with a small number
of participants and to have the distinct advantage of
being technology-agnostic, meaning it can be used to
evaluate a wide range of hardware and software sys-
tems (Brooke, 2013; Brooke, 1996; Tullis and Stet-
son, 2004; Kortum and Acemyan, 2013).
Next, a general-feedback, open-ended question-
naire was administered to collect the participants’
comments on HORM-DT. The participants were
asked the following two questions: i) what they liked
about their use of and experience with HORM-DT
and ii) what they did not like about it. Participants
were free to describe their experiences textually, and
there were no follow-up questions from the experi-
menter’s side.
The main task completion time was measured to
document the time needed for participants to repro-
duce a model for a cybersecurity-related user journey.
Finally, informal observations took place by the
experimenter that focused on modelling accuracy;
that is, the similarity between the final user-submitted
models and the original models provided by the ex-
perimenter. The observations were transcribed in a
qualitative manner as high level descriptions of simi-
larities and differences.
4.4 Procedure
Due to COVID-19 restrictions, the entire evaluation
procedure took place online via the participants’ pre-
ferred video-conferencing platform (e.g., Zoom, MS
Teams or Skype), with them sharing their screens so
that the experimenter could observe their progress.
First, the participants were presented with an in-
troduction to the study, then they provided informed
consent and filled out a questionnaire on demograph-
ics and their experience with diagramming tools.
The experimenter then presented two tasks. The
first was about reproducing a journey diagram model
of an online service user journey with a problematic
IVAPP 2023 - 14th International Conference on Information Visualization Theory and Applications
208
point/deviation (Figure 6a). It served as the introduc-
tory task, and so it was used to introduce participants
to the tool and become familiar with it by perform-
ing simple actions. For the second task, the partici-
pants had to reproduce a swimlane diagram model of
a phishing attack against an SME employee (Figure
6b). This was the main task of the evaluation study
and, based on pilot testing, its estimated completion
time was 25—45 minutes.
The models that had to be reproduced in the two
tasks were sent to the participants at the time of the
study via the video-conferencing platform as exported
images.
There was no time limit for the completion of the
tasks, and no external help was allowed for the con-
struction of the diagrams. At the end of each task,
the participants had to export their reproduced mod-
els in PNG image format and send them to the exper-
imenter via the video-conferencing platform. There
was no control or set threshold regarding the mod-
elling accuracy of the tasks (produced vs original) as
a requirement for completing the study.
Finally, participants completed the SUS and gen-
eral feedback questionnaires.
4.5 Statistical Analysis
All data were analysed using the IBM SPSS version
29 software. The significance level was set at p <
0.05. Descriptive analysis was used to depict the de-
mographic data of the participants and to analyse the
SUS values. Welch’s t-test was used to detect dif-
ferences between the inexperienced and experienced
users’ SUS scores and main task completion times.
The data from the general feedback questions were
transcribed and then analysed through open and ax-
ial coding, which enabled core concepts, themes and
ideas to be identified. Two researchers coded the data
independently; the inter-rater reliability was assessed;
and any disagreements were identified, discussed and
settled.
5 RESULTS
Twenty-nine participants (N = 29, mean age: 27.76,
SD: 4.48, male/female: 17/12) evaluated HORM-DT.
Sixteen participants (N = 16, mean age: 26.38, SD:
3.56, male/female: 10/6) had never used diagram-
ming tools before (i.e., inexperienced users), and thir-
teen participants (N = 13, mean age: 29.46, SD: 5.03,
male/female: 7/6) had experience with diagramming
tools such as Draw.io/Diagrams.net, MS Visio, Miro
and the diagramming functions of MS Word and MS
PowerPoint (i.e., experienced users).
The HORM-DT interface scored a mean SUS
value of 80.69 (SD: 12.64). This value provided a
usability evaluation of between ‘good’ and ‘excel-
lent’, equivalent to a B grade (Brooke, 2013; Brooke,
1996). When the results were analysed and con-
trolled for participants’ previous experience with di-
agramming tools, the following results were identi-
fied (Table 1). Experienced users (N = 13) awarded
a mean SUS value of 82.5 (SD: 14.43), and inexperi-
enced users (N = 16) awarded a mean SUS value of
79.22 (SD: 11.24). No statistically significant differ-
ences were found between the experienced and inex-
perienced users’ SUS scores, based on Welch’s t-test:
t(22.4) = 0.67, p = 0.509.
Table 1: SUS results for HORM-DT.
Participants Mean SUS score
(SD)
All (N = 29) 80.69 (12.64)
– Experienced (N = 13) 82.5 (14.43)
– Inexperienced (N = 16) 79.22 (11.24)
The mean main task completion time (Table 2)
was 35.62 minutes (SD: 17.22); for experienced users
(N = 13), it was 29.38 minutes (SD: 13.2), and for
inexperienced users (N = 16), it was 40.69 minutes
(SD: 18.79). No statistically significant differences
were found between the experienced and inexperi-
enced users’ main task completion time, based on the
Welch’s t-test: t(26.53) = 1.9, p = 0.069.
Table 2: Main task completion time results for HORM-DT.
Participants Mean main-task
completion time
(SD)
All (N = 29) 35.62 min. (17.22)
– Experienced (N = 13) 29.38 min. (13.2)
– Inexperienced (N = 16) 40.69 min. (18.79)
Table 3 presents the participants’ comments col-
lected from the two general feedback questions, to-
gether with the frequency of their occurrence. The
participants’ comments were further characterised as
positive (P) and negative (N).
Based on observations regarding modelling accu-
racy, all participants reproduced the diagrams with
very high accuracy, and only some typographical er-
rors and small positioning differences in lines and
shapes featured in their produced models.
The HORM Diagramming Tool: A Domain-Specific Modelling Tool for SME Cybersecurity Awareness
209
Table 3: Participants comments as collected from the two general-feedback questions.
Comment Count
P HORM-DT is simple and easy to use. 18
N It can be challenging to use the arrows to connect objects and form straight lines. 9
P Icons and graphical objects are clearly designed and categorised. 6
P Templates are useful and easily accessible. 4
P HORM-DT is suitable for novice users. 3
P It is easy to connect objects. 3
N When inserting text, the default font size is too small. 2
N HORM-DT does not contain all the libraries of Diagrams.net. 1
N Some icons are not immediately visible and need to be searched for. 1
6 DISCUSSION
In this work, HORM-DT was designed and devel-
oped to be a standalone, open and customised tool for
the application of extended CJML within the HORM
framework. The tool’s UI and functionality were
based on the Diagrams.net software and then signifi-
cantly extended and redesigned to fit the customised
needs of the modelling language domain. To that end,
a heuristic evaluation of the Diagrams.net UI took
place, and new icon libraries and templates were cre-
ated and new interaction steps regarding model devel-
opment were introduced, among other changes (Sec-
tion 3.2). The usability of the tool was central to the
study, paving the way for using HORM-DT when ex-
amining the effectiveness of the modelling language
for raising SME cybersecurity awareness.
Since HORM-DT featured satisfactory and good
usability values, based on the SUS score, it can be said
that the chosen design and development process con-
tributed to reaching the goal. The usability, simplicity
and user-friendliness of the tool were also praised by
the participants. The clearly designed icons, graphi-
cal objects and useful templates may have contributed
to that, based on the participants’ comments. On the
other hand, the drawing of lines was negatively rated,
since Diagrams.net features a very simplistic way of
connecting items with lines (by connecting predefined
points); however, for new users, this might not have
been obvious and required a starting tutorial.
As for the comparison of the perceived usability
and main task completion time between experienced
and inexperienced users (Tables 1 and 2), there were
naturally differences; however, those differences were
not found to be statistically significant. Therefore, it
may be said that HORM-DT can facilitate use even
by users that have no prior experience with diagram-
ming tools and fulfil the goal of enabling modelling
by users with diverse technical backgrounds (Section
1).
Finally, in this work, the decision was made to
use existing open-source code (i.e., Diagrams.net), in-
stead of starting from the ground up. This happened
for the following two reasons: i) Diagrams.net is a
widely used tool with a user-friendly UI, and it is
used frequently in the modelling-language domain (as
mentioned in Section 3.1), and ii) a green coding strat-
egy based on reusing existing code (Verdecchia et al.,
2021) was followed. In terms of the latter strategy,
this work alone may not have had an enormous en-
vironmental impact, but it has helped create a coding
attitude towards the environmental goal of reducing
software’s energy consumption, which can be applied
in future projects and contribute to raising awareness
about the topic.
6.1 Implications
In this work, the following elements can be gener-
alised and inform the designs of other practitioners
and researchers in the field:
A modelling tool for mapping cybersecurity-
related user journeys is openly deployed for use
by other practitioners and researchers in their
cybersecurity-related use cases. The tool features
satisfactory usability and can facilitate modelling
by users with diverse technical backgrounds.
Moreover, the tool’s code is openly distributed for
interested parties to further extend and build upon.
An effective and resource-efficient methodolog-
ical approach is introduced to measure the per-
ceived usability of modelling tools when the un-
derlying formalism remains the same, albeit with
an updated UI. The proposed methodology fol-
lows formalism-agnostic logic when the UI of a
modelling-language tool is updated so that i) the
evaluation results solely address and focus on the
updated UI elements and their usability perfor-
mance, and ii) resources such as time and cost are
saved during the evaluation process, since partic-
IVAPP 2023 - 14th International Conference on Information Visualization Theory and Applications
210
ipants are not required to become accustomed to
the tool’s formalism in order to use it for the pur-
poses of the study.
An extension of the aforementioned methodologi-
cal approach is also introduced to measure the per-
ceived usability coming from users with diverse
attributes (in this case, prior experience with dia-
gramming tools). Therefore, in cases in which a
modelling language tool targets a heterogeneous
user group, a within-group comparative approach
based on the differentiating attribute can benefit
the usability measurements and the conclusions
that come out of it.
6.2 Limitations
This evaluation study of HORM-DT had the follow-
ing limitations:
COVID-19 restrictions affected the evaluation
process. Although the tasks were digital and
could take place remotely, the qualitative part of
the evaluation study might have been carried out
in a more effective way with physical attendance,
such as through personal interviews. The restric-
tions also affected the recruitment process, lead-
ing to convenience sampling and contributing to
the next point.
Since the majority of the study participants were
students from the University of Oslo (i.e., con-
venience sampling), a participant sample of a
young age (mean age: 27.76, SD: 4.48) was re-
cruited. The study participants’ technical back-
ground and prior experience with diagramming
tools may have been affected by their young age.
The participants, especially those without prior
experience with diagramming tools, may have
needed detailed guidance to become familiar with
drawing lines when using HORM-DT. Neverthe-
less, the logic behind drawing lines and connect-
ing elements became clear after a while, but a
more detailed starting tutorial could alleviate the
few negative experiences on that matter, as docu-
mented in Table 3.
7 CONCLUSION
In this work, HORM-DT was introduced as a tool
for modelling cybersecurity-related journeys to help
SMEs to raise cybersecurity awareness and take the
first step towards defining or updating their cyberse-
curity practices. The tool was examined in terms of
the usability its UI offers to users, producing satisfac-
tory and promising results. Apart from the actual tool,
which has been openly deployed online and its code
is freely accessible, the work also introduced a cost-
effective methodological approach for evaluating the
usability of modelling tools that have updated UIs but
still feature the same formalism. In the future, further
work will be conducted on i) developing a starting
tutorial (e.g., an introductory video) on how simple
HORM-DT diagrams can be made, with an empha-
sis on drawing lines and connecting objects; ii) ex-
ploring the effectiveness of the overall HORM mod-
elling language, as facilitated by HORM-DT, for rais-
ing SME cybersecurity awareness and involving var-
ious target groups, such as cybersecurity experts and
regular, non-technical employees; and iii) developing
a text editor as an extension of the tool that would en-
able users to textually describe their models, which
would be automatically translated to diagrams.
ACKNOWLEDGEMENTS
We would like to thank the IVAPP 2023 review-
ers for their constructive feedback. This research
is funded by the European Commission through the
CyberKit4SME project (www.cyberkit4sme.eu) un-
der Grant Agreement 883188. CyberKit4SME will
provide cybersecurity tools that help SMEs become
aware of, analyse, and manage cybersecurity and data
protection risks.
REFERENCES
Aboussoror, E. A., Ober, I., and Ober, I. (2013). Signifi-
cantly increasing the usability of model analysis tools
through visual feedback. In International SDL Forum,
pages 107–123. Springer.
Ag
´
ardi, A. (2022). Relontouml model of the archaeological
findings. Production Systems and Information Engi-
neering, 10(1):64–71.
Arctic Wolf (2017). The state of mid-market
cybersecurity: Findings and implications.
https://2p167arhj4lo70dn1q26fm1c-wpengine.netdn
a-ssl.com/wp-content/uploads/AW Brief Midmarket
Cybersecurity Survey.pdf (Accessed 17 Nov 2022).
Bangor, A., Kortum, P., and Miller, J. (2009). Determin-
ing what individual SUS scores mean: Adding an
adjective rating scale. Journal of Usability Studies,
4(3):114–123.
Bellamy, R. K., Erickson, T., Fuller, B., Kellogg, W. A.,
Rosenbaum, R., Thomas, J. C., and Wolf, T. V. (2007).
Seeing is believing: Designing visualizations for man-
aging risk and compliance. IBM Systems Journal,
46(2):205–218.
The HORM Diagramming Tool: A Domain-Specific Modelling Tool for SME Cybersecurity Awareness
211
Benz, M. and Chatterjee, D. (2020). Calculated risk? A cy-
bersecurity evaluation tool for SMEs. Business Hori-
zons, 63:531–540.
Boletsis, C., Halvorsrud, R., Pickering, J. B., Phillips, S. C.,
and Surridge, M. (2021). Cybersecurity for SMEs:
Introducing the Human Element into Socio-technical
Cybersecurity Risk Assessment. In VISIGRAPP (3:
IVAPP), pages 266–274.
Brooke, J. (1996). SUS - A quick and dirty usability scale.
In Jordan, P., Thomas, B., Weerdmeester, B., and Mc-
Clelland, I., editors, Usability Evaluation in Industry,
pages 189–194. Taylor & Francis.
Brooke, J. (2013). SUS: A Retrospective. Journal of Us-
ability Studies, 8(2):29–40.
Ch
´
avez-Feria, S., Garc
´
ıa-Castro, R., and Poveda-Villal
´
on,
M. (2021). Converting UML-based ontology concep-
tualizations to OWL with Chowlk. In European Se-
mantic Web Conference, pages 44–48. Springer.
Ch
´
avez-Feria, S., Garc
´
ıa-Castro, R., and Poveda-Villal
´
on,
M. (2022). Chowlk: from UML-based ontology con-
ceptualizations to OWL. In European Semantic Web
Conference, pages 338–352. Springer.
Dimitrakos, T., Ritchie, B., Raptis, D., and Stølen, K.
(2002). Model based security risk analysis for web
applications: the CORAS approach. In EuroWeb 2002
Conference, pages 1–13.
Eterovic, T., Kaljic, E., Donko, D., Salihbegovic, A., and
Ribic, S. (2015). An Internet of Things visual do-
main specific modeling language based on UML.
In 2015 XXV International Conference on Informa-
tion, Communication and Automation Technologies
(ICAT), pages 1–5. IEEE.
Fair, N., Phillips, S., Erdogan, G., and Tverdal, S. (2022).
Information security & risk management: trustwor-
thiness and human interaction. In Tutorial proceed-
ings of the 16th International Conference on Research
Challenges in Information Science, pages 1–48.
RCIS. https://www.rcis-conf.com/rcis2022/files/T1-
NicholasFairEtAl.pdf.
Halvorsrud, R., Boletsis, C., and Garcia-Ceja, E. (2021).
Designing a modeling language for customer jour-
neys: Lessons learned from user involvement. In 2021
ACM/IEEE 24th International Conference on Model
Driven Engineering Languages and Systems (MOD-
ELS), pages 239–249. IEEE.
Halvorsrud, R., Haugstveit, I. M., and Pultier, A. (2016a).
Evaluation of a modelling language for customer jour-
neys. In 2016 IEEE Symposium on Visual Languages
and Human-Centric Computing (VL/HCC), pages 40–
48. IEEE.
Halvorsrud, R., Kvale, K., and Følstad, A. (2016b). Improv-
ing service quality through customer journey analysis.
Journal of Service Theory and Practice, 24(6):840–
867.
Halvorsrud, R., Lee, E., Haugstveit, I. M., and Følstad, A.
(2014). Components of a visual language for service
design. In ServDes. 2014 Service Future, Proceedings
of the fourth Service Design and Service Innovation
Conference, Lancaster University, United Kingdom,
9-11 April 2014.
Haugstveit, I. M., Halvorsrud, R., and Karahasanovic, A.
(2016). Supporting redesign of C2C services through
customer journey mapping. In Service Design Ge-
ographies. Proceedings of the ServDes. 2016 Confer-
ence, number 125, pages 215–227. Link
¨
oping Univer-
sity Electronic Press.
Jaimez-Gonz
´
alez, C. and Mart
´
ınez-Samora, J. (2020).
DiagrammER: A Web Application to Support the
Teaching-Learning Process of Database Courses
Through the Creation of ER Diagrams. International
Journal of Emerging Technologies in Learning (iJET),
15(19):4–21.
JGraph (2022). Features of diagrams.net and draw.io.
https://diagrams.net/features (Accessed 17 Nov
2022).
Khan, M. I., Tanwar, S., and Rana, A. (2020). The Need
for Information Security Management for SMEs. In
2020 9th International Conference System Modeling
and Advancement in Research Trends (SMART), pages
328–332. IEEE.
Kortum, P. and Acemyan, C. Z. (2013). How low can you
go? Is the System Usability Scale range restricted?
Journal of Usability Studies, 9(1):14–24.
Kullman, K., Buchanan, L., Komlodi, A., and Engel, D.
(2020). Mental model mapping method for cyber-
security. In International Conference on Human-
Computer Interaction, pages 458–470. Springer.
Lund, M. S., Solhaug, B., and Stølen, K. (2011). Risk anal-
ysis of changing and evolving systems using CORAS.
In International School on Foundations of Security
Analysis and Design, pages 231–274. Springer.
Meshkat, L., Miller, R. L., Hillsgrove, C., and King, J.
(2020). Behavior modeling for cybersecurity. In
2020 Annual Reliability and Maintainability Sympo-
sium (RAMS), pages 1–7. IEEE.
Meszaros, J. and Buchalcevova, A. (2017). Introducing
OSSF: A framework for online service cybersecurity
risk management. computers & security, 65:300–313.
Miranda, T., Challenger, M., Tezel, B. T., Alaca, O. F.,
Bari
ˇ
si
´
c, A., Amaral, V., Goul
˜
ao, M., and Kardas, G.
(2018). Improving the usability of a mas dsml. In In-
ternational workshop on engineering multi-agent sys-
tems, pages 55–75. Springer.
Nair, A., Ning, X., and Hill, J. H. (2021). Using recom-
mender systems to improve proactive modeling. Soft-
ware and Systems Modeling, 20(4):1159–1181.
Nielsen, J. (2020). 10 Usability Heuristics for User Inter-
face Design . https://www.nngroup.com/articles/ten-
usability-heuristics/ (Accessed 17 Nov 2022).
Nurse, J. R., Creese, S., Goldsmith, M., and Lamberts, K.
(2011). Trustworthy and effective communication of
cybersecurity risks: A review. In 2011 1st Work-
shop on Socio-Technical Aspects in Security and Trust
(STAST), pages 60–68. IEEE.
Paulsen, C. (2016). Cybersecuring small businesses. Com-
puter, 49(8):92–97.
Ponsard, C. and Grandclaudon, J. (2019). Guidelines and
Tool Support for Building a Cybersecurity Aware-
ness Program for SMEs. In International Conference
IVAPP 2023 - 14th International Conference on Information Visualization Theory and Applications
212
on Information Systems Security and Privacy, pages
335–357. Springer.
Raptis, D., Dimitrakos, T., Gran, B. A., and Stølen, K.
(2002). The CORAS approach for model-based risk
management applied to e-commerce domain. In Ad-
vanced Communications and Multimedia Security,
pages 169–181. Springer.
Rauer, M. (2019). Draw.io diagramming in Con-
fluence is currently the most successful app in
the Atlassian Marketplace. https://blog.seibert-
media.com/2019/04/18/draw-io-diagramming-in-
confluence-is-currently-the-most-successful-app-
in-the-atlassian-marketplace/ (Accessed 17 Nov
2022).
Shojaifar, A. (2019). SMEs Confidentiality Issues and
Adoption of Good Cybersecurity Practices. In IFIP
Summer School on Privacy and Identity Management,
pages 1–8.
Silva, J., Bari
ˇ
sic, A., Amaral, V., Goul
˜
ao, M., Tezel, B. T.,
Alaca, O. F., Challenger, M., and Kardas, G. (2018).
Comparing the usability of two multi-agents systems
DSLs: Sea ml++ and DSML4MAS study design. In
3rd International Workshop on Human Factors in
Modeling, volume 2245, pages 770–777.
Skubisz, C., Reimer, T., and Hoffrage, U. (2009). Com-
municating quantitative risk information. Annals
of the International Communication Association,
33(1):177–211.
Slovic, P. (1999). Trust, emotion, sex, politics, and science:
Surveying the risk-assessment battlefield. Risk analy-
sis, 19(4):689–701.
Stickdorn, M., Hormess, M. E., Lawrence, A., and Schnei-
der, J. (2018). This is service design doing: applying
service design thinking in the real world. O’Reilly
Media, Inc.”.
Stølen, K. (2001). CORAS A Framework for Risk Analy-
sis of Security Critical Systems. In supplement of the
2001 International Conference on Dependable Sys-
tems and Networks, pages D4-D11.
Stølen, K., den Braber, F., Dimitrakos, T., Fredriksen, R.,
Gran, B. A., Houmb, S.-H., Lund, M. S., Stamatiou,
Y., and Aagedal, J. (2002). Model-based risk assess-
ment – the CORAS approach. In iTrust Workshop.
Surridge, M., Meacham, K., Papay, J., Phillips, S. C., Pick-
ering, J. B., Shafiee, A., and Wilkinson, T. (2019).
Modelling compliance threats and security analysis of
cross border health data exchange. In International
Conference on Model and Data Engineering, pages
180–189. Springer.
Symantec (2019). Symantec 2019 internet security threat
report. https://docs.broadcom.com/doc/istr-24-2019-
en (Accessed 17 Nov 2022).
The National Center for the Middle Market (2016).
Cybersecurity and the middle market: The
importance of cybersecurity and how mid-
dle market companies manage cyber risks.
https://www.middlemarketcenter.org/Media/-
Documents/the-importance-of-cybersecurity-and-
how-middle-market-companeis-manage-cyber-
risks NCMM Cybersecurity Report FINAL.pdf
(Accessed 17 Nov 2022).
Tullis, T. S. and Stetson, J. N. (2004). A comparison
of questionnaires for assessing website usability. In
Proceedings of the Usability Professional Association
Conference, pages 1–12. Usability Professional Asso-
ciation.
Vakaliuk, T. A., Korotun, O. V., and Semerikov, S. O.
(2021). The selection of cloud services for ER-
diagrams construction in IT specialists databases
teaching. In CTE Workshop Proceedings, volume 8,
pages 384–397.
Verdecchia, R., Lago, P., Ebert, C., and De Vries, C.
(2021). Green IT and green software. IEEE Software,
38(6):7–15.
Von Der Assen, J., Franco, M. F., Killer, C., Scheid, E. J.,
and Stiller, B. (2022). CoReTM: An Approach En-
abling Cross-Functional Collaborative Threat Model-
ing. In 2022 IEEE International Conference on Cyber
Security and Resilience (CSR), pages 189–196. IEEE.
Vraalsen, F., Lund, M. S., Mahler, T., Parent, X., and
Stølen, K. (2005). Specifying legal risk scenarios us-
ing the CORAS threat modelling language. In Inter-
national Conference on Trust Management, pages 45–
60. Springer.
The HORM Diagramming Tool: A Domain-Specific Modelling Tool for SME Cybersecurity Awareness
213