Persona Studies 2018, vol. 4, no. 2

ARE PERSONAS DONE? EVALUATING
THE USEFULNESS OF PERSONAS IN

THE AGE OF ONLINE ANALYTICS

JON I SALMI N EN, BE RN AR D J. JANS E N, JISU N AN,
HAE WOON KWA K A N D SOON-GY O JU NG

ABSTRACT

In this research, we conceptually examine the use of personas in an age of
large-scale online analytics data. Based on the criticism and benefits outlined in
prior work and by practitioners working with online data, we formulate the major
arguments for and against the use of personas given real-time online analytics data
about customers, analyze these arguments, and demonstrate areas for the
productive employment of data-driven personas by leveraging online analytics data
in their creation. Our key tenet is that data-driven personas are located between
aggregated and individual customer statistics. At their best, digital data-driven
personas capture the coverage of the customer base attributed to aggregated data
representations while retaining the interpretability of individual-level analytics;
they benefit from powerful computational techniques and novel data sources. We
discuss how digital data-driven personas can draw from technological
advancements to remedy the notable concerns voiced by scholars and practitioners,
including persona validation, inconsistency problem, and long development times.
Finally, we outline areas of future research of personas in the context of online
analytics. We argue that to survive in the rapidly developing online customer
analytics industry, personas must evolve by adopting new practices.

KEY WORDS

Data-Driven Personas; Online Analytics; Customer Segmentation

INTRODUCTION

The abundance of social media data has increased the difficulty of making sense of data
(Järvinen 2016; Saggi & Jain 2018; Salminen, Milenković & Jansen 2017), while at the same time
making it possible to automatically infer customers attributes from social media that were
previously accessible only by survey research. Researchers have begun utilizing publicly
available social media posts to infer customer attributes, including personality traits, political
orientation, brand liking, and needs and wants (Ardehaly & Culotta 2015; Del Vecchio, Mele,
Ndou, & Secundo 2017; Jung, An, Kwak, Salminen, & Jansen 2017; Volkova, Bachrach, & Durme
2016). Free expression in social media provides opportunities to learn about the needs and
traits of groups and individuals (Owusu et al. 2016). Overall, the development of computational
techniques and the availability of online data has resulted in an increased interest in data-
driven personas: (a) to describe the content consumption patterns of diverse online audiences

Salminen et al.

(Salminen, Şengün, et al. 2018), and (b) to use online data, namely online analytics and social
media posts, in persona generation (An, Kwak & Jansen 2017; Zhang, Brown & Shankar 2016).

However, at the same time, the usefulness of personas has been questioned. There is a
plethora of alternative online analytics tools (e.g., Google Analytics, SimilarWeb), services (e.g.,
comScore, HitWise) and metrics (Boghrati et al. 2017; Clarke & Jansen 2017; Järvinen &
Karjaluoto 2015) that one can employ to understand customers. Moreover, companies have
gained access to individual-level data that performs well for many marketing purposes,
including customer relationship management (Zerbino et al. 2018), providing tailored
experiences and recommendations (Ronen, Yom-Tov & Lavee 2016), one-to-one targeting of
online ads (Miralles-Pechuán, Ponce & Martínez-Villaseñor 2018), and enabling the creation of
sophisticated customer segments (Jansen et al. 2017; Rundle-Thiele, Dietrich & Kubacki 2017).
Although prevalent in many fields, the use of automated techniques is especially pertinent in the
field of digital marketing which is shifting toward a higher degree of personalization, micro-
targeting, and one-to-one marketing (Bleier & Eisenbeiss 2015). Thus, there are concerns about
personas providing real value in such an environment.

In this research, we evaluate the use and usefulness of personas given the easy
availability of online individual customer data for digital marketing use cases. Traditionally,
personas are used in replacement of actual one-to-one data about customers (Howard 2015), as
mental models that help keeping customers in mind in the absence of having real customers
available when making decisions about design (Nielsen 2013), software development (Pruitt &
Grudin 2003), and marketing (Russell & Toklu 2011). However, with the widespread availability
of online analytics data and numerous programs, techniques, and platforms to process the
gathered data, a research question arises: Are personas still valid as a marketing tool in the era of
online analytics?

We address this question through a conceptual inquiry and analysis, drawing from two
streams of literature: (a) persona studies (for drawing the benefits and shortcomings of
personas) and (b) digital marketing research (for drawing the use of digital marketing). We aim
to conceptually combine these streams to better understand the role of personas in the era of
online analytics and digital marketing. Therefore, this work is conceptual research discussing the
value of personas in the era of online analytics. Our purpose is to review the related literature for
better positioning personas given the context of online analytics, which is currently lacking in
the extant persona literature. With our inquiry, we aim to avoid the advocacy issue mentioned
by Matthews, Judge & Whittaker (2012, p. 1220): “[persona] literature […] generally takes a
position of advocacy and lacks objectivity.” That is, we explore both the possible strengths and
the shortcomings of personas in an objective manner.

First, we identify the benefits to which personas are associated in the literature. Second,
we summarize the traditional criticisms of personas. Third, we formulate new critical
arguments against the use of personas given the widespread availability of online analytics data.
We then address these arguments, examining their relevance to traditional and digital data-
driven personas, and discuss the ability of technology to solve the major shortcomings of
personas. Finally, we conclude by presenting ideas for the future of personas in the era of online
analytics.

Our key contributions are twofold. First, we present novel criticism relating to the use of
personas for customer-related decision making, addressing concerns from both digital analytics
practitioners and academic scholars. Second, we reposition personas in the age of online
analytics by conceptually distinguishing between traditional data-driven personas and digital
data-driven personas, and analyzing the implications of both for solving persona challenges. We,

Persona Studies 2018, vol. 4, no. 2

therefore, establish and discuss the role of personas in the context of online analytics, which has
not previously been done in the persona literature.

METHODOLOGICAL APPROACH

Being a conceptual and survey work, this manuscript relies on prior literature as a source of
evidence. To this end, we perform a literature review to identify the key benefits and criticisms
of personas in the academic literature. To find the works of this domain, we conducted searches
on Google Scholar and Science Direct with the base keyword “personas” and then scanned the
abstracts of the articles for cue words indicating weaknesses, downsides, or critiques of
personas. In this literature query, we focused on finding critical papers, as most persona studies
list the benefits but more rarely include criticism (Matthews, Judge & Whittaker 2012). The
literature on persona benefits is, therefore, easier to find. In total, we identify 32 relevant
papers, many of which are conference articles in the field of computer science (sub-fields e.g.
human-computer interaction, design, and software development). We read these articles in
detail and use them to discover related articles discussing the strengths and weaknesses of
personas (a technique referred to as snowball sampling (Provan & Milward 1995).

WHY ARE PERSONAS USEFUL?

Cooper (1999) introduced personas as a design technique for understanding and
communicating the goals and needs of different user types. Since then, personas have been
employed by designers, software developers, and marketers, among other decision-maker
groups (Nielsen 2013; Pruitt & Adlin 2006; Mulder & Yaar 2006). Personas crystallize a specific
user type, often focusing on core users in the absence of an immediate contact to the end user
(Floyd, Jones & Twidale 2008). A user can refer to a user of a software system, such as a website
or mobile application, a player of a game, customer of a product or service, audience of online
content, target segment for marketing campaigns, a patient of public health services, and so on
(Pruitt & Grudin 2003; LeRouge et al. 2013; Ma & LeRouge 2007; Scott 2007; Dong, Kelkar &
Braun 2007; Nacke, Drachen & Göbel 2010; Vahlo & Koponen 2018). In the remainder of this
work, we use the concept of ‘customers’ when referring to these groups. In this research, we
examine personas specifically in the context of online analytics, adopting the perspective that
personas are created to provide real value for their end users (Cooper 1999; Gudjónsdóttir
2010), such as achieving more user-friendly designs or more empathetic advertising texts.
Evaluating personas, therefore, has to take place in relation to their value in use (Kaartemo,
Akaka & Vargo 2017).

The benefits of personas according to the literature are summarized in Table 1. The
communicational benefits arise from summarizing customer information into an intuitive
format of representation that can be communicated with little effort (Holtzblatt, Wendell &
Wood 2005) within organizations, teams, departments, and with external stakeholders
(Matthews, Judge & Whittaker 2012). In theory, personas provide an engaging description of the
end users’ needs and wants, in the form of another human being that is more memorable than
numbers (Goodwin 2009; Hill et al. 2017). At their best, personas become shared mental models
that individuals rely upon when making decisions (Nielsen 2013) that concern the specific user
type (Cooper 1999). This enables the decision makers to discuss experiences and backgrounds
different from their own and realize that the customer preferences may deviate from their own
preferences (Miaskiewicz & Kozar 2011).

Salminen et al.

Table 1: Benefits associated with the use of personas

Category Description

Communication Personas facilitate user-oriented communication within and
between teams in the organization.

Psychology Personas enhance the immersion required for designing ‘for a
person’ instead of fuzzy and complex target groups.

Transformation Personas challenge existing assumptions about customers and
orientate trade-off decisions when customers have conflicting
needs.

Focus Personas help focus design decisions on user goals and needs
rather than on system attributes and features.

The psychological benefits are rooted in identification with the personas (Miaskiewicz,
Sumner & Kozar 2008), whereby decision makers can obtain an empathic understanding of
users, immersing themselves in real situations experienced by others. Decision makers can use
this ability to predict customer behavior under different circumstances (Pruitt & Grudin 2003).
This mental modeling relies on human beings’ innate ability of empathy and immersion
(Krashen 1984), and is, therefore, a powerful agent for enhanced motivation and purpose. At
best, personas can give a higher sense of meaning to one’s work. Consider the psychological
difference between creating a software product to the nameless target group of 24–35 year-old
women, compared to creating a mobile application for Jane, a stressed single mum who wants to
better manage her time.

Transformational benefits relate to challenging the established perceptions about the
users within the organization (Miaskiewicz & Kozar 2011). Because the creation of personas is
based on gathering real evidence of the users (Pruitt & Grudin 2003), the results can deviate
from the existing preconceptions and truisms within the organization. When there is friction
between the perceived and real goals of the users, accurate persona representations can reduce
this perceptional gap by conveying factual information about users’ needs and wants (Pruitt &
Adlin 2006). If the organization and the decision makers are open to re-aligning their
perceptions, the persona exercise can prevent and rectify false conceptions of the end users
(Matthews, Judge & Whittaker 2012).

Finally, personas can facilitate focusing on the most important audiences (Miaskiewicz &
Kozar 2011). This helps decision makers with strategic agenda to prioritize certain customers
over others, and thus resolve conflicting needs and wants among the customer base. For
example, if Persona A wants Feature set x, while Persona B wants Feature set y; by considering
the overall strategy of which users the organization wants to serve (a practice known as
customer portfolio management) (Johnson & Selnes 2004), we can define the optimal product
features to focus on (Cooper 1999; Ma & LeRouge 2007). Thus, personas help to prioritize
product requirements and help determine if the right problems are being solved while curbing
the self-centering bias of the decision makers (Matthews, Judge & Whittaker 2012).

Persona Studies 2018, vol. 4, no. 2

CRITICISM OF PERSONAS

The literature includes a lot of substantial criticism ofpersonas. To better understand and
dissect this criticism, we have categorized it into three sections that are roughly compatible
with the typical lifecycle of a persona project: first personas are created (creation of personas),
then assessed by decision makers (evaluation of personas), and finally applied in real scenarios
and use cases (use of personas). Various critical arguments arise in the literature at each of these
stages. These arguments are summarized in Table 2 and discussed afterward.

Table 2: Established criticism of personas from prior literature.

Category Key issues Authors

Creation of
personas

a) persona creation takes a long time
b) personas are expensive to create
c) personas can be biased by their creators
d) personas are based on non-

representative data

Pruitt and Grudin (2003);
Vincent and Blandford
(2014); Hill et al. (2017)

Evaluation of
personas

e) personas lack credibility
f) personas are not accurate or verifiable
g) the information in personas is not

relevant for decision makers
h) personas are inconsistent

Chapman and Milham
(2006); Bødker et al. (2012);
Matthews et al. (2012)

Use of
personas

i) not using the created personas
j) using personas for politics and power

play
k) using personas to justify preconceptions
l) personas change in time

Rönkkö et al. (2004); Rönkkö
(2005); Chapman and
Milham (2006)

Hill et al. (2017) point out that (a) creating quality personas takes considerable time and
effort. According to Vincent and Blandford (2014), persona creation can take months. In a
similar vein, Pruitt and Grudin (2003) advocate an in-depth research effort for persona creation,
typically lasting months. It is seen that for personas to be accurate, considerable investigative
work is required. This tends to result in (b) persona projects being expensive, in the range of tens
of thousands of American dollars (Marsden & Haag 2016; Miaskiewicz, Sumner & Kozar 2008).
Consequently, as Rönkkö (2005) found, the amount of effort may lead to questioning the return
on investment of persona projects. Moreover, the high cost of persona creation tends to exclude
them from the reach of small businesses and startups, as pointed out by Salminen, Jansen, An,
Kwak, and Jung (2018).

There are also concerns about the validity of personas, or how well the personas match
the reality. A commonly mentioned validity concern is that (d) personas are based on insufficient
or non-representative data (Chapman & Milham 2006). Personas that are built based on
relatively few qualitative interviews may not represent the underlying user groups in a
statistically valid manner. For example, the resulting personas may suffer from bias when the
interviewed subjects are chosen based on availability rather than representativeness of the
entire customer base. Overall, (c) personas risk inheriting organizational tensions and individual
biases, including political and strategic ambitions of their creators (Hill et al. 2017; Massanari
2010; Rönkkö 2005). Vincent and Blandford (2014, p. 1098) argue that “[persona creation] has

Salminen et al.

been adapted, depending on what people want to accomplish and why.” Since persona
representations are not easily verifiable, a deliberate selection bias can take place. This bias can
also be involuntary so that the creators of the personas are projecting their own prejudices
unconsciously in the persona description.

Additionally, since persona creation work is typically qualitative, (e) personas lack the
credibility of numbers and are, to some, interpretative and subjective instead of rigorous and
believable (Chapman & Milham 2006). Even when using the best practices of qualitative inquiry,
number-oriented decision makers may consider personas as ‘nice narratives’ instead of serious
decision-making instruments, resulting in resistance for the use of personas (Massanari 2010).
For example, the participants in the study of Matthews et al. (2012) found personas misleading,
abstract, and unrealistic. In a similar vein, Bødker et al. (2012, p. 93) report that “as soon as the
project started [...] it became a concern that the 12 personas seemed distant from actual citizens,
very general and difficult to activate.” There can be many reasons for such interpretations. For
example, the subjective experiences and impressions of decision makers may conflict with
personas (Marsden & Haag 2016). Furthermore, other information about users, such as direct
customer feedback, can conflict with persona information, because the full complexity and
range of the customer base deviates from the idealized personas (Chapman & Milham 2006). In
such cases, decision-makers need to consider the credibility of personas against other sources
of data. This may result in a willingness to hold on to one’s existing beliefs (Delfabbro 2004),
trusting one’s own observations and insights instead of more abstract personas.

Moreover, there is no objectively right or wrong answer on which information to include
in the persona profile (Bødker et al. 2012; Chapman & Milham 2006). Some decision-makers, for
example, might prefer data relating to customer journey while others are more interested in
psychographics. Therefore, (g) information selection for personas is arbitrary and may not be of
use in a given scenario or use case. The information should be based on the information needs
of the end users of personas (Sinha 2003). These needs vary across industries and use cases,
even between job roles within the same organization. For example, “marketing personas” would
include information such as consumptions patterns and consumer motivations, goals, and likes
and dislikes (Thoma & Williams 2009), whereas online content producers would prefer
information on content consumption patterns (Nielsen et al. 2017).

Additionally, (h) personas are said to be inconsistent, meaning that they are created by
combining information from several unrelated data sources, without ensuring that the
individual pieces of information are commensurable (Matthews, Judge & Whittaker 2012).
Bødker et al. (2012) refer to personas as “Frankenstein’s monsters”, postulating that they can be
patched up from any information available. Due to the above reasons, (f) the accuracy of
personas is difficult to validate. If decision makers in an organization are aware of this risk, they
will not trust the persona representations and will downplay their use in real decision-making
situations (Chapman & Milham 2006). The lack of trust is aggravated when decision makers do
not personally participate in the persona creation (Long 2009), and, thus, the lack the
psychological ownership of the persona artifacts (Bødker et al. 2012). Overall, these adverse
dynamics assert strain on the credibility of personas. If the attitude of the decision makers is not
favorable to the use of personas, the potential benefits remain unachieved.

Finally, we identified five threats for the actual use of personas. First, there are
situations where (i) personas are fully developed but then left without meaningful use. For
example, Rönkkö et al. (2004) report a case where a considerable amount of time was used to
develop personas that were never implemented. In a similar vein, Matthews, Judge and
Whittaker (2012) found that personas had little impact on the actual design work. Friess (2012)

Persona Studies 2018, vol. 4, no. 2

conducted an ethnographic study among designers and found a serious imbalance between the
creation and use of personas in real decision-making situations. Chapman and Milham (2006)
see the lack of validation resulting in political conflicts where (j) choosing a persona is a question
of opinion instead of a fact. For example, the created personas may be rejected in favor of the
manager’s pre-existing beliefs about the customers (Chapman & Milham 2006; Pruitt & Grudin
2003), to frame the strategic and operational discussions in the organization (Rönkkö 2005), or
(k) interpreted to confirm existing beliefs rather than to seek new explanations (Salminen, Jung,
et al. 2018).

Rönkkö (2005) also found that personas were used as a form of after-the-fact
justification, so design choices based on other inputs were later communicated to other team
members as if they were based on the personas. Rönkkö et al. (2004, p. 115) describe a case
where the persona evolved in time and was eventually seen differently by stakeholders of the
company: “The [...] persona who originated as a middle-age businessman ended up as a less
clearly definable figure, e.g. a younger careerist of both the male and the female sex, and a
diversity of different professions whose common characteristics was mobility, e.g. salesman,
plumber, nurse, policeman, veterinary.” This practice in effect nulls the acclaimed benefit of
aligning user understandings within the organization. Finally, (l) use of personas is hindered by
changing customer behavior; as the customer behavior changes, personas should be updated to
reflect these changes (Jung et al. 2017). However, being that data collection is typically
expensive, the updating may not be possible, and personas risk expiring rapidly in real use.

Overall, the above challenges risk creating situations where the persona benefits remain
largely theoretical and do not materialize in real use cases (Friess 2012; Marsden & Haag 2016).
In addition, there are novel concerns arising from the use of personas in comparison to utilizing
online analytics data to understand customers.

NEW CRITICISM

We now move toward introducing the context of online analytics. In addition to the established
criticism laid out in the previous section, there have been newly-found criticisms of personas,
arising from the availability of online analytics data, metrics, and techniques for business
purposes. To demonstrate the logic of this new criticism, we present the following quotations
retrieved from recent online writings by practitioners dealing with online customer data:

“Whereas personas were once a good starting point to identify ‘buckets’ of
customers, the limitations of persona-based marketing have become apparent as
the consumer decision-making journey veered from its predictable linear path
and increased in complexity.”i

“Personas tend to be exhaustive where it’s not needed (demographics, names,
pictures are not necessary most of the times), while they fail to summarise the
complex variety of needs and usage scenarios that real users express in real life
situations.”ii

“The idea of a persona or an average customer was the typical way that
marketers would think about their customer base. But now with advancements
in technology, with modeling, with more available skill sets, they are able to
understand and predict future behavior at more granular levels, and it’s a
dramatic shift that’s happening.”iii

“You’re still clinging to generic user personas in the age of Big Data? LOL."iv

Salminen et al.

The above critical statements are chosen to illustrate the anti-persona sentiment taking place
among some of the data-oriented business professionals. It is not a comprehensive sample, nor
do we argue that all business professionals proficient with online analytics data and tools would
perceive personas as useless. However, there are several recent blog posts that dispute the
usefulness of personas and therefore it is worthwhile to bring this criticism to scholarly
attention and analyze it objectively in the proper frame of context. The crux of the criticism can
be summed up in three categories:

increased complexity (Quotes 1 and 4) – personas are not able to capture the
diversity and nuances of the increasingly large online audiences. Typically, one
would create only a few personas (less than 10) to describe the core users which
may not be enough in the era of online analytics and fragmented consumer
behavior.

redundant information (Quote 2) – personas are overly focused on superficial
demographic information instead of focusing real needs and wants of the
customer base. Also, the information presented tends to be static and not
dynamic.

lack of prediction (Quote 3) – personas are descriptive and not predictive; they
cannot be used for prediction, unlike other analytics tools.

To understand this criticism, we must bear in mind that personas are inherently connected to
decision making about customers. In this sense, they are analytical tools and fall under the
scope of other analytics solutions when applied in practical use cases. As such, practitioners
using personas are questioning them in comparison to other analytical tools. For example, it is
now commonplace to target individual users within digital marketing. Online advertising
platforms, such as Facebook Ads and Google AdWords, have constructed social graphs and
knowledge graphs (Venkataramani et al. 2012) with each user a node with descriptive
properties that can be used for advertisement targeting. Moreover, there have been substantial
algorithmic advances that enhance targeting and optimization of advertising (Graepel et al.
2010; Wang & Yuan 2015). Techniques such as multi-armed bandits define the search space and
find the best matches given an overall target group or population (Chatwin 2013). Given that
users can be targeted and analyzed individually, what purpose is there for aggregated data
representations, such as personas? To answer this question, we examine the ability of data-
driven personas.

EVALUATING DATA-DRIVEN PERSONAS IN THE LIGHT OF CRITICISM: DISTINGUISHING
BETWEEN TRADITIONAL AND DIGITAL DATA-DRIVEN PERSONAS

Cooper’s (1999) initial idea was for personas to be data-driven, i.e., based on real insights about
the users. Later, other scholars working with personas have confirmed this view of personas
originating from comprehensive investigative data collection among real customers (Chapman
et al. 2008; Howard 2015; Pruitt & Grudin 2003). However, the conceptual difference between
manual and digital data-driven personas has been poorly established in the prior literature,
even though this distinction is central for understanding the role of personas amidst digital
data. Therefore, we separate between traditional data-driven personas (TDDPs) and digital
data-driven personas (DDDPs) and focus on analyzing their strengths for solving the criticism
proposed by scholars and practitioners.

What are DDDPs, then? The primary descriptor of DDDPs is that they bridge persona
creation between quantitative data and computational techniques. Several examples can be

Persona Studies 2018, vol. 4, no. 2

found in the literature. For example, Chapman et al. (2008) use conjoint analysis to reduce a
data-set to persona-like representations. Zhang, Brown, and Shankar (2016) use click-stream
data to generate personas of website visitors. (An, Kwak & Jansen 2017) develop a methodology
and system for completely automatic persona generation using YouTube audience statistics.
These novel approaches illustrate ways to combine online analytics data and personas in a way
that draws from the benefits of large-scale online data and retains the core benefits of personas
in showing the data as another human being.

Online analytics data and computational techniques for processing the data provide at
least four major advantages for persona creation (Salminen et al. 2017): (1) the possibility to
automatically collect large volumes of data through application programming interfaces (APIs),
(2) the availability of behavioral data (not only survey and interview responses), providing
better grounds for statistical methods (Chapman et al. 2008), (3) scalability, meaning that data
analysis algorithms and automatic systems can process millions of user interactions from
millions of content pieces, (4) near real-time responsiveness, enabling customer insights to
change as the underlying data changes. These features also make DDDPs different from mere
“quantitative personas” suggested in the literature (Mesgari, Okoli & de Guinea 2018). The
concept of quantitative persona captures the statistical aspect of data analysis, but it does not
correctly capture the aspects of automation and large-scale data analysis associated with the
use of online analytics data characterized by high volume, velocity, veracity, and variety (Storey
& Song 2017).

At its best, automatic data collection and analysis is cost-efficient and behaviorally
accurate across the whole user base, providing excellent foundations for the creation of data-
driven personas. Additionally, DDDPs do not necessarily need to be either quantitative or
qualitative, but they can draw from both types of data, as demonstrated by Salminen, Şengün, et
al. (2018) with their hybrid personas created using quantitative online analytics data and
qualitative insights.

How are DDDPs, then, able to solve persona challenges? And how do they compare
against TDDPs? We perform a conceptual analysis evaluating these questions for each point of
criticism laid out in the previous sections. After this, we discuss the findings and evaluate the
benefits. Table 3 provides a comparison of the ability of TDDPs and DDDPs to address the
criticism.

If we contrast DDDPs against the criticism of personas laid out in the previous sections, we find
that they have the potential to solve many acute problems. First, automation enables rapid
persona creation. Whereas the creation of personas using manual methods, such as
ethnography and surveys, can take several months, digital persona generation system are able
to run the required calculations in the matter of a few hours (Jung et al. 2017). Second, the
personas can be generated by inferring latent patterns of users’ behavior, e.g. video viewing or
website browsing (An, Kwak & Jansen 2017). This technique is robust against personal biases of
human creators and produces personas that are based on behavioral data instead of self-stated
data that has been shown vulnerable to respondent bias (Fisher 1993). Third,
representativeness of the sample is not an issue, when the persona generation is based on the
whole user base. For example, Salminen, Şengün, et al. (2017) generated personas from
YouTube data of a major online news media company consisting of millions of viewers. Using
online analytics data potentially solves the trade-off of relying on either qualitatively rich but
non-verifiable data or using numbers that are accurate but lose the immersion of another
human being.

Salminen et al.

Table 3: Evaluation of TDDPs and DDDPs against persona criticism.

Applies to

Category Key issues TDDPs DDDPs

Creation of
personas

a lot of effort and time needed to
create quality personas

Yes, manual data
collection and analysis

No, data collection and
analysis can be
automated

personas are expensive to create Yes, because they
require manual labor

No, because persona
generation can be
automated and
replicated

personas can be biased by their
creators

Yes, because
information selection is
made subjectively

No, because algorithms
decide the information
shown

personas are based on non-
representative data

Yes, because sampling is
limited

No, because one can
sample the whole
customer base

Evaluation of
personas

personas lack credibility Potentially yes Potentially yes

personas are not accurate or
verifiable

Yes, qualitative analysis
is difficult to replicate
systematically

No, the results can be
statistically evaluated

the information in personas is not
relevant for decision makers

Potentially yes Potentially yes

personas are inconsistent Potentially yes Potentially yes

Use of personas not using the created personas Potentially yes Potentially yes

using the personas for politics and
power play

Potentially yes Potentially yes

using personas to justify one’s
preconceived notions

Potentially yes Potentially yes

personas change in time Yes, because data
collection and analysis
would need to be
repeated

No, because personas
can be updated to reflect
changes in data

Personas in
online analytics
context

increased complexity Yes, because the number
of personas is limited

No, because the
personas can capture a
large number of
patterns

redundant information Potentially yes Potentially yes

lack of prediction Potentially yes No, because the
underlying data can be
used for prediction

Persona Studies 2018, vol. 4, no. 2

Fourth, the “file drawer effect” (Rönkkö et al. 2004) becomes an aggravated issue for TDDPs as
time goes by because of the underlying customer behavior changes over time but the manually
collected data does not change to reflect these changes. DDDPs are responsive to those changes
because the personas can be periodically updated to reflect the most recent behaviors. If the
data is directly obtained from the APIs of online platforms, this update process can be
completely automated, as demonstrated by (An, Kwak & Jansen 2017). Regarding the DDDPs in
the context of online analytics, several studies support our argument that DDDPs can handle the
increasing complexity of audiences and customer bases. For example, (Liapis et al. 2015) are
able to infer dozens of behavioral gaming patterns from their dataset. Salminen, Şengün, et al.
(2018) summarize the audience of a large social media channel, consisting of viewers from
more than 200 countries, to five representative personas. (Kwak, An & Jansen 2017) identify
hundreds of unique content consumption patterns among channel users that they use to
generate personas.

An example of prediction is demonstrated by Jung, Salminen, An, Kwak, and Jansen (2018). They
use online analytics data for persona generation, and then predict the interest of a persona to a
given video using an underlying topic matching algorithm. Even though some prediction tasks
could be done using TDDPs (e.g., “Would Martin like our content?”), the method of doing so
inevitably involves a prominent level of subjectivity, whereas using DDDPs the algorithm treats
the prediction as a numerical problem.

LIMITATIONS OF DDDPS

From the evaluation in Table 3, we see that DDDPs have the greatest potential for solving
persona creation and digital context problems. Some of the evaluation issues can also be
addressed. In contrast, they do not seem to provide a considerable advantage for use of
personas, as the application is subject to organizational and individual biases. Moreover,
consistency has been found an issue also in DDDPs (Salminen, Nielsen, et al. 2018) as in TDDPs
(Bødker et al. 2012). In fact, inconsistency might be even heightened in automatic persona
generation, as there is a subjective safeguard for making sure that the information pieces
selected by the machine are topically consistent (An, Kwak & Jansen 2017). Moreover, even in
the context of DDDPs, information selection remains a challenge, as Salminen, Jung, et al. (2018)
observed in their user study.

Despite the progress made in developing DDDPs, there remain many open challenges,
such as reliance on current audience data, lack of depth, and a discerning lack of basic attributes
in many of the approaches (i.e., not generating persona profiles but behavioral archetypes).
Another major limitation of the current DDDPs methodologies is that none of them include
deeper information and insights about the users, such as customer pain points, motivations,
needs and wants that are essential for the depiction of full, rounded personas (Nielsen 2013).
Understanding the deeper motivations of customers is an essential question for marketers
(Dichter 1964). Furthermore, while the existing DDDPs may be efficient in modeling current
audiences, the decision makers might be interested in potential customers (Thoma & Williams
2009). This interest can be explained by the expansive goals of a typical marketing organization;
that is, marketers are pressured to find novel audiences and markets. However, when the
DDDPs are generated from current audience data (Jung et al. 2018), such information is not
readily available. Thus, we conjecture that the greater the need for reaching new audiences, the
riskier it is to use DDDPs for decision making. Moreover, the use of existing online analytics data
can lead to confirmation bias. For example, a decision maker may only target Women, age 25–34
with his efforts, so when DDDPs reveal to him that the group is indeed his core customers, he

Salminen et al.

carries on saying “I was right,” maintaining his targeting and never, in fact, trying out other
target groups. In other words, DDDPs must be interpreted carefully.

Furthermore, a major limitation of DDDPs is the lack of participation of the team using
the personas. The process of persona creation has been found valuable per se, as participation
increases the decision makers’ interest in and understanding of personas (Molenaar 2017). This
aspect is lacking in DDDPs that are created by algorithms “in distance”, representing a potential
threat for adoption and active use of the persona among decision makers (Matthews, Judge &
Whittaker 2012). It is, however, possible to create so-called hybrid personas (Miaskiewicz,
Sumner & Kozar 2008; Salminen et al. 2017) that combine quantitative and qualitative aspects.
Finally, like for TDDPs, another major limitation for DDDPs is that the benefits postulated in the
existing persona literature remain potential, depending on the decision makers’ actual
willingness to use the personas. Table 4 evaluates the applicability of persona benefits to TDDPs
and DDDPs.

Table 4: Applicability of persona benefits to TDDPs and DDDPs

Applies to

Category Description TDDPs DDDPs

Communication Personas facilitate user-oriented
communication within and between
teams in the organization.

Potentially yes Potentially
yes

Psychology Personas enhance the immersion
required for designing ‘for a person’
instead of fuzzy and complex target
groups.

Potentially yes Potentially
yes

Transformation Personas challenge existing
assumptions about customers and
orientate trade-off decisions when
customers have conflicting needs.

Potentially yes Potentially
yes

Focus Personas help focus design decisions
on user goals and needs rather than on
system attributes and features.

Potentially yes Potentially
yes

The benefits of personas are a question of value in use (Kaartemo, Akaka & Vargo 2017).
That use varies by use case and user of personas. Generally, the same challenges in transcending
the theoretical value of personas to practice apply to both TDDPs and DDDPs. However, as
alternatives to numerical online analytics data, personas do have some distinct advantages. For
example, dealing with numbers poses cognitive challenges for individuals who often cannot
recall many numbers at a time (Miller 1956), whereas human attributes are more easily
remembered (Mulken, André & Müller 1998). Therefore, DDDPs seem to provide an ample
alternative for presenting numerical data, even though they are not perfect.

Persona Studies 2018, vol. 4, no. 2

DISCUSSION

In this research, we examined the role of personas given the widespread availability of online
analytics data, and its perceived useful for business purposes. We reviewed scholarly criticism
of personas and extended it with fresh perspectives from practitioners working with online
analytics data. Although our inquiry is conceptual in its nature, these quotations served as a
basis for understanding the challenges and disadvantages of personas in the modern customer
analytics environment. To better address the question of personas’ usefulness in this
environment, we dissected the concept of personas into TDDPs and DDDPs. This separation was
crucial to make the analysis because the two types of personas are different in their ability to
address the challenges associated with personas in the context of online analytics.

The main contributions of this research include

a) identifying new critical arguments against personas in the context of online analytics,
previously not discussed in persona studies, and appending this criticism to the
continuum of established persona criticism; and

b) conceptually differentiating between traditional and digital data-driven personas and
separately analyzing their ability to address the established and novel criticism of
personas, with the finding that DDDPs possess considerable strengths in regard to both
types of criticism.

Regarding the answer to our research question, namely can personas provide value in
the age of online analytics, we answer that the real value of DDDPs is provided by giving faces to
data, as an alternative way to present online analytics information. Ultimately, however, the
usefulness of personas comes down to specific use cases (Cooper 1999). In general, personas
are useful for tasks requiring a qualitative understanding of customers and numbers are useful
for getting a general overview and, in the case of machine-based decision-making (i.e.,
marketing automation), making individual level optimization. To this end, we postulate that
individual data is optimal for automated decision making, whereas aggregate data such as
personas work best for human decision making, especially relating to decisions at the strategic
level.

The claimed irrelevance of personas seems to be based on the confusion of their use in
an age of online analytics data. It seems that the criticism presented by online analytics
practitioners is based on understanding personas as TDDPs, while overlooking the potential of
DDDPs. This insight further supports the purpose of conceptually separating these two types of
personas, and clearly communicating their differences to end users of online analytics data.

Data-driven personas are not a novel idea, as the purpose of persona creation has
always been to use real customer information to generate realistic user characterizations
(Cooper 1999). However, using computational techniques, such as machine learning, provide
tremendous opportunities toward this end (Salminen, Jansen, et al. 2018). At the same time,
there are a plethora of open research questions to answer. Moreover, the persona research
related to DDDPs tends to be fragmented, while it would make sense for researchers to
collaborate and validate the works of one another to make evolutionary progress in this field.

The core benefits of using personas for design, system development, and marketing have
not changed. However, for tasks such as targeting or recommendation engines, individual level
analytics are likely to perform more efficiently. While these methods are likely to excel in those
use cases, their application to other use cases, such as strategic decision making, is more limited.
As with any analytical technique, the use of personas is relative to the problem one seeks an

Salminen et al.

answer to. In this light, the question “Are personas useful?” should be rephrased as “When are
personas useful?” Finding the answer requires persistent conceptual and empirical research and
exploring new contexts such as online analytics.

CONCLUSION

In conclusion, personas remain a viable option even in the era of online analytics data. It is
possible to combine automatic data collection and other computational techniques to create
accurate persona profiles that can also be used for advanced purposes such as prediction.
However, some of the challenges of TDDPs are inevitably inherited such as the end user relating
to the actual use of personas and their perceived credibility. Therefore, more research and
development work is needed to overcome these challenges and to show the tangible value of
personas in actual use.

i See http://www.cmo.com/features/articles/2017/1/16/why-personas-dont-work-and-what-
innovators-are-doing-differently.html#gs.zxQ5E9s

ii See https://www.humaneinterface.net/article/are-personas-really-useful

iii See http://knowledge.wharton.upenn.edu/article/160811b_kwradio_fader-mariychin-mp3-
zodiac/

iv See http://thecontextofthings.com/2016/08/25/user-personas/

END NOTES

WORKS CITED

An, J, Kwak, H & Jansen, BJ 2017, ‘Personas for Content Creators via Decomposed Aggregate
Audience Statistics’, in Proceedings of Advances in Social Network Analysis and Mining
(ASONAM 2017), Sydney, Australia.

Ardehaly, EM & Culotta, A 2015, ‘Inferring latent attributes of Twitter users with label
regularization’, retrieved 29 October 2018 from
http://cs.iit.edu/~culotta/pubs/ehsan15inferring.pdf.

Bleier, A & Eisenbeiss, M 2015, ‘Personalized Online Advertising Effectiveness: The Interplay of
What, When, and Where’, Marketing Science, vol. 34, no. 5, retrieved 18 May 2016, from
http://pubsonline.informs.org/doi/abs/10.1287/mksc.2015.0930.

Bødker, S, Christiansen, E, Nyvang, T & Zander, P-O 2012, ‘Personas, people and participation:
challenges from the trenches of local government’, Proceedings of the 12th Participatory
Design Conference: Research Papers - Volume 1 (PDC '12), K Halskov, H Winschiers-
Theophilus, Y Lee, J Simonsen, and K Bødker (Eds.). ACM, New York, NY, USA, pp. 91-
100. DOI=http://dx.doi.org/10.1145/2347635.2347649

Boghrati, R, Hoover, J, Johnson, KM, Garten, J & Dehghani, M 2017, ‘Conversation level syntax
similarity metric’, Behavior Research Methods, vol. 50, no. 3, pp. 1055-1073, doi:
10.3758/s13428-017-0926-2.

Chapman, CN, Love, E, Milham, RP, ElRif, P & Alford, JL 2008, ‘Quantitative Evaluation of
Personas as Information’, Proceedings of the Human Factors and Ergonomics Society
Annual Meeting, vol. 52, no. 16, pp. 1107–1111.

http://cs.iit.edu/%7Eculotta/pubs/ehsan15inferring.pdf
http://pubsonline.informs.org/doi/abs/10.1287/mksc.2015.0930

Persona Studies 2018, vol. 4, no. 2

Chapman, CN & Milham, RP 2006, ‘The Personas’ New Clothes: Methodological and Practical
Arguments against a Popular Method’, Proceedings of the Human Factors and Ergonomics
Society Annual Meeting, vol. 50, no. 5, pp. 634–636.

Chatwin, RE 2013, ‘An overview of computational challenges in online advertising’, American
Control Conference (ACC) 2013, IEEE, pp. 5990–6007, retrieved 29 October 2018 from
http://ieeexplore.ieee.org/abstract/document/6580778/.

Clarke, TB & Jansen, BJ 2017, ‘Conversion potential: a metric for evaluating search engine
advertising performance’, Journal of Research in Interactive Marketing, vol. 11, no. 2, pp.
142–159.

Cleary, PD 1997, ‘Subjective and Objective Measures of Health: Which is Better When?’, Journal
of Health Services Research, vol. 2, no. 1, pp. 3–4.

Cooper, A 1999, The Inmates Are Running the Asylum: Why High Tech Products Drive Us Crazy
and How to Restore the Sanity. Sams - Pearson Education, Indianapolis, IN.

Del Vecchio, P, Mele, G, Ndou, V & Secundo, G 2017, ‘Creating value from Social Big Data:
Implications for Smart Tourism Destinations’, Information Processing & Management.
vol. 54, no. 5, pp. 847-860, doi: https://doi.org/10.1016/j.ipm.2017.10.006

Delfabbro, P 2004, ‘The stubborn logic of regular gamblers: Obstacles and dilemmas in cognitive
gambling research’, Journal of Gambling Studies, vol. 20, no. 1, pp. 1–21.

Dichter, E 1964, Handbook of consumer motivations: the psychology of the world of objects.
McGraw-Hill.

Dong, J, Kelkar, K & Braun, K 2007, ‘Getting the most out of personas for product usability
enhancements’, Usability and Internationalization. HCI and Culture. UI-HCII 2007.
Lecture Notes in Computer Science, N Aykin (eds), vol 4559. Springer, Berlin,
Heidelberg, doi: https://doi.org/10.1007/978-3-540-73287-7_36.

Fisher, RJ 1993, ‘Social Desirability Bias and the Validity of Indirect Questioning’, Journal of
Consumer Research, vol. 20, no. 2, pp. 303–315.

Floyd, IR, Jones, MC & Twidale, MB 2008, ‘Resolving Incommensurable Debates: A Preliminary
Identification of Persona Kinds, Attributes, and Characteristics’, Artifact, vol. 2, no. 1, pp.
12–26.

Friess, E 2012, ‘Personas and Decision Making in the Design Process: An Ethnographic Case
Study’, in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems,
CHI ’12, ACM, New York, NY, USA, pp. 1209–1218, retrieved from
http://doi.acm.org/10.1145/2207676.2208572.

Goodwin, K 2009, Designing for the Digital Age: How to Create Human-Centered Products and
Services. Wiley, Indianapolis, IN.

Graepel, T, Candela, JQ, Borchert, T & Herbrich, R 2010, ‘Web-scale bayesian click-through rate
prediction for sponsored search advertising in microsoft’s bing search engine’, in
Proceedings of the 27th international conference on machine learning (ICML-10), pp. 13–
20, retrieved September 20, 2017, from
http://machinelearning.wustl.edu/mlpapers/paper_files/icml2010_GraepelCBH10.pdf.

Gudjónsdóttir, R 2010, Personas and scenarios in use, Skolan för datavetenskap och
kommunikation, Kungliga Tekniska högskolan, Stockholm.

Hill, CG, Haag, M, Oleson, A, Mendez, C, Marsden, N, Sarma, A & Burnett, M 2017, ‘Gender-
Inclusiveness Personas vs. Stereotyping: Can We Have it Both Ways?’, CHI
'17 Proceedings of the 2017 CHI Conference on Human Factors in Computing, pp. 6658–
6671, retrieved May 28, 2017, from
http://dl.acm.org/citation.cfm?doid=3025453.3025609.

Holtzblatt, K, Wendell, JB & Wood, S 2005, Rapid Contextual Design: A How-to Guide to Key
Techniques for User-centered Design, Elsevier.

Howard, TW 2015, ‘Are Personas Really Usable?’, Communication Design Quarterly Review, vol.
3, no. 2, pp. 20–26.

Jansen, BJ, An, J, Kwak, H, Salminen, J & Jung, S-G 2017, ‘Viewed by Too Many or Viewed Too
Little: Using Information Dissemination for Audience Segmentation’, Proceedings of the

http://ieeexplore.ieee.org/abstract/document/6580778/
https://doi.org/10.1016/j.ipm.2017.10.006
https://doi.org/10.1007/978-3-540-73287-7_36
http://doi.acm.org/10.1145/2207676.2208572
http://machinelearning.wustl.edu/mlpapers/paper_files/icml2010_GraepelCBH10.pdf
http://dl.acm.org/citation.cfm?doid=3025453.3025609

Salminen et al.

Association for Information Science and Technology, vol. 54, no. 1, pp. 189–196, doi:
https://doi.org/10.1002/pra2.2017.14505401021.

Järvinen, J 2016, ‘The use of digital analytics for measuring and optimizing digital marketing
performance’, Jyväskylä Studies in Business and Economics 170, retrieved September 8,
2017, from https://jyx.jyu.fi/dspace/handle/123456789/51512.

Järvinen, J & Karjaluoto, H 2015, ‘The use of Web analytics for digital marketing performance
measurement’, Industrial Marketing Management, vol. 50, pp. 117–127,
https://doi.org/10.1016/j.indmarman.2015.04.009.

Johnson, MD & Selnes, F 2004, ‘Customer portfolio management: toward a dynamic theory of
exchange relationships’, Journal of Marketing, vol. 68, no. 2, pp. 1–17.

Jung, S, Salminen, J, An, J, Kwak, H & Jansen, BJ 2018, ‘Automatically Conceptualizing Social
Media Analytics Data via Personas’, Proceedings of the Twelfth International AAAI
Conference on Web and Social Media (ICWSM 2018), retrieved 29 October 2018 from
https://www.aaai.org/ocs/index.php/ICWSM/ICWSM18/paper/view/17810/17089.

Jung, S-G, An, J, Kwak, H, Ahmad, M, Nielsen, L & Jansen, BJ 2017, ‘Persona Generation from
Aggregated Social Media Data’, in Proceedings of the 2017 CHI Conference Extended
Abstracts on Human Factors in Computing Systems, CHI EA ’17, pp. 1748–1755, retrieved
29 October 2018 from
http://www.bernardjjansen.com/uploads/2/4/1/8/24188166/jansen_personas_chi20
17.pdf.

Jung, S-G, An, J, Kwak, H, Salminen, J & Jansen, BJ 2017, ‘Inferring social media users’
demographics from profile pictures: A Face++ analysis on Twitter users’, in Proceedings
of 17th International Conference on Electronic Business, Dubai, retrieved 29 October
2018 from https://aisel.aisnet.org/iceb2017/22/.

Kaartemo, V, Akaka, MA & Vargo, SL 2017, ‘A Service-Ecosystem Perspective on Value Creation:
Implications for International Business’, in Value Creation in International Business,
Palgrave Macmillan, Cham, pp. 131–149, retrieved 11 February 2018, from
https://link.springer.com/chapter/10.1007/978-3-319-39369-8_6.

Krashen, SD 1984, ‘Immersion: Why it works and what it has taught us’, Language and Society,
vol. 12, no. 1, pp. 61–64.

Kwak, H, An, J & Jansen, BJ 2017, ‘Automatic Generation of Personas Using YouTube Social
Media Data’, Proceedings of the 50th Hawaii International Conference on System Sciences,
pp. 833–842, retrieved 29 October 2018 from
http://www.bernardjjansen.com/uploads/2/4/1/8/24188166/jansen_personas_hciss2
016.pdf.

LeRouge, C, Ma, J, Sneha, S & Tolle, K 2013, ‘User profiles and personas in the design and
development of consumer health technologies’, International Journal of Medical
Informatics, vol. 82, no. 11, pp. e251-268.

Liapis, A, Holmgård, C, Yannakakis, GN & Togelius, J 2015, ‘Procedural Personas as Critics for
Dungeon Generation’, in AM Mora & G Squillero (eds), Applications of Evolutionary
Computation, Springer International Publishing, Cham, pp. 331–343, retrieved 14 March
2018 from http://link.springer.com/10.1007/978-3-319-16549-3_27.

Long, F 2009, ‘Real or imaginary: The effectiveness of using personas in product design’, in
Proceedings of the Irish Ergonomics Society Annual Conference, Irish Ergonomics Society
Dublin.

Ma, J & LeRouge, C 2007, ‘Introducing User Profiles and Personas into Information Systems
Development’, AMCIS 2007 Proceedings, retrieved 29 October 2018 from
http://aisel.aisnet.org/amc is2007/237.

Marsden, N & Haag, M 2016, ‘Stereotypes and Politics: Reflections on Personas’, in Proceedings
of the 2016 CHI Conference on Human Factors in Computing Systems, CHI ’16, pp. 4017–
4031, retrieved 18 March 2018, from http://doi.acm.org/10.1145/2858036.2858151.

Massanari, AL 2010, ‘Designing for imaginary friends: information architecture, personas and
the politics of user-centered design’, New Media & Society, vol. 12, no. 3, pp. 401–416.

https://doi.org/10.1002/pra2.2017.14505401021
https://jyx.jyu.fi/dspace/handle/123456789/51512
https://doi.org/10.1016/j.indmarman.2015.04.009
https://www.aaai.org/ocs/index.php/ICWSM/ICWSM18/paper/view/17810/17089
http://www.bernardjjansen.com/uploads/2/4/1/8/24188166/jansen_personas_chi2017.pdf
http://www.bernardjjansen.com/uploads/2/4/1/8/24188166/jansen_personas_chi2017.pdf
https://aisel.aisnet.org/iceb2017/22/
https://link.springer.com/chapter/10.1007/978-3-319-39369-8_6
http://www.bernardjjansen.com/uploads/2/4/1/8/24188166/jansen_personas_hciss2016.pdf
http://www.bernardjjansen.com/uploads/2/4/1/8/24188166/jansen_personas_hciss2016.pdf
http://link.springer.com/10.1007/978-3-319-16549-3_27
http://aisel.aisnet.org/amcis2007/237
http://doi.acm.org/10.1145/2858036.2858151

Persona Studies 2018, vol. 4, no. 2

Matthews, T, Judge, T & Whittaker, S 2012, ‘How Do Designers and User Experience
Professionals Actually Perceive and Use Personas?’, in Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems, CHI ’12, pp. 1219–1228, retrieved
29 October 2018 from http://doi.acm.org/10.1145/2207676.2208573.

Mesgari, M, Okoli, C & de Guinea, AO 2018, ‘Creating Rich and Representative Personas by
Discovering Affordances’, IEEE Transactions on Software Engineering, doi:
10.1109/TSE.2018.2826537

Miaskiewicz, T & Kozar, KA 2011, ‘Personas and user-centered design: How can personas
benefit product design processes?’, Design Studies, vol. 32, no. 5, pp. 417–430.

Miaskiewicz, T, Sumner, T & Kozar, KA 2008, ‘A latent semantic analysis methodology for the
identification and creation of personas’, in Proceedings of the SIGCHI Conference on
Human Factors in Computing Systems, pp. 1501–1510, retrieved from
http://dl.acm.org/citation.cfm?id=1357290.

Miller, GA 1956, ‘The magical number seven plus or minus two: some limits on our capacity for
processing information’, Psychological Review, vol. 63, no. 2, pp. 81–97.

Miralles-Pechuán, L, Ponce, H & Martínez-Villaseñor, L 2018, ‘A novel methodology for
optimizing display advertising campaigns using genetic algorithms’, Electronic
Commerce Research and Applications, vol. 27, pp. 39–51.

Molenaar, L 2017, ‘Data-driven personas: Generating consumer insights with the use of
clustering analysis from big data’, Thesis, retrieved 17 March 2018, from
http://resolver.tudelft.nl/uuid:12d7f261-20b4-4656-93d7-fed2b437aefb.

Mulder, S & Yaar, Z 2006, The User is Always Right: A Practical Guide to Creating and Using
Personas for the Web, New Riders.

Mulken, S van, André, E & Müller, J 1998, ‘The Persona Effect: How Substantial Is It?’, in People
and Computers XIII, Springer, London, pp. 53–66, retrieved 28 February 2018, from
https://link.springer.com/chapter/10.1007/978-1-4471-3605-7_4.

Nacke, LE, Drachen, A & Göbel, S 2010, ‘Methods for evaluating gameplay experience in a serious
gaming context’, International Journal of Computer Science in Sport, vol. 9, no. 2, pp. 1–12.

Nielsen, L 2013, Personas - User Focused Design, Springer Science & Business Media, Springer-
Verlag London.

Nielsen, L, Jung, S-G, An, J, Salminen, J, Kwak, H & Jansen, BJ 2017, ‘Who Are Your Users?:
Comparing Media Professionals’ Preconception of Users to Data-driven Personas’, in
Proceedings of the 29th Australian Conference on Computer-Human Interaction, OZCHI
’17, pp. 602–606, retrieved 28 January 2018, from
http://doi.acm.org/10.1145/3152771.3156178.

Owusu, RA, Mutshinda, CM, Antai, I, Dadzie, KQ & Winston, EM 2016, ‘Which UGC features drive
web purchase intent? A spike-and-slab Bayesian Variable Selection Approach’, Internet
Research, vol. 26, no. 1, pp. 22–37.

Provan, KG & Milward, HB 1995, ‘A preliminary theory of interorganizational network
effectiveness: A comparative study of four community mental health systems’,
Administrative Science Quarterly, vol. 40, no. 1, pp. 1–33.

Pruitt, J & Adlin, T 2006, The Persona Lifecycle: Keeping People in Mind Throughout Product
Design. Morgan Kaufmann, Boston.

Pruitt, J & Grudin, J 2003, ‘Personas: Practice and Theory’, in Proceedings of the 2003 Conference
on Designing for User Experiences, DUX ’03, pp. 1–15, retrieved from
http://doi.acm.org/10.1145/997078.997089.

Ronen, R, Yom-Tov, E & Lavee, G 2016, ‘Recommendations meet web browsing: enhancing
collaborative filtering using internet browsing logs’, 2016 IEEE 32nd International
Conference on Data Engineering (ICDE), pp. 1230–1238, doi:
10.1109/ICDE.2016.7498327.

Rönkkö, K 2005, ‘An Empirical Study Demonstrating How Different Design Constraints, Project
Organization and Contexts Limited the Utility of Personas’, in Proceedings of the
Proceedings of the 38th Annual Hawaii International Conference on System Sciences -

http://doi.acm.org/10.1145/2207676.2208573
https://doi.org/10.1109/TSE.2018.2826537
http://dl.acm.org/citation.cfm?id=1357290
http://resolver.tudelft.nl/uuid:12d7f261-20b4-4656-93d7-fed2b437aefb
https://link.springer.com/chapter/10.1007/978-1-4471-3605-7_4
http://doi.acm.org/10.1145/3152771.3156178
http://doi.acm.org/10.1145/997078.997089
https://doi.org/10.1109/ICDE.2016.7498327

Salminen et al.

Volume 08, HICSS ’05, IEEE Computer Society, retrieved 29 October 2018 from
http://dx.doi.org/10.1109/HICSS.2005.85.

Rönkkö, K, Hellman, M, Kilander, B & Dittrich, Y 2004, ‘Personas is Not Applicable: Local
Remedies Interpreted in a Wider Context’, in Proceedings of the Eighth Conference on
Participatory Design: Artful Integration: Interweaving Media, Materials and Practices -
Volume 1, PDC 04, pp. 112–120, retrieved 29 October 2018 from
http://doi.acm.org/10.1145/1011870.1011884.

Rundle-Thiele, S, Dietrich, T & Kubacki, K 2017, ‘Why We Need Segmentation When Designing
Social Marketing Programs’, in Segmentation in Social Marketing, Springer, Singapore,
pp. 197–214, retrieved 9 May 2018, from
https://link.springer.com/chapter/10.1007/978-981-10-1835-0_13.

Russell, S & Toklu, C 2011, ‘Acquiring and Applying Market Knowledge for Large Software
Purchases: Products, Personas, and Programs’, in Knowledge Management for Process,
Organizational and Marketing Innovation: Tools and Methods, IGI Global, pp. 195–215.

Saggi, MK & Jain, S 2018, ‘A survey towards an integration of big data analytics to big insights for
value-creation’, Information Processing & Management, vol. 54, no. 5, pp. 758–790.

Salminen, J, Jansen, BJ, An, J, Kwak, H & Jung, S-G 2018, ‘Research Roadmap for Automatic
Persona Generation: Principles and Open Questions’, Poster presented at The Fifth
Machine Learning and Data Analytics (MLDAS 2018) Symposium, Doha, Qatar.

Salminen, J, Jung, S-G, An, J, Kwak, H & Jansen, BJ 2018, ‘Findings of a User Study of
Automatically Generated Personas’, in Extended Abstracts of the 2018 CHI Conference on
Human Factors in Computing Systems, CHI EA ’18, p. LBW097:1–LBW097:6, retrieved 26
April 2018, from http://doi.acm.org/10.1145/3170427.3188470.

Salminen, J, Milenković, M & Jansen, BJ 2017, ‘Problems of Data Science in Organizations: An
Explorative Qualitative Analysis of Business Professionals’ Concerns’, in Proceedings of
International Conference on Electronic Business (ICEB 2017), Dubai.

Salminen, J, Nielsen, L, Jung, S-G, An, J, Kwak, H & Jansen, BJ 2018, ‘“Is More Better?”: Impact of
Multiple Photos on Perception of Persona Profiles’, in Proceedings of ACM CHI Conference
on Human Factors in Computing Systems (CHI2018), Montréal, Canada.

Salminen, J, Şengün, S, Kwak, H, Jansen, BJ, An, J, Jung, S, Vieweg, S & Harrell, F 2017, ‘Generating
Cultural Personas from Social Data: A Perspective of Middle Eastern Users’, in
Proceedings of The Fourth International Symposium on Social Networks Analysis,
Management and Security (SNAMS-2017), Prague, Czech Republic.

— 2018, ‘From 2,772 segments to five personas: Summarizing a diverse online audience by
generating culturally adapted personas’, First Monday, vol. 23, no. 6, retrieved 3 June
2018, from http://firstmonday.org/ojs/index.php/fm/article/view/8415.

Scott, DM 2007, The New Rules of Marketing, John Wiley and Sons, Hoboken, New Jersey.
Sinha, R 2003, ‘Persona Development for Information-rich Domains’, in Extended abstracts of the

2003 Conference on Human Factors in Computing Systems (CHI 2003), Ft. Lauderdale,
Florida, USA.

Storey, VC & Song, I-Y 2017, ‘Big data technologies and management: What conceptual modeling
can do’, Data & Knowledge Engineering, vol. 108, March 2017, pp. 50–67.

Thoma, V & Williams, B 2009, ‘Developing and Validating Personas in e-Commerce: A Heuristic
Approach’, in Human-Computer Interaction – INTERACT 2009, Lecture Notes in Computer
Science, Springer, Berlin, Heidelberg, pp. 524–527, retrieved 29 October 2018 from
https://link.springer.com/chapter/10.1007/978-3-642-03658-3_56.

Vahlo, J & Koponen, A 2018, ‘Player Personas and Game Choice’, in N Lee (ed), Encyclopedia of
Computer Graphics and Games, Springer International Publishing, Cham, pp. 1–6,
retrieved March 6, 2018, from http://link.springer.com/10.1007/978-3-319-08234-
9_149-1.

Venkataramani, V, Amsden, Z, Bronson, N, Cabrera III, G, Chakka, P, Dimov, P, Ding, H, Ferris, J,
Giardullo, A, Hoon, J & others 2012, ‘Tao: how Facebook serves the social graph’, in
Proceedings of the 2012 ACM SIGMOD International Conference on Management of Data,

http://dx.doi.org/10.1109/HICSS.2005.85
https://link.springer.com/chapter/10.1007/978-981-10-1835-0_13
http://doi.acm.org/10.1145/3170427.3188470
http://firstmonday.org/ojs/index.php/fm/article/view/8415
https://link.springer.com/chapter/10.1007/978-3-642-03658-3_56
http://link.springer.com/10.1007/978-3-319-08234-9_149-1
http://link.springer.com/10.1007/978-3-319-08234-9_149-1

Persona Studies 2018, vol. 4, no. 2

pp. 791–792, retrieved 29 October 2018 from
http://dl.acm.org/citation.cfm?id=2213957.

Vincent, CJ & Blandford, A 2014, ‘The challenges of delivering validated personas for medical
equipment design’, Applied Ergonomics, vol. 45, no. 4, pp. 1097–1105.

Volkova, S, Bachrach, Y & Durme, BV 2016, ‘Mining User Interests to Predict Perceived Psycho-
Demographic Traits on Twitter’, in 2016 IEEE Second International Conference on Big
Data Computing Service and Applications (BigDataService), pp. 36–43, retrieved 29
October 2018 from https://ieeexplore.ieee.org/document/7474353.

Wang, J & Yuan, S 2015, ‘Real-Time Bidding: A New Frontier of Computational Advertising
Research’, in Proceedings of the Eighth ACM International Conference on Web Search and
Data Mining, WSDM ’15, pp. 415–416, retrieved 28 February 2016, from
http://doi.acm.org/10.1145/2684822.2697041.

Zerbino, P, Aloini, D, Dulmin, R & Mininno, V 2018, ‘Big Data-enabled Customer Relationship
Management: A holistic approach’, Information Processing & Management, vol. 54, no. 5,
pp. 818–846.

Zhang, X, Brown, H-F & Shankar, A 2016, ‘Data-driven Personas: Constructing Archetypal Users
with Clickstreams and User Telemetry’, in Proceedings of the 2016 CHI Conference on
Human Factors in Computing Systems, CHI ’16, , pp. 5350–5359, doi:
10.1145/2858036.2858523.

http://dl.acm.org/citation.cfm?id=2213957
https://ieeexplore.ieee.org/document/7474353
http://doi.acm.org/10.1145/2684822.2697041
https://doi.org/10.1145/2858036.2858523

Joni Salminen, Bernard J. Jansen, Jisun An,
Haewoon Kwak and Soon-gyo Jung
Abstract
Key Words
Introduction
Methodological approach
Why are personas useful?
Criticism of personas
New criticism
Evaluating data-driven personas in the light of criticism: Distinguishing between traditional and digital data-driven personas
Limitations of DDDPs
Discussion
Conclusion
Works Cited