=Paper=
{{Paper
|id=Vol-1573/Paper_1_BCSS2016
|storemode=property
|title=Building an App for Behavior Change: Case RightOnTime
|pdfUrl=https://ceur-ws.org/Vol-1573/Paper_1_BCSS2016.pdf
|volume=Vol-1573
|authors=Piiastiina Tikka,Bereket Woldemicael,Harri Oinas-Kukkonen
|dblpUrl=https://dblp.org/rec/conf/persuasive/TikkaWO16
}}
==Building an App for Behavior Change: Case RightOnTime==
https://ceur-ws.org/Vol-1573/Paper_1_BCSS2016.pdf
3 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
Building an App for Behavior Change: Case
RightOnTime
Piiastiina Tikka, Bereket Woldemicael, and Harri Oinas-Kukkonen
Oulu Advanced Research on Service and Information Systems Research Group (OASIS),
University of Oulu, Finland
piiastiina.tikka@oulu.fi, berektka@gmail.com, harri.oinas-
kukkonen@oulu.fi
Abstract. The present paper describes and illustrates the process of developing
a mobile application intended as a Behaviour Change Support System for peo-
ple who want to improve their punctuality and time-keeping skills. Design sci-
ence approach together with Persuasive Systems Design (PSD) model was used.
The presented article focuses on the software development process from the
perspective of using the PSD model as the central design framework and design
science approach (including user evaluation) as the guiding development pro-
cess.
Keywords: Behavior change; timing strategies; persuasive systems design; de-
sign research, systems development, software design.
1 Introduction
Much of the current literature on persuasive systems focuses on describing the behav-
ioral challenge and outcomes after interventions, but the system used in the process is
seldom described in clear detail as regards software architecture, leaving developers
and researchers alike with limited insights as regards the design of the systems. Such
lack of visibility results in so-called black box thinking [1] where it is difficult to
assess what the intended behavior change of a system is, and how the software artifact
has been constructed in order to support that behavior change objective.
The present paper describes the process of designing and developing a persuasive
mobile device app. The design and development work is based on the theoretical
frameworks of the Persuasive Systems Design (PSD) model and it employs Design
Science Research Method (DSRM) [2].
In this design science research study the selected problem domain for the applica-
tion is chronic tardiness and challenges with punctuality: the inability to do things in a
timely manner. Such issues can contribute to a person’s general mental well-being
through social pressures and the stress involved; punctuality problems can also affect
productivity and success at work or in education [3,4]. The presented software (SW)
artifact, a mobile app, provides its users with tools for reflecting upon their actions
and behavior as well as a way of practicing time awareness.
4 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
2 Background
2.1 Persuasive Systems Design model
According to the Persuasive Systems Design (PSD) model [5] the development of a
persuasive system has three main components: understanding key issues of persuasive
design (presented as seven postulates), analysis of the Persuasion Context, and the
design of system qualities. The seven postulates describe the underlying notions that
guide persuasive system development, for example, designers and developers to real-
ize that persuasive systems should always be open, that they should aim at unobtru-
siveness, and that persuasion is often incremental [5].
A crucial stage of persuasive systems development is the accurate and careful
analysis of the Persuasion Context. The context includes determining what is the in-
tent of persuasion (what the system is expected to achieve), what is the persuasion
event, and who is the persuader (also part of the intent of persuasion). The PSD model
stipulates that persuasion intent includes the persuader (stakeholders deciding on the
system components) and the change type (does the system seek change in behavior,
attitude, or both) [5]. The persuasion event in the model accounts for Use Context,
User Context and Technology Context. The analysis of the Persuasion Context, then,
advises the selection of persuasive design features. The analysis will be presented in
section 4.1.
2.2 Behavioral problem domain: punctuality
Up to 20% of people in the USA have issues with punctuality [6]. The reasons for
tardiness are various, and such behavior cannot simply be attributed to lack of respect
or consideration. The skills to break the bad habits are learnable by practically anyone
[3]. A close relation of tardiness, procrastination, can be viewed as a behavior pattern
that can lead to negative outcomes such as stress, illness and inferior performance [4].
The social costs of tardiness can also be significant if, for example, a person’s career
prospects start to suffer because he or she is seen as always late and therefore unrelia-
ble [3].
Factors that are common among those people who do not have problems with
keeping time include a realistic mindset as regards time, a readiness to plan schedules
to tolerate unforeseen delays, a tendency to be organized and to analyze their actions,
and not minding having little downtime between activities [6]. Typical tardy personal-
ity types include those who enjoy the adrenaline rush brought on by leaving a piece of
work until the very last moment before a deadline, and those who enjoy the ego-
stroking effect of trying to get as many things done as possible in as little time as
possible [3].
These considerations have been taken into account in the Persuasion Context anal-
ysis.
5 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
3 Method: Design Science Research
The presented development project follows the Design Science Research (DSR), as
outlined by Hevner et al. [2]. The research problem in the present paper is to see how
to apply the full range of PSD design principles to a specified behavioral problem
domain and thus make the role of the information system involved more explicit and
transparent, avoiding black box thinking [1]. Additional motivation comes from inter-
est in producing a tool that can help end-users manage time better and learn to be
more punctual in their everyday life. The app is available on iOS devices for free
through Apple Store.
4 Building a BCSS: RightOnTime app
The software artifact developed to address the identified behavior change targets in
the area of time management and punctuality was named RightOnTime in order to
suggest timeliness and successful time management. The main features of the app
include appointment/event planning (‘Activity’) and a time awareness practice (‘Prac-
tice’) section where users can train their time-awareness. In addition, the app includes
push notification tips and hints that provide the user with suggestions, encouragement
and further information regarding punctuality and achieving better time-keeping
skills. The tips and hints are intended as support material for the user’s behavior
change process and support the strategy of direct persuasion. The messages are not
tailored to individual users but are general and based on the most common time per-
ception issues identified in the guiding literature [3].
4.1 Persuasion Context analysis
The initial steps of the development process involved the analysis of the Persuasion
Context: intent, event, and strategy [1].
Intent. The behavior change intent of the app is to assist end-users to achieve change
in their time management and punctuality skills as regards both behavior and possibly
also attitude levels [7].
Event. Persuasion event consists of the concepts of Use, User, and Technology Con-
text.
Use Context: The focus of activities in the app is on everyday routines and events
– whether private, education or work related – and the content and selection of activi-
ties is entirely up to the user. The application encourages planning time use, reducing
the larger task of learning punctuality into smaller chunks of individual tasks, and
then reflecting on action after set events and activities. These activities stem from the
literature [3] suggesting that in many cases tardiness is a result of unrealistic and
overly optimistic expectations of time use, and that with better awareness and strate-
gies tardiness problems can be overcome. In terms of SW features, Use Context cate-
6 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
gory comprises the overall environment (user interface, data storage, etc.) as dictated
by the problem domain.
User Context: The application is intended for individuals who want to develop
their time-management skills and become more punctual in their everyday life. The
potential users are expected to use the system because they have experienced social
discomfort as a result of their inability to be punctual, or have found that better time
management skills (and punctuality) would benefit them in their professional (or
equivalent) aspirations. Looking at some characteristics of punctual people [3][6] and
turning them into practical actions anyone can do, the present project focused on a)
learning time-awareness in order to increase the above-mentioned realistic mind-set,
b) planning time use, and c) routinely analyzing activities. Each of these user motiva-
tions was addressed by supporting increasing awareness and reflection: users were
directed to pay attention to their daily appointments and events and then reflect on
their own actions [8] afterwards; they were also given means to learn time awareness
by measuring how long certain tasks actually take. The tips and hints sent to the users
as push notifications were based on existing literature on how to conquer chronic
tardiness [3].
Technology Context: The application is developed for mobile iOS devices (name-
ly, iPhone but it also runs on iPads). The benefit of developing an app for mobiles is
in making the BCSS available at all times, thereby matching the Use Context as close-
ly as possible. Developing the system for iOS devices necessarily ruled out users of
other mobile platforms (Android, Windows, Sailfish, BlackBerry, etc.) at this point in
time, but the design principles and persuasive feature set applied in the present project
can also be applied to other platforms.
Strategy. The analysis of feasible approaches, namely, planning, reflection and
awareness, indicates that the app will use a direct route in the delivery of the persua-
sive content. Focus on features that require the user to evaluate his or her actions (re-
flect) promotes the direct route as the user has to elaborate on the actions, that is,
engage in intellect based debate over actions and their consequences instead of rely-
ing on easily-available heuristics or unfounded assumptions [9]. For example, accord-
ing to DeLonzor [3] a typical misconception might be that we think that a journey that
in reality takes 40 minutes can be done in 30 minutes because there once was an occa-
sion where it happened. Relying on the one instance of 30 minutes uses, essentially,
an availability heuristic. By reflecting (elaborating) on the actual time the journey
takes as measured and recorded, the system user can re-adjust the time expectation.
In addition, overall message quality (including knowing that its source is reliable)
promotes use of central processing.
4.2 Persuasive software features
Every section of the application includes a description of its purpose and guidance on
how to use it. The app also provides a selection of references to background infor-
mation on chronic tardiness.
7 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
What is more interesting for our purposes here, based on the Persuasion Context
analysis, a set of feasible persuasive features was selected from three of the four PSD
design feature categories [5]: Primary task support, Dialogue support, and Credibility
support. No social support features were implemented. The main features are listed
and described in Table 1.
Table 1. Persuasive features implemented in the RightOnTime app.
Primary Reduction The overall objective of “be more punctual” can seem
task sup- too big to manage as such, and therefore the objective is
port split into smaller goals and objectives: individual tasks.
Also, by having control over each activity and practice
task, the user is allowed to select less critical tasks, thus
reducing the possible stress involved if, for example,
failure at a task has some serious consequences.
Self- User can do self-monitoring based on self-assessments
monitoring regarding their own tasks (timeliness and satisfaction).
The app also requests user to perform reflection by
showing summaries of their daily assessments. In addi-
tion, the Practice feature supports observing one’s own
behavior through set tasks so that it is possible to learn
better time estimation.
Computer- Reminders The app provides reminders for the user-set tasks in it-
human self. The reminders considered here concern the push
dialog notifications for users to complete their self-assessments
support and to view their assessment summaries.
Suggestion Tips and Hints push notifications are provided daily
(three per day). The content of the notifications provide
information, encouragement and guidance as regards
punctual behavior.
System Real- “About” section provides information about the origin,
credibility World Feel purpose and designers of the app. The development team
support members’ credentials and contact information are in-
cluded in the “About” section of the app in order to
communicate openly who is behind the app and to en-
courage users to ask for more information if necessary.
4.3 Software architecture
Persuasion Context manifests itself in architecture terms as components for persua-
sion event (Use Context, User Context and Technology Context) and as a layer of
components for persuasion message and route (system-user interaction and system-
mediated messages, in the case of RightOnTime) [5][10].
8 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
The architecture, seen in Figure 1, shows the four relevant components. The User
context involves information about the user, and based on this information appropri-
ate persuasion strategies can be employed. Use Context enables the user interface and
data storage to employ the appropriate persuasion strategies. User-system Interaction
includes means for the user to set tasks, complete assessments, view results, etc., and
it allows the system to record the necessary data about the users and by the users.
System-mediated Messages component defines what content and which route is used
for conveying behavior change messages to the user. [5][10]
Fig. 1. Persuasion Context software architecture for RightOnTime app, modified from [10]:
RightOnTime app does not use Social interaction features.
The technology context (iOS) in the case of RightOnTime has guided the applied
software pattern in practice. Apple’s application development interface uses Model
View Controller (MVC) as a design pattern1, which involves three groups: the model,
the view and the controller. The model handles objects related to data, view handles
objects related to user interface and controller is a mediator between view and model
object. In our application, model object contains data related to calendar event, local
notifications, and user settings. Such data are persistent in the application and have no
explicit relationship with the application’s view object. The interaction between the
application’s model and view objects takes place through the controller object. In our
application there are three persistent data storages: device’s database (SQLite), Cal-
endar and Local notification. A user’s setting and user’s self-evaluations data are
stored in SQLite. iOS application framework provides access to iOS calendar data-
base, and we utilized iOS calendar to store data related to user’s events. Data such as
notification time and notification message are kept in local notification.
All graphics, the user interface, and interactions are handled by view object. The
user can view and update model objects. However, view and model objects do not
have any explicit connection and all interaction between view and model object is
1
Model-View-Controller. Apple Inc.
https://developer.apple.com/library/mac/documentation/General/Conceptual/DevPedia-
CocoaCore/MVC.html
9 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
handled through the controller object. The controller object plays an intermediary role
between the application’s view object and the model object. It learns updates from a
user and notifies the update to the model object. It is also used as a tunnel to display
the model object to the user.
The software was implemented in Objective-C programing, which is the primary
language for developing iOS and OX software2. Objective-C follows syntax of primi-
tive C language, and it has also object-oriented features. In addition to plain objective
C-programing language, we have also utilized Apple’s foundation framework to de-
velop our application. The app’s interface follows iOS human interface guidelines.3,
and the look and feel aims at being “design neutral” in terms of use of any custom
components. Instead the look and feel of the app relies on Apple’s framework com-
ponents to enhance the sense of familiarity and trustworthiness that comes with hav-
ing used the device already.
5 Artifact demonstration
Based on the Persuasion Context analysis, the selected persuasive features (Table 1)
were delivered in three main elements of the application: Activity, Practice and Tips
and Hints. Visually the app displays a tabular interface with four main sections:
Home, Practice, Self-evaluation and Setting sections (Figure 2). Navigation from one
section to another is done by selecting a relevant tab button.
Home (view for Activity feature) and Practice tabs: The user can add delete or
modify tasks. When there are no tasks on the relevant day, the page shows a brief
description of the purpose of the view. The description also includes instructions for
how to set new events (helpful for first time use).
Tips and Hints: The tips and hints are shown as push notifications (Figure 3). The
timing of these notifications can be random (default) or the user can fix set times for
three notification per day under Settings (Figure 4).
2
Programming with Objective-C. Apple Inc.
https://developer.apple.com/library/mac/documentation/Cocoa/Conceptual/ProgrammingWit
hObjectiveC/Introduction/Introduction.html
3
iOS Human Interface Guideline, Apple Inc.
https://developer.apple.com/library/ios/documentation/UserExperience/Conceptual/MobileH
IG/
10 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
Fig. 2. RightOnTime Home (left) and Practice (center) tabs with example activities, and an
example of a tip notification (right).
Self-Evaluation and performance summary views: Upon completion of tasks, the
user is prompted to evaluate the timeliness and perceived effort involved in the task,
and at a specified time in the day the system prompts the user to review a summary of
that day’s activity assessments (Figure 3).
Fig. 3. RightOnTime self-evaluation views: list of activities (left), self-evaluation view (center)
and a summary of evaluations of one activity (right).
Setting: Under settings the user can opt to use the system under default settings (ran-
dom time for all notifications, including requests to provide assessments), or custom
settings where user sets times for all notifications and interactions with the system
(Figure 4).
11 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
Fig. 4. Settings: Continuous in active mode produces reminders to complete the self-
assessments within a 15 minute window after end-time of any user-set task, and to produce
three Tips and Hints notifications during the daytime. The randomness is not absolute, but the
notifications come during morning, midday and afternoon/evening periods, and not at night.
When Continuous mode is in off position, the user can set a time for when he wishes to com-
plete the self-assessments, and when to receive the three Tips and Hints notifications. In each
case, summary time can be set by the user.
6 Design science research evaluation
An evaluation of the persuasiveness of the app’s main features of Primary task sup-
port and Computer-human dialogue support was conducted by means of employing
five respondents to use the system for five days. After the experiment period the par-
ticipants evaluated their experiences on a Likert-type scale and by providing open-
ended responses. The questions for the Likert-type ratings were adapted from a per-
ceived persuasiveness questionnaire [11], and these questions were followed by an
open-ended text field where participants could elaborate on their rating. All partici-
pants responded on an electronic form. The question items were analyzed using a
Goal-Question-Metric approach [12]. In addition to the GQM approach, the respond-
ents also completed a desirability test [13] based on an approach developed at Mi-
crosoft [14], in which participants were presented with 118 adjectives (60% positive,
40% negative) and they then had to choose five that best described the system. The
desirability test is intended to indicate overall satisfaction with a system.
Table 2. GQM items derived from the PSD feature selection (Table 1).
Goal Question Metric
Primary Reduction of larger behav- Does the application help Likert
task sup- ior targets into smaller and users to produce target /open
port more manageable tasks. behavior? ended
comments.
12 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
The system should help the Does the application ena-
user understand his or her ble self-monitoring?
present time management
problems and eventual de-
velopments through self-
monitoring.
Computer- Remind user of timeliness Does the app help users to Likert
human tasks and of reflection- perform required tasks in /open
dialog supporting features (self- the application? ended
support monitoring tasks). comments
The system should increase Does the system increase Likert
the users knowledge as users’ knowledge of punc- /open
regards punctual habits and tual behavior and how to ended
effective time-keeping dur- achieve it? comments
ing the period of system
use.
Primary task support. Predominantly, the respondents indicated that the app did
offer encouragement and support with the primary features of reduction and self-
monitoring, thus yielding a positive response on the Primary task support fea-
tures. In the open comments the respondents acknowledged that setting specific tasks
helped in seeing time allocation in more concrete terms and to think about what activ-
ities need longer or shorter time allocation. Long-term benefits of task setting as a
means of observing time use were noted upon, but there was also concern that setting
tasks daily for a longer time might become tedious.
Computer-human dialogue support. All but one participant stated that the remind-
ers from the system were helpful in performing the required actions (assessments,
reflection). The remaining one expressed a neutral stand on this issue. The division
was stronger for the users’ assessment on whether they were able to learn meaningful
things about timely behavior from the app: all but one agreed (or strongly agreed),
where one respondent strongly disagreed. For this respondent the reminder was an
irritant and did not have any function in supporting the main task. Therefore, the
overall response to the computer-human dialog support features in the app was posi-
tive, but the potential of reminders and notifications to irritate users should not be
ignored.
Desirability test. Each participant selected five adjectives from a large, pre-set group
of 118 adjectives. Of the total 25 selections done by the five respondents in the pre-
sent evaluation three words were selected by four respondents: usable, personal, and
clear. Innovative was selected three times; helpful and accessible were selected twice.
Relevant, ordinary, effective, creative, complex, and boring were all selected once.
13 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
While usable, personal and clear are positive attributes for the app, as are helpful,
accessible, relevant, effective and creative, we can see that these positive impressions
on the use and novelty of the system are also juxtaposed with the minority assess-
ments of ordinary, complex, and boring (design aesthetics elements, largely).
To summarize the results, the features from Primary Task Support category (reduc-
tion, self-monitoring, and rehearsal) and Dialog Support (reminders, suggestion) offer
encouragement, support and increased knowledge and awareness in the problem do-
main (that is, higher elaboration [9]). Responses also included concerns over potential
irritation that can arise from reminders and notifications, and that continued daily use
might become “tedious” for some users. Based on the desirability test and open com-
ments, one area for improvement for the system would be the design aesthetics.
7 Conclusion and Discussion
This paper presented the process of developing a persuasive mobile app based on PSD
principles of Persuasion Context analysis leading to selection of persuasive features of
the system. The evaluation of the produced app focused on the match between select-
ed features and perceived persuasiveness.
We learned that using the PSD model is a feasible means of defining and describ-
ing persuasive software features that match the problem domain. The implications
from the development project and this finding are that following a systematic analysis
process of the Persuasion Context we can design systems in a way that link intention
to outcome in a traceable manner that also allows analysis not only to see whether the
system in general affects the outcome, but more precisely which features contribute to
the behavior change outcomes and how. Such increased detail in the description of
persuasive information systems offers a means of advancing technology development
and helps the field in establishing visibility and transparency as regards the technolog-
ical component in the behavior change process.
As for future research and development of the app, the unobtrusiveness of the sys-
tem interaction (whether system messages and requests come at a user-defined time or
at random times) has been studied [15]. Further research into the effectiveness of this
app would have to include a more longitudinal approach with participants who admit
to having time management issues and who wish to address these issues in practice.
Further development of the app would most likely be based on how the research ad-
vises the selection and refinement of the persuasive feature set of the app.
References
1. Oinas-Kukkonen, H.: Behaviour Change Support Systems: A Research Model and Agen-
da. In: 5th International Conference, PERSUASIVE 2010, Copenhagen, Denmark (2010).
2. Hevner A. R., March S. T., Park, J., Ram, S.: Design science in information systems re-
search. MIS Quarterly, 28(1), 75-105 (2004)
14 Fourth International Workshop on Behavior Change Support Systems (BCSS’16):
Building an App for Behavior Change: Case RightOnTime
3. DeLonzor, D.: Never be Late Again: 7 Cures for the Punctually Challenged. Post Madison
Pub (2003)
4. Tice, D. M., Baumeister, R. F.: Longitudinal study of procrastination, performance, stress,
and health: The costs and benefits of dawdling. Psychological science, 454-458 (1997)
5. Oinas-Kukkonen, H., Harjumaa, M.: Persuasive systems design: Key issues, process mod-
el, and system features. Communications of the Association for Information Systems,
24(1), 28 (2009)
6. Vozza S.: 4 Habits Of Punctual People.
http://www.fastcompany.com/3025074/work-smart/4-habits-of-punctual-
people (2014)
7. Oinas-Kukkonen, H.: A foundation for the study of behavior change support systems. Per-
sonal and Ubiquitous Computing, 17(6), 1223–1235 (2013)
8. Ploderer, B., Reitberger, W., Oinas-Kukkonen, H., van Gemert-Pijnen, J.: Social interac-
tion and reflection for behaviour change. Personal and ubiquitous computing, 18(7), 1667-
1676 (2014)
9. Petty, R.E., Cacioppo, J.T.: The Elaboration Likelihood Model of Persuasion. Advances in
Experimental Social Psychology, Vol. 19 (1986)
10. Alahäivälä, T., Oinas-Kukkonen, H., Jokelainen, T.: Software Architecture Design for
Health BCSSs: Case Onnikka. In: Berkovsky, S. and Freyne, J. (Eds), Persuasive 2013,
LNCS 7822, pp. 3-14, 2013. Springer Verlag Berlin Heidelberg.
11. Lehto, T., Oinas-Kukkonen, H.: Explaining and predicting perceived effectiveness and use
continuance intention of a behavior change support system for weight loss. Behaviour &
Information Technology, 34(2), 176-189 (2015)
12. Basili, V.R., Caldiera, G., Rombach, H. D.: Goal question metric paradigm. In: Encyclo-
pedia of Software Engineering, pp. 528-532. John Wiley & Sons, Inc. (1994)
13. Benedek, J., Miner, T.: Measuring desirability: New methdods for evaluating desirability
in usability lab setting. In: Proceedings of Usability Professionals Association, 8-12.
(2003)
14. Barnum, C.M., Palmer, L.A.: More than a feeling: Understanding the desirability factor in
user experience. In: CHI’10 Extended Abstracts on Human Factors in Computing Systems,
4703-4716 (2010)
15. Tikka, P., Oinas-Kukkonen, H.: RightOnTime: The Role of Timing and Unobtrusiveness
in Behavior Change Support Systems. In: A. Meschtscherjakov et al. (Eds.):
PERSUASIVE 2016, LNCS 9638, pp. 1–12, (2016)