Content uploaded by Roberta Tassi
Author content
All content in this area was uploaded by Roberta Tassi on May 02, 2020
Content may be subject to copyright.
HAL Id: hal-01901090
https://hal.archives-ouvertes.fr/hal-01901090
Submitted on 22 Oct 2018
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Digital Methods for Service Design
Roberta Tassi, Agata Brilli, Donato Ricci
To cite this version:
Roberta Tassi, Agata Brilli, Donato Ricci. Digital Methods for Service Design: Experimenting with
data-driven frameworks. ServDes2018 - Service Design Proof of Concept, Jun 2018, Milano, Italy.
Service Design and Innovation Conference, 2018. �hal-01901090�
!
ServDes2018 - Service Design Proof of Concept!
Politecnico di Milano!
18th-19th-20th, June 2018
Digital Methods for Service Design
Experimenting with data-driven frameworks
Abstract
From logs and information left in online spaces to data points self-generated by connected
devices, digital traces have become more and more diffused over the past years. Along with
some big-data approaches, Digital Methods of research - treating the actual content of users’
manifestation online (i.e. tweets, Instagram pictures, comments) - offer the opportunity to
better understand people and behaviors through their online activities. This paper
investigates how Digital Methods can be repurposed as a full-fledged approach for the
Service Design practice, by offering a method to outline service design frameworks from a
corpus of web data. This quantitative methods, in combination with the traditional
qualitative approaches, leverage the continuous exchange of information that is happening in
the digital space and suggest the possibility to automate parts of the data collection and
analysis processes in support of service design activities. Grafting on several case studies -
we will explain how Digital Methods could be used to identify and describe a set of personas
by extracting and interpreting data from their online activities, and we will inquire into the
application of the same methodological approach to map other frameworks - such as
experience journeys or system maps - that are critical to Service Design.
KEYWORDS: Service Design; Digital Methods; Personas; Service Design Tools
1.Design disciplines in transition
In the last two decades, we are acknowledging an entire disciplinary field experimenting with
new “ways of thinking and doing” (Manzini 2016) in the face of growing environmental,
technical and political issues in our society (Cross 2011; Ehn et al. 2014). Using design to
address those type of challenges is now a global phenomenon and is raising important
questions around design itself as the discipline seeks to make sense of its new role in the
world (Yee et al. 2013). The Service Design practice is playing an important role in this
transition by offering an approach that helps entire organizations shift towards a user-
centered or customer-centered mindset, and transform the way they offer their services as
well as they way they operate. This transition also leads towards an increasing system-level
thinking, enabling organizations to approach and tackle broader issues (Tonkinwise, 2015).
The design practice utilizes a deep understanding of people and communities to understand
the continuously evolving context, but also need to consider that those communities are
place-based and globally-connected, in a continuous exchange of technology, information
and culture. On one hand this require to expand not just the overall approach and process,
but also to elaborate on the tools and techniques that support our understanding of the user,
the system and the constantly evolving context around them (Ostrom 2015). On the other
hand, more germane fields of study to such complex socio-technical issues and problems
(i.e. Science and Technology Studies, Political Sciences, Media Studies and Public Affairs) are
experimenting new methodologies fitting into the so called digital-turn, expanding the notion
of design research to the online domain. Research approaches based on the digital traces left
over the Web and conducted in the framework of digital (Rogers 2009) and quali-quantitative
methods (Venturini & Latour 2009) are opening the possibility to collect and analyze a
wealth of data to observe and describe such complex environments.The hypothesis of the
present paper is that a promising way to cross these two tendencies is to continue and
reinforce the circulation of approaches and methods between Design and Social Sciences, re-
imagining the use of digital data and methods in specific, controversial and complex Design
Research and Service Design contexts.
2. The evolution of Service Design frameworks
The design of new services is an activity that should be able to link the techno-productive
dimension (What is the realm of the possible?) to the social (What are the explicit areas of
demand and what the latent ones?) and cultural dimension (Manzini, 1993). This definition
given by Manzini at the start of the scientific debate around service design suggested that the
methodological approach of industrial designers should have been expanded in order to
embrace the possibility of designing services (Morelli, 2002). The new practice was asking
designers to deal with an intangible subject matter and a constant need of engaging other
stakeholders in the process: it was not immediately clear what techniques could have better
supported those purposes and several tools have been tried and borrowed from Social
Science, HCI Marketing and Business Management (Tassi, 2008).
!
Figure 1 - The origin of service design tools
In the years, we saw a gradual convergence around a set of frameworks that have become
essential assets for design practitioners, such as the human archetypes or personas to
represent the key user behaviors, the user or customer journey map to describe the
experience of interacting with a service, or the system and ecosystem maps to describe the
wider context and what the different players exchange in the service delivery.These types of
Service Design frameworks have helped bringing the service design approach and user-
centered mindset to organization, and have become an essential component of systemic
transformation processes. However, the expansion of the design scope and context of
action is raising some new methodological questions around methods and tools. The existing
frameworks are mostly relying on a qualitative approach to describe users, experiences and
systems - but if design is transitioning to focusing on a local-based globally-connected
community in continuous exchange of information, values and culture (Tonkinwise, 2015),
we could consider better integrating data coming from all the relevant sources available
nowadays, and embracing that evolving context and dimension with a mixed of quantitative
and qualitative approaches. Furthermore, we could also look at a partial automation of the
data collection and exploration processes to provide new ways to efficiently observe specific
! 2
Authors!
Paper Title !
Linköping University Electronic Press
behaviors over time and reflect on their constant evolution, through self-generated
frameworks.
3. The opportunity to work with Digital Methods
We argue that using digital data and analyse them both qualitatively and quantitatively can be
extremely useful in complex social, technical and economic contexts where design is called to
intervene. Proving this hypothesis requires to address different lingering challenges facing
design theory and practice (The Design Collaborative 2014): How to cope with a
heterogeneous and conflicting spectrum of values and interests? How to collaborate with
other disciplines. How to stabilise specific research methods and protocols? How to test
them in large scale empirical experiments? While we will try to tackle the latter two questions
further in the article, the first two ones are related to the controversial nature of the issues
faced by design intervention. This is the case of all those issues where their very same
definition is questioned by various actors and is redefined by the means of new technologies,
governance settings and social representation (Venturini 2009). To study them, a specific
research methods emerged, called Controversy Mapping (Latour 2007). Controversy
Mapping proposes a data-acquisition protocol drawn on the theories and practices of Digital
Methods of research (Rogers 2009, 2013). They exploit the wide range of traces that are left
on the Web by the very actors of the issue under analysis. Digital methods further a social
research approach taking advantage of the empirical capacities embedded in online activities
(Schneider & Foot 2004) with their unique dynamic nature - a mixture of ephemeral and
permanent elements - (Hewson 2003). Digital Methods differ from the big-data research
programs. The emphasis of Digital Methods is not in the magnitude of digital data analyzed
but in the critical affordances deployed by the data-acquisition protocol. Digital Methods
protocols are deriving significant findings from relatively small, ad-hoc designed, data-sets
(Marres and Weltevrede 2013). By following a series of iterative steps and refinement
procedures, the final formatted data are carved out of informational disarrays and unformed
mass of online digital objects. This process has the advantage of avoiding the risk of
projecting pre-existent categories on the issue in analysis. Being a challenge of data selection
and curation, they provide, through the redaction of their protocols, an evident and traceable
inspection of the qualitative decisions taken to compose datasets and corpora.
4. Data-Driven Personas case studies
To propose an example of fertile synergy between Service Design and Digital Methods and
set the ground for a replicable empirical methodology, a full scale test has been conducted by
focusing on one specific design framework. The initial hypothesis for this research is that it is
possible to understand the behaviors, needs and expectations of the users we are designing
for by collecting and studying their online traces. Distilling relevant information out of these
online traces can lead to the identification of clusters of users to be then described as
personas.
4a.The process of distilling Data-Driven Personas
The Data-Driven Personas method could be summarised into four macro-phases (FIG. 2).
! 3
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 2 - The four steps of Data-Driven Personas
1. Data collection: aims at defining the nature and scope of the data harvested as well
as their limitations, by defining the research protocol used to generate the corpus of
data.Before the harvest starts, a deep reflection is required concerning what data
could be relevant for the exploration and when they should be collected. Similarly to
the moment in which service designers and researchers craft a research plan to set
up some interviews with users or observation sessions, the first step is the
identification of a relevant space in the online context where users are discussing a
certain topic. A deep understanding of the bias induced by the kind of user traces
(e.g. hyperlinks, tags and hashtags, threads, ranks or edits) and by the kind of
platforms that are offering them (e.g. Facebook, Twitter, Wikipedia, blogging
platform, search engines) is needed. According to the specific theme and objective
of the exploration, the protocol could for example rely on observing how people
generally talk about a topic on a specific social network, versus look at search results
in the existing engines.
2. Exploration of the discursive space: aims at finding an entry point to analyse the
investigated topic, displaying the constellation of debates emerged from the
collected data, and highlight research insights and patterns.The corpus of data
extracted from the web is visualised in order to get a synoptic view on the issue
under analysis and identify the main components. Similarly to the synthesis moment
in which service designers or researchers start mapping their data point and insights
in order to quickly identify affinities and patterns, this visualisation is aimed at
providing an overview of all the data points collected, and start explore them. For
example, this is the phase that allows to start identifying the clusters of people who
relate to the investigated theme in the same way, and can lead to outline a first set of
personas (Cooper, 2014).
3. Patterns and insights refinement: aims at iterating on the emerging clusters of
information in order to enrich their understanding and further detail the research
insights.While the exploration of the discursive space allows to quickly highlight the
most relevant topics and clusters composing a complex issue, through a deeper
analysis it is possible to closely identify the cohesive groups of users and needs
behind those clusters, refining the set of personas and enriching their understanding.
To achieve this objective, it is necessary to detect some distinctive features
characterizing each cluster, by analysing the verbal space or visual imagery that are
associated to their debates.
4. Mapping and storytelling: aims at outlining the complete description of each
persona, making use of the most relevant qualitative and quantitative aspects related
to the data emerged during the research process.Vivid descriptions of user types
enable to bring fictional user profiles to life (Cooper, 1998; Grudin & Pruitt, 2003)
and - as regards to Data-Driven Personas - this can be done by using the data
generated by users during their online activities. An advantage of the Data-Driven
method is that researchers can build the profiles in a semi-automatic way by using
existing information to narrate the various aspects of the user profile. At this stage,
! 4
Authors!
Paper Title !
Linköping University Electronic Press
data visualisation shifts its objectives from an analytical research tool to a way of
synthesising the main aspects of each persona - or eventually their experience
journey.
Figure 3 - A representation of the process used to distill Data-Driven Personas
4b. Case Study: Naturpradi
NATURPRADI is a research project aimed at observing and describing the effects of the
many initiatives endorsed by the Paris municipality to revegetate the city. These initiatives are
trying to produce smart solutions to a growing range of issues created by urban growth.
Nevertheless, there is no agreement on the imaginaries and technical practices that should be
included into this new urban nature (Gandy 2006).To observe, monitor and, eventually,
produce elements of reflections for future urban policies, the NATURPRADI project is
mapping the symbolic and material elements of the urban nature debate (Ricci et al. 2017).
The research project is aimed at exposing the different social, political and technological
issues associated to urban nature, its actors, and the controversies caused by alignment and
misalignment of interests.To achieve its objective, NATURPRADI started a Digital Method
campaign by collecting digital-native content produced on Twitter. The online news and
social networking platform has been chosen since it is broadly used by a variety of actors
getting spontaneously organised around discussion topics by using hashtags. Twitter,
presenting the concrete opportunity for “empirical sociocultural research” (Burgess and
Bruns 2015), has become, over the years and despite its transformation, an object of study
and a data source for research scopes (Rogers 2013b).The core of the NATURPRADI
project is to investigate and elicit different viewpoints and perspectives on urban nature, how
they are sustained by specific communities and populated by identifiable users, each of them
proposing an instantiated vision of the future vegetation in Paris. For this reason,
NATURPRADI has provided the great opportunity to test a new process moving from user-
generated content to personas, following the four-step process defined above.
1. Data Collection !
After having chosen Twitter for collecting the manifestations of interest, the
! 5
Authors!
Paper Title !
Linköping University Electronic Press
Streaming API has been adopted to retrieve online data related to the topic of
1
nature in Paris. This process requires to acknowledge and consciously embrace
Twitter limitation , specificity and embedded politics (Gillespie 2010; VanDijck
2
2013), as they are technically, rhetorically and culturally expressed (Gillespie 2014):
only their clear understanding allows later on in the process to mitigate and validate
the results of the research. Among the different approaches for Twitter corpora
building (see Mayr and Weller 2017) it has been chosen one based on key expression
query. Through a collaborative and participatory procedure among the members of
the NATURPRADI consortium, a list of 158 expressions (FIG. 4) has been used to
capture the tweets in which they were mentioned. To assure a territorial specificity to
our corpus we queried only for French word. Furthermore, all the keywords were
queried by adding the word “Paris” .
3
!
Figure 4. The set of keywords used for the data collection on Twitter
2. Data Exploration !
The data collected were used to generate a series of graphs showing how different
users are connected to each other and the specific words used in the tweets. In the
map, we can recognize clusters of conversations, such as the central institutional
cluster featuring linked to the municipality (@Paris, @Anne_Hidalgo,
@PKOMITES and @vegetalisons). On the left of this group, there is a cluster
concerning urban farming and bio-agriculture activities, where we see for example
the initiative of the municipality named Parisculteurs. On the right of the
institutional accounts, lays a cluster about participative initiatives deployed by the
municipality of Paris to promote the citizens’ engagement, like the Permis de
Végétaliser. Another relevant cluster is located under the institutional accounts,
featuring ecology and recycling related topics. Looking at the right edge of the
discursive space, there is the cluster where all the famous green areas of Paris are
mentioned. Finally, at the opposite edge, can be detected the cluster about
innovative agricultural techniques and startups.On the basis of this quali-quantitative
interpretation of the graph, sustained by a visual analysis of the network, five key
clusters can be identified (FIG. 5) :!
These API offers the possibility to retrieve only live data, imposing a bandwidth limita:ons coming into effect when the
1
requested tweets exceed the 1% of the all traffic flowing in the plaDorm.
A limitation affecting Twitter based researches is linked to its representativeness (see Blank 2016). Although
2
Twitter is widely used all across the world, its adoption rate changes accordingly to different social milieux and
the way it is used may differ significantly from country to country. In the NATURPRADI project there is no
assumption about the possible exact extension of the observed digital population to the general one.
To assure that the final corpus would not been biased by tweets not related to Paris or to the urban nature, a
3
further curatorial procedure has been applied. Through a custom and open-source software (the source code is
available here: https://github.com/medialab/catwalk.), every tweet has been read by the research team and
evaluated in terms of its pertinence. This approach, distinguishes the NATURPRADI project from many other big
data ones. Furthermore, the close reading of the tweets enabled us to have a constant overview of the state of
the discussion, gaining a deep understanding of the dynamics of the issue. This aspect resulted to be extremely
useful in the analysis and interpretation of the data.
! 6
Authors!
Paper Title !
Linköping University Electronic Press
!
A. Technological Development, featuring innovation initiatives and project in the
agricultural field;!
B. Urban-Agriculture, featuring bio-agricultural projects developed inside the city of
Paris;!
C. Co-design of Public nature, featuring all the debates around the participatory
activities endorsed by the municipality;!
D. Ecological attitude, featuring the concerns about the ecological transition like the
domestic recycle of wastes.!
E. Relaxed Contemplation, featuring the discussion about outdoor activities around
Parisian gardens.!
!
The exploration of the discursive space and the description of the key-clusters
allowed to identify the most vibrant and relevant topics for the users. But who are
the groups of people populating those clusters? And what is their approach towards
urban nature? Drawing upon this overlapped delimitations, the process has moved
towards a more precise description and characterisation of the users present in each
cluster.
!
Figure 5 - The user-object networks
3. Clusters Refinement !
For each cluster, a list of keyword has been produced to identify the users belonging
to one or more of them. For example, a user is identified as part of the Ecological
Attitude group whenever he or she used at least one of the keywords compost,
ecologique, dechet, tri. If that same user wrote any of the keywords related to the
other clusters, she would also appear in those communities (Fig. 6). The so obtained
corpus is then used to understand if, besides debating about the same topic, they
also debate in a similar way. Just like when the researcher carries out field
! 7
Authors!
Paper Title !
Linköping University Electronic Press
investigations to collect more insights about how people live, the visualisation and
interpretation of different aspects of their online activities allows to progressively
validate the cohesion of communities. This iterative process consists in visualising
the multiple dimension of the corpus (e.g. images, texts, links) and then interpreting
the results to understand if there are similar groups which can be merged together
and considered as a unique behaviour or, on the contrary, if inside a cluster more
than one distinctive behaviour can be discerned. In our test, we have focused on the
two main elements of a tweet, its textual content and the possible images attached
to it.
!!
Figure 6 - The selected keywords for each cluster used to retrieve the
communities of users which used these words in their tweets
Digging into the textual sphere
Analyzing the most frequent and relevant vocabulary elements that a group of
people use to discuss a given issue is extremely relevant to identify both the
commonalities and distinctive traits of each cluster. In our test, the visualisation of
the textual sphere shows the 150 most recurrent terms for each group, sorted from
the most to the least frequent (FIG. 7). The size of each bubble is proportional to
the frequency of the word. The color of the bubbles describes how much each word
is shared with other clusters: the lightest the color, the most shared the word; the
darkest the color, the least share the word - which means that it is uniquely used by a
specific community. While the most used words by the Technological development,
Urban agriculture and Co-design of public nature have proved to be in accordance
with the initial depiction of these communities, the interpretation of this
visualization lead to an interesting observation concerning the Relaxed
contemplation and the Ecological attitude communities. In the Relaxed
contemplation cluster, the names of several famous French photographers occurred
among the most frequent words. This could suggest the presence of a smaller
community within that cluster, with an interest in photography and in the historic
representation of the city. Whereas the Ecological attitude community seemed
mostly linked to the recent news of the urinal-vases installed in Paris by the
municipality, and their textual sphere appeared similar to the vocabulary used by the
Urban-agriculture community: this suggested that the two clusters share parts of
same debates and are likely representations of the same attitude.
! 8
Authors!
Paper Title !
Linköping University Electronic Press
!
!
Figure 7 - The textual sphere visualisation shows the 150 most used
terms for each community
Digging into the visual sphere
Analyzing the images that are produced and shared by the users enables a quick
introduction to the imagery of each cluster. For example, by interpreting the visual
elements, we understood that the Relaxed contemplation cluster mostly share
content about the most famous Parisian architectures and green areas (Fig. 8) . This
seems to be coherent with the fact that this cluster is composed mainly by tourists
! 9
Authors!
Paper Title !
Linköping University Electronic Press
and those who appreciate the aesthetic role of the Parisian nature. In the lower part
of the network there is a significant group of historic images, which corroborates
the presence of a sub-community of Nostalgic users. Repeating the process for the
other clusters, the visual sphere analysis helps to understand the cohesion of the
identified groups (Fig. 9-12) . The Start-up entrepreneur works on agricultural
research and innovation, the Sustainability aware consumer is interested in locally
grown and produced products, the Overactive neighbor participates in every
municipality greening initiative, the Forever tourist always looks at paris with
enchanted eyes, the Nostalgic remembers the better time of Parisian nature with a
bitter smile (FIG. 13).
Figure 8 - The visual sphere of the “Relaxed contemplation” community
! 10
Authors!
Paper Title !
Linköping University Electronic Press
Figure 9 - The visual sphere of the “Technological development” community
!
Figure 10 -. The visual sphere of the “Urban-agriculture” community
! 11
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 11 - The visual sphere of the “Co-design of public nature” community
!
Figure 12 - The visual sphere of the “Ecological attitude” community
! 12
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 13 - The analysis of the textual and visual imagery of each community allows to
individuate the unique behaviours of the research
4. Mapping and Storytelling !
Finally, each persona has been outlined with the data produced by the group of
users from whom that personas was created. The narration starts with the picture
and the name randomly picked from the real users belonging to that persona. The
keywords which initially brought to the definition of the community and then of the
relative persona are listed as the most connoting hashtags. A tweet has been selected
from the data corpus, in order to represent the usual way that persona would talk
about urban nature in Paris (the persona’s quote or motto). The Twitter descriptions
of users are used to narrate how each persona would describe themselves: a bubble
chart visualise the most occurred terms. The most recurring images of each
personas are also part of the narration, showing their visual imagery. The
relationship between personas, as well as their similarity, is represented by a diagram
showing how many users are unique to that personas and how many are shared with
other personas, since a user could be present in more than one cluster. Each persona
can be also located on the initial map of the overall discourse, telling us if - in their
relationship with nature - they show a more contemplator or expert approach.
Finally the tweets activity over time of each personas can help understand the
engagement with the topic (FIG. 14-18). The final personas could be used as a
starting point to imagine different services that the municipality of Paris could
propose - related to nature - or different needs concerning the existing services.
! 13
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 14 - Description of the “Start-up entrepreneur” persona
! 14
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 15 - Description of the “Overactive neighbor” persona
! 15
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 16 - Description of the “Sustainability aware consumer” persona
! 16
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 17 - Description of the “Forever tourist” persona
! 17
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 18 - Description of the “Nostalgic” persona
4c. Case Study: Co-design workshop on gender violence
! 18
Authors!
Paper Title !
Linköping University Electronic Press
In the context of the World Usability Day hosted in Rome in November 2017, the same
approach piloted during the NATURPRADI project has been adopted to facilitate a
collaborative workshop around the topic of Data-Driven Personas. The topic we approached
this time was the one of gender violence in the Italian context. The research question we
asked ourselves was how we could map all the different attitudes towards the topic of gender
violence in order to identify opportunities to design new information services that could
reach a multitude of subjects. This project offered the opportunity to validate and refine the
approach introduced with NATURPRADI, by having other designers playing with the same
method during the workshop.
1) Data Collection!
As for the NATURPRADI project, Twitter was identified as the appropriate space
for the analysis since in the Italian context it collects a lot of opinions on this topic,
in relationship to specific news and public debates. A set of keywords were
identified (FIG. 19) in order to obtain the data corpus, following exactly the same
strategy adopted during the previous experimentation. The data collection took
place during the period of October 1st to 15th, in 2017. !
!
Figure 19 - The set of keywords used for the data collection on Twitter
2) Data Exploration!
The overall map representing the online discourse on Twitter was used as a starting
point for the collaborative session. Participants were asked to look at the map and
highlight the key clusters they could recognize in the visualization, based on the
affinity of themes and keywords used to talk about gender violence. The clusters
covered a wide range of perspectives (FIG. 20):!
!
A. The activist community, featuring the organizations and associations that belong
to the Italian feminist movement, who exclusively talk about violence against women
and gender rights.!
B. The area of short-term reactions, linked to single episodes coming out in the
news, which tend to disappear fast, till the moment in which the next violence-
related fact happens.!
C. The anti-migrant space, where the vocabulary associated to gender violence is
blended with racism and hanger against foreigners.!
D. The area of social and ethical reflections, where we can find a generic variety of
conversations around the topic, featuring media and intellectuals.!
E. A small area of institutional prevention, giving space to the institutions that are
trying to tackle the violence issue through education programs and awareness
initiatives.
! 19
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 20 - The tweets user-object networks
3) Cluster Refinement!
In a second moment, participants were asked to filter a cluster of reference and start
digging into it. By qualitatively reading some of the tweets and checking some of
the relevant twitter profiles emerging from that specific debate, they could
understand if the cluster was really sharing the same approach or if it was hiding
different nuances. Based on that they could decide whether to map one or more
personas within each group, and start characterizing that persona (FIG. 21) with an
identifying attribute and quote.
!
!
Figure 21 - A first description of the identified persona
4) Mapping and Storytelling!
Finally, each persona could be described by combining a set of automated
! 20
Authors!
Paper Title !
Linköping University Electronic Press
information extracted from the corpus (e.g. random image profile, tag cloud of the
profile description, most used keywords, most used images) and some qualitative
information derived from the understanding developed by the team during the
exercise (e.g. a qualitative description of the behaviour, needs and challenges of that
persona). By putting together all the results obtained by the different teams during
the workshop session, we had the map of seven key personas (FIG. 22-23).
!
Figure 22 - Personas description
!
Figure 23 - The process of ideas generation, starting from the persona identified
4d. Case Study: Glamour
! 21
Authors!
Paper Title !
Linköping University Electronic Press
The initial corpus of data used to distill the Data-Driven Personas doesn’t necessarily need
to come from social networks. Web analytics could provide interesting insights around user
behaviours, as well as a dataset built ad hoc through a quantitative survey or diary study.
In this last case study we worked with the Glamour team of Condé Nast to setup a model
aimed at observing user behaviours with a mixed qualitative/quantitative research approach.
We first looked at their web analytics, which were showing clear patterns in terms of readers’
engagement and reaction to the content offered by the online magazine. It was easy to detect
different types of behaviours by collecting and aggregating that information, but still hard to
understand the motivations, needs and wishes behind those behaviours. That has led to the
decision of adding another step of research, by distributing a dedicated diary study to be
filled by a large number of readers - with the aim of collecting photos and comments in
answer to specific research questions, and use that material to look deeper into the clusters
emerged from the initial analysis of web data.
1) Data Collection!
The analysis of patterns related to the interaction with app and web content has
initially lead to the identification of several clusters, representing the different levels
of users’ engagement in relationship to the type of content they consume. The
clusters helped screen for a sample of research participants (50 in total), with the
objective to ask them to share their experiences, stories and desires through a digital
diary. The diaries generated a large amount of information (photos and comments),
that was analysed through visualisations and maps in order to see differences and
similarities in the response of the participants. !
2) Data Exploration!
The first step was analysing all the comments in relation to the specific topics
discussed in the diary. For example, they were invited to talk about their idea of
make-up and their make-up routine, and we mapped all the users and words used to
describe those aspects. They were also invited to share their idea of self-care and
their self-care habits, and we mapped all the users and words used to describe those
aspects in a second visualisation. By looking at those two maps, we started to see
some emerging clusters, in particular a clear distinction among all the women who
interpreted those topics more on the makeup and aesthetic side of beauty, versus all
the women who talked more about wellness and beauty as an intimate way of taking
care of themselves (FIG. 24). We could also see the groups of users talking about
these topics in a very general way, and the groups of users characterized by their
own vocabulary developed thanks to a deeper interest and expertise in the topic.!
! 22
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 24 - The user-object network about the concept of make-up
3) Cluster Refinement!
In a second moment we looked at all the images (we had a total 1950 images took by
the participants) in order to understand if the clusters identified through the initial
semantic maps were cohesive, and start understanding more about each of them
(FIG. 25). The refinement was conducted by simultaneously looking at the digital
map and all the printed photos sticked on the walls, to immediately double-check the
interpretation of the visual map with the entire set of materials produced by each
cluster of participants. The refinement exercise has led to the identification of seven
main personas.
! 23
Authors!
Paper Title !
Linköping University Electronic Press
!
Figure 25 - The images produced by the users to represent their concept of beauty
4) Mapping and Storytelling!
Finally, the personas were described by extracting some of the images and quotes
from the diary of each cluster. In this case we favored a curated selection of that
material (instead of an automatic extrapolation of those items) because we wanted
to make sure the outcome was visually capturing the essence of each persona in an
appealing and exhaustive way (Fig. 26).
!
Figure 26 - The “State of Beauty” personas
! 24
Authors!
Paper Title !
Linköping University Electronic Press
At the very end of the process, a final round of individual interviews was conducted on a
small selection of participants (one person per each type of persona) leading to an enriched
understanding of their needs and motivations and to the closure of the study. This approach
suggests the possibility of an integration of Digital Methods into a more traditional
ethnographic research process. The final outcome was a set of personas as well as a
replicable model that allows to bridge a quantitative and qualitative understanding, and to
leverage web analytics to structure an efficient ethnographic field-work.
5. Limits and opportunities
The new Data-Driven personas method aims to expose how Digital Methods can be
integrated into design practices, deploying new “techniques which communicate, interact,
empathize and stimulate the people involved” (Giacomin, 2014). Digital Methods could
allow to scale up the magnitude of data and information collected. The proposed approach
offers significant advantages in terms of time and costs, if compared to traditional
qualitative research and analysis techniques: it allows to quickly collect and analyse a wide
dataset and develop key insights even before activating the field-research and start investing
on it. Nevertheless, there are some activities in the process that shouldn’t be underestimated,
such as:
●Data collection: setting up the necessary infrastructure for collecting data might take
some time. Depending on the scale of the data to be collected, simple solutions like
storing it into spreadsheets or plain text files (i.e. CSV) might be not appropriate and
the setup of a proper and efficient database might be required. Furthermore, the
API provided by the digital platform, as well as the interface through which the data
might be scraped, tend to change rapidly. This may affect the quality of the
harvesting, or at least, require a continuous monitoring, tracking and adjustment of
the collection procedures. Working with digital data means to respect the ever-
changing privacy policies and terms of use of the platforms involved in the
research. Along with the respect for these standards, an ethical reflection on how to
handle personal identifiable information is always needed.
●!
Data cleaning: in some cases sorting noise out of the stream of data collected can
be done in a quick way (e.g. filtering out objects that are less frequently encountered,
or conversely, the ones that are mentioned too much). In other cases, as for the
NATURPRADI project, a careful reading of the collected data is necessary.
Regardless the specific strategy adopted for reviewing and cleaning the dataset, a
constant control of the data harvested is always necessary. While this operation
helps the researcher explore the material they are going to work with, it also requires
the setup of an appropriate infrastructure (i.e. from generating reports containing
random samples of the data to reading the single data points one by one).!
●Data visualisation: distilling information out of a dataset is more and more simple
thanks to the growing numbers of techniques, libraries and software. Nevertheless
the ultimate scope of the visualization, exploratory visualisations and procedures are
needed in order to continuously offer different views on the data through multiple
and non-exclusive visual models, especially in the first part of the process. The
production of interactive visualisation should be preferred to the static ones, to
better support the exploration of the views and ease the identification of key
insights and learnings.!
●Data interpretation: what Digital Methods offers is a better understanding of the
user space and a reduction of the risk of projecting pre-existing categories, or
missing unknown parts of the debate. The insights and clusters that emerge during
this type of analysis need to consider possible limitations and bias. Collecting data
! 25
Authors!
Paper Title !
Linköping University Electronic Press
over social media, for example, implies to cope with the digital divide issue and with
the different platform culture (e.g the more or less marked propension to use
hashtags) that might be present in different geographical location. The outcome of
this type of analysis needs to be seen as part of a wider range of research and
design thinking activities aimed at understanding the people and context of use. The
insights can be used for example to inject new hypothesis in a user research phase,
or to enrich the existing knowledge with a different perspective.
Digital Methods need to be considered an opportunity in integration, and not in
replacement, of current service design tools and techniques. For example, along the process
of creating Data-Driven Personas, the researchers may use the emerging clusters as a way to
define potential participants for a set of in-depth interviews. The interviews will provide
both a validation of the analyses carried out previously and, above all, add a deeper
qualitative layer to the understanding of the different personas. Following this example,
Data-Driven Personas can ease the preparation of a field-research, by raising important
themes upfront and offering an alternative strategy to recruit research participants.
The Data-Driven Personas protocol is a first attempt to provide a sustainable and replicable
approach to effectively apply Digital Methods to support the service design process. This
protocol is applicable to all those cases where the research environment involves a
community of users who drive a series of debates inside an online space. Other protocols
could be explored in the future, investigating the possibility to derive other key Service
Design frameworks - such as experience journeys and system maps - from the analysis of the
online discourse. The expansion towards additional frameworks would require to think about
spaces where to find that information, ways of clustering the data collected and strategies for
refinement and visualization, by relying on a four-steps model similar to the one introduced
and discussed in this paper.
References
Berry, D. M., Borra, E., Helmond, A., Plantin, J.-C., Walker Rettberg, J., & Walker, J.
(2015). The data sprint approach: exploring the field of Digital Humanities through
Amazon’s application programming interface. Digital Humanities Quarterly, 9(4).
Blank, G. (2016). The Digital Divide Among Twitter Users and Its Implications for Social
Research. Social Science Computer Review. http://doi.org/10.1177/0894439316671698
Blyth, S., & Kimbell, L. (2011). Design Thinking and the Big Society: From solving personal
troubles to designing social problems An essay exploring what Design can offer those working on social
problems and how it needs to change. London.
! 26
Authors!
Paper Title !
Linköping University Electronic Press
Broberg, O., Andersen, V., & Seim, R. (2011). Participatory ergonomics in design
processes: The role of boundary objects. Applied Ergonomics, 42(3), 464–472. http://
doi.org/10.1016/J.APERGO.2010.09.006
Brown, T. (2008). Design Thinking. 2008. Harvard Business Review, 6.
Brown, T. (2009). Change by Design: How Design Thinking Transforms Organizations
and Inspires. Innovation Harper Business. New York, USA.
Bruns, A., & Burgess, J. (2015). Twitter hashtags from ad hoc to calculated publics. In N.
Rambukkana (Ed.), Hashtag publics : the power and politics of discursive networks (p. 293).
Cooper, A., Reimann, R., Cronin, D., & Noessel, C. (2014). About face: the essentials of
interaction design. Indianapolis: Wiley.
Cooper, A. (2004). The inmates are running the asylum:[Why high-tech products drive us crazy and
how to restore the sanity]. Indianapolis, IN: Sams.
Cross, N. (2011). Design thinking: Understanding how designers think and work. Oxford: Berg.
Diana, C., Pacenti, E., & Tassi, R. (2012, September). Visualtiles: Communication tools
for (service) design. In Conference Proceedings ServDes. 2009; DeThinking Service; ReThinking
Design; Oslo Norway 24-26 November 2009 (No. 059, pp. 65-76). Linköping University
Electronic Press.
Ehn, P., Nilsson, E. M., & Topgaard, R. (2014). Making futures : marginal notes on innovation,
design, and democracy. Cambridge, Mass.: The MIT Press.
Gandy, M. (2002). Concrete and Clay. Cambridge, Mass.: The MIT Press.
Gandy, M. (2006). Urban nature and the ecological imaginary. In N. Heynen, M. Kaika, &
E. Swyngedouw (Eds.), In the Nature of Cities: Urban political ecology and the politics of Urban
Metabolism (pp. 63–75). New York: Routledge.
Giacomin, J. (2014). What Is Human Centred Design? The Design Journal, 17(4), 606–623.
http://doi.org/10.2752/175630614X14056185480186
Gillespie, T. (2010). The politics of “platforms.” New Media & Society, 12(3), 347–364.
http://doi.org/10.1177/1461444809342738
Gillespie, T. (2014). The Relevance of Algorithms. In T. Gillespie, P. J. Boczkowski, & K.
A. Foot (Eds.), Media technologies: essays on communication, materiality, and society. Cambridge,
Massachusetts: MIT Press.
Gothelf, J., & Seiden, J. (2016). Lean UX: Designing Great Products with Agile Teams. O'Reilly
Media, Inc.
Hanington, B. (2003). Methods in the Making: A Perspective on the State of Human
Research in Design. Design Issues, 19(4), 9–18. http://doi.org/
10.1162/074793603322545019
Hewson, C. (2003). Conducting research on the Internet. The Psychologist, 16(6).
Ingold, T. (2015). Design Anthropology is not and cannot be Ethnography. In Research
Network for Design Anthropology.
Kimbell, L. (2015). Title Applying Design Approaches to Policy Making: Discovering Policy Lab
Type Report. Brighton.
Latour, B. (2007). La cartographie des controverses. Technology Review.
Le Dantec, C. A. (2016). Designing publics. Cambridge, Mass.: The MIT Press.
! 27
Authors!
Paper Title !
Linköping University Electronic Press
Macnaghten, P., & Urry, J. (1998). Contested natures. London: Sage.
Manzini, E. (1993). Il Design dei Servizi. La progettazione del prodotto-servizio, in “Design
Management, (7).
Manzini, E. (2016). Design Culture and Dialogic Design. Design Issues, 32(1), 52–59.
http://doi.org/10.1162/DESI_a_00364
Marres, N., & Weltevrede, E. (2013). SCRAPING THE SOCIAL? Journal of Cultural
Economy, 6(3), 313–335. http://doi.org/10.1080/17530350.2013.772070
Mauri, M. (2014). An interview with Richard Rogers: repurposing the web for social and
cultural research. Retrieved November 6, 2017, from http://www.densitydesign.org/
2014/05/an-interview-with-richard-rogers-repurposing-the-web-for-social-and-cultural-
research/
Mayr, P., & Weller, K. (2017). Think before you collect: Setting up a data collection
approach for social media studies. In L. Sloan & A. Quan-Haase (Eds.), The SAGE
handbook of social media research methods (p. 679). London.
Morelli, N. (2002). Designing product/service systems: A methodological exploration1.
Design issues, 18(3), 3-17.
Muller, M. J. (2003). Participatory Design: The Third Space in HCI. In A. Sears & J. A.
Jacko (Eds.), Human-computer interaction. Development process (pp. 165–185). Boca Raton: CRC
Press.
Norman, D. A., & Draper, S. W. (1986). User centered system design. New Perspectives on
Human-Computer Interaction. New Perspectives on Human-. Hillsdale: L. Erlbaum Associates.
Ostrom, A. L., Parasuraman, A., Bowen, D. E., Patricio, L., & Voss, C. A. (2015). Service
research priorities in a rapidly changing context. Journal of Service Research, 18(2), 127-159.
Pruitt, J., & Grudin, J. (2003, June). Personas: practice and theory. In Proceedings of the
2003 conference on Designing for user experiences (pp. 1-15). ACM.
Ricci, D. (2010). Seeing what they are saying: Diagrams for socio-technical controversies.
In D. Durling, R. Bousbaci, L.-L. Chen, P. Gauthier, T. Poldma, S. Rowoth-Stokes, & E.
Stolterman (Eds.), DRS2010 - Design & Complexity proceedings. Montreal.
Ricci, D., Colombo, G., Meunier, A., & Brilli, A. (2017). Designing Digital Methods to
monitor and inform Urban Policy. The case of Paris and its Urban Nature initiative. In
3rd International Conference on Public Policy (ICPP3).
Rogers, R. (2009). The End of the Virtual: Digital Methods. Amsterdam: Amsterdam
University Press.
Rogers, R. (2013a). Debanalizing Twitter. In Proceedings of the 5th Annual ACM Web Science
Conference on - WebSci ’13 (pp. 356–365). New York, New York, USA: ACM Press. http://
doi.org/10.1145/2464464.2464511
Rogers, R. (2013b). Digital Methods. Cambridge Mass.: The MIT Press.
Schneider, S. M., & Foot, K. A. (2004). Web Sphere Analysis: An Approach to Studying
Online Action Forthcoming in Virtual Methods: Issues in. In Social Science Research on the
Internet. Oxford: Berg.
Seemann, J. (2012). Hybrid Insights : Where the Quantitative Meets the Qualitative.
Rotman Magazine, 58–61.
! 28
Authors!
Paper Title !
Linköping University Electronic Press
Steen, M. (2012). Human-centered design as a fragile encounter. Design Issues, 28(1),
72-80.
The Design Collaborative. (2014). DesignX: A Future Path for Design - jnd.org.
Retrieved October 1, 2016, from http://www.jnd.org/dn.mss/designx_a_future_pa.html
Tonkinwise, C. (2015). Design for Transitions‒from and to what?. Design Philosophy Papers,
13(1), 85-92.
VanDijck, J. (2013). The culture of connectivity : a critical history of social media. Oxford: Oxford
University Press.
Venturi, G., Troost, J., & Jokela, T. (2006). People, Organizations, and Processes: An
Inquiry into the Adoption of User-Centered Design in Industry. Internation Journal of
Human-Computer Interaction, 2(21), 219–238.
Venturini, T. (2009). Diving in magma: How to explore controversies with actor-network
theory. Public Understanding of Science, 0963662509102694v1.
Venturini, T., & Latour, B. (2009). The Social Fabric: Digital Traces and Quali-quantitative
Methods. In Proceedings of Future En Seine. Paris.
Venturini, T., Munk, A., & Meunier, A. (2016). Data-Sprinting: a Public Approach to
Digital Research. In C. Lury, P. Clough, M. Michael, R. Fensham, S. Lammes, A. Last, &
E. Uprichard (Eds.), Interdisciplinary Research Methods.
Venturini, T., Ricci, D., Mauri, M., Kimbell, L., & Meunier, A. (2015). Designing
Controversies and their Publics. Design Issues, 31(3), 74–87. http://doi.org/10.1162/
DESI_a_00340
Yee, J., Jefferies, E., & Tan, L. (2013). Design transitions. Amsterdam: BIS Publishers.
! 29
Authors!
Paper Title !
Linköping University Electronic Press