Willingness to Accept for Instagram Accounts. First Empirical Evidence
©2016
Textbook
93 Pages
Summary
Finding an incentive compatible method to assess account values can be seen as the bedrock of social media research across all platforms and is of crucial importance for researches and practitioners alike. This study presents a new method of applying the willingness to accept (instead of the commonly applied willingness to pay for establishing account values on Instagram), by modifying a randomised Vickrey Auction. Primary research among 1024 participants and 409 Instagram users measured the willingness to accept, in relation to demographic variables, account and interaction metrics. The average account was valued at 100 € median, correlating significantly with participants' income and prevalently with the number of followers. Other significant correlations were found in the duration since sign up, number of posts, average number of likes and comments as well as the decision to establish a micro blogging business. Findings are discussed by regarding the limitations and implications for Instagram's business model in terms of a Freemium model, insurances companies offering privacy enhancing features and ad campaign pricing when users engage in brand collaborations.
Excerpt
Table Of Contents
List of Figures
Figure: Title:
Source:
Page:
1
Overview of research fields
Own
Illustration
8
2
Relative and absolute growth of members/users
Own
Illustration 10
3
Characteristics of network industries applied to
Instagram
Own Illustration
11
4
Comparison of takeover prices within the industry
Own
Illustration 14
5
Formula to determine the optimal price for spon-
sored content
Own Illustration
18
6
Incentive compatibility
Own
Illustration
26
7
Question 1
Own
Illustration
29
8
Measuring Consumer's WTA
Own Illustration
31
9
Validation process of account performance indica-
tors
Own Illustration
32
10
Snowball Effect of Survey Responses Own
Illustration
33
11
Anonymity of the survey
Own Illustration
36
12 Variable:
user
Frequencies
Own
Illustration 37
13
Figure 13: Variable: sex Frequencies Own
Illustration
38
14
Variable: sex chi square test with Gondorf
(2015)
Own Illustration
39
15
Variable: age descriptive statistics
Own Illustration
39
16
Age pyramid of frequencies
Own Illustration
40
17
Variable: country bar chart of frequencies
Own Illustration
41
18
Variable: marital frequencies
Own Illustration
42
19
Variable: education frequencies in comparison
between the UK
Own Illustration
43
20
Variable: income frequencies
Own Illustration
44
21 Key
Findings
Demographic
Variables
Own
Illustration 44
22
Variables: followers, followings, posts - frequen-
cies
Own Illustration
45
23
Variable: category frequencies and percentage
Own Illustration
46
24
Variable: business_income frequencies and per-
centage
Own Illustration
47
25
Key Findings: Account Metrics and Business
Own Illustration
47
26 Variable:
wta
descriptive
statistics
Own
Illustration 49
27 Variable:
wta_2
frequency
Own
Illustration 49
28
Variable: wta_2 Kolmogorov-Smirnov Test
Own Illustration
50
29
Key Findings Key Findings: WTA
Own Illustration
51
30
Variable: wta Pearson correlation with age
Own Illustration
51
31
Variable: wta Spearman correlations with demo-
graphic variables
Own Illustration
52
32
Key Findings: Demographic Correlations
Own Illustration
52
33
Variable: wta Pearson correlation with follower,
followings and posts
Own Illustration
53
34
Variable: wta Spearman correlations with ac-
count metrics
Own Illustration
53
35
Key Findings: Account Metrics Correlations
Own Illustration
54
36 Regression
Analysis
Own
Illustration 54
37
Linear Regression Graph
Own Illustration
55
38 Final
formula to determine the optimal price
Own
Illustration
56
List of Abbreviations
B
r
Random Bid
B
S
Bid of the Seller
BDM
Becker DeGroot Marschak (Becker et al., 1964)
E
Expected value
HTML
IQR
HyperText Markup Language
Interquartile Range (box plot)
N
F
Number of Instagram Followers
p* Optimal
price
PSM
Price Sensitivity Meter (van Westendorp, 1976)
RP Risk
Premium
RQ Research
Question
r
x;y
Correlation between x and y
u
s
Derived Utility of the Seller
v
s
True Value of the Possessed Good (WTA)
WTA
Willingness to Accept
WTP Willingness
to
Pay
Change in two Figures
Profit
1 Introduction
The radical novelty of the Web 2.0 has shifted focus from website administration
generated content to user generated content and allowed for the emergence of user
generated businesses in the same movement. Whilst a vast majority of users can be
described as "lurkers" and passive users, following the 1% rule, only one in one
hundred people actually actively participated in creating content on the internet ten
years ago (van Mierlo, 2014). Fast forward 10 years, the share of active users post-
ing videos, sharing pictures, writing blogs has now risen to 9% in 2009. The emer-
gence of social networks, allowing for more enticing opportunities to interact with
other users or brands and low barriers to sharing content over 3G and now 4G tech-
nology can be held accountable for this development. In the same period the first in-
cidences of social media entrepreneurship emerged and allowed users to gain reve-
nue from social media sites in return for various forms of advertising and related in-
come sources. The suddenly arising business-critical issues of privacy concerns
(e.g. losing the account over unauthorised access or technological errors) and tech-
nology based questions have been researched in a platform specific manner, with
most research revolving around Twitter and Facebook users. Despite its rapid
growth and recent overtake by Facebook, reliable economic studies on social pic-
ture sharing service Instagram are not represented in contemporary research.
Therefore, this study should serve as the first foundation in this area and analyse
the monetary value behind Instagram accounts. By closing this existing research
gap, insurances may be able to expand their current range of services and insure
Instagram related business models. At the same time, Instagram receives the op-
portunity to launch a fee-based premium version of their existing platform, which fur-
ther secures the value and privacy of one account, in the form of a Freemium model.
The third beneficiaries of the study are Instagram users, who derive information on
how to find the optimal price for sponsored content on the network.
The dominant methodology to determine price preferences, analogous to Facebook
or Twitter research, is to statistically determine a users willingness to pay (WTP,
purchasing price) to keep and use the existing account, through a survey. Willing-
ness to accept (WTA, selling price) as the slightly higher counterpart mechanism is
expected to complement the survey over the endowment effect. The effect postu-
lates that there is a "sentimental" non-rational value attached to the account, overes-
timating its actual value. Yet the overestimation may drive the price a customer is
willing to pay for an insurance and privacy protection (WTA > WTP).
6
To allow incentive compatible
1
measurement of the WTA (instead of the WTP), the
renowned Becker-DeGroot-Marschak mechanism is modified and applied to suit the
study requirements. Following data collection all established demographic variables
such as age, gender and education and key account metrics such as followers,
number of posted pictures shall then be used in a second step to define statistically
significant correlations to further break down and define WTA composition.
The scientific approach in measuring account values and metric will allow insight for
three main stakeholders of this study: Instagram, may now act as provider of addi-
tional security through a Freemium model and thus may be able to secure additional
revenue, aside from advertising. Knowing a close to accurate value of a profile in-
surance companies will be able to insure users and corporate accounts against loss
of business and the intellectual property created. Professional social media users,
who sustain a significant proportion of their monthly income through Instagram, will
be able to determine an adequate market price for advertising opportunities, taking
into account positive and negative changes in followers.
In order to diligently and systematically address the below research question this
dissertation will be divided into the following conceptual chapters. An analysis of key
findings in current literature will allow us to establish an up to date understanding of
network industries generally with special focus on social networks, their characteris-
tics, pricing options and applicable mechanism in Chapter 2. Chapter 3 will then jus-
tify and design an appropriate research plan, touching upon ethical issues and po-
tential risks and their mitigation during the data collection. When analysing sample
data, focus will be set on demographic variables in descriptive statistics and correla-
tions with key account metrics and socio-demographics to develop a round under-
standing of the topic. Concluding, all findings will be discussed in regards to limita-
tions, implications and potential further research questions.
1
For the individual to disclose their true preferences and values
7
Explore the following research questions:
RQ1 Mathematically develop an incentive compatible method to expose user's
willingness to accept (WTA) for existing Instagram accounts.
RQ2
Calculate descriptive statistics of the willingness to accept (WTA) for existing
Instagram accounts.
RQ3 Find significant correlations (to different levels) of the willingness to accept
(WTA) with demographic variables and key account metrics in order to inves-
tigate the current level of development of businesses on Instagram.
Figure 1: Overview of research fields (own illustration)
8
2 Literature
Review
2.1
Alignment and Social Networks
Past years have seen a trend in the usage of social media platforms, where users
are able to interact by self or peer- generated content in a virtual environment
(Kaplan and Haenlein, 2010). Traditionally, information in the online context has
been shared unidirectional from few content creators to attract are large number of
recipients by using simple HTML web pages (Cormode and Krishnamurthy, 2008).
Guided by several technological changes and increasing information processing ca-
pacity ("Moore's law") of servers since 2004 this proportion has progressively shifted
towards the number of content creators (Fox and Pierce, 2009, Kambil, 2008,
O'reilly, 2007). In the so-called Web 2.0 era content and information flows differ from
previous times as users may contribute to or customise an online service or platform
using a profile or personal user account (O'reilly, 2007, Weinberg, 2009). This sub-
stantial and on-going change also empowered millions of people to start online
businesses that create value by using social media platforms, blogs, video sharing
platforms or other web applications characterised through significantly lowering set
up costs and barriers to enter an online market than previous offline markets
(O'Reilly, 2005). Generally, social media platforms feature the highest diffusion rates
among content creators as they aim to target a broad audience with strategic focus
on direct network effects (Kietzmann et al., 2011, Lin and Lu, 2011). While "older"
networks such as Facebook (established in 2004) and Twitter (established in 2006)
have frequently been addressed in studies in the field of social, business and eco-
nomic science, the novel social picture sharing service Instagram (established in
2010) has been underestimated in its growth potential by researchers in the past.
The current growth rates support this hypothesis (Figure 2) and further indicate a
knowledge gap in regards to the social network Instagram. In addition, the rise in the
Alexa ranking, which can be seen as the dominant web analytical service, of 31
places to position number 26 (19.05.15) highlights the rapid diffusion of Instagram at
the moment (Alexa, 2015, Lo and Sedhain, 2006). Bearing that in mind, Instagram
may aptly be described by the "Early Majority" phase of the diffusion process,
whereby Facebook and Twitter are allocated in the "Late Majority" or in the turning
point of Rogers (2010) renowned classification.
9
Figure 2: Relative and absolute growth of members/users in own illustration
(Alexa, 2015, GWI, 2014)
Two papers shall especially be highlighted at this point to reflect the current state of
knowledge in regards to account value in social networks.
First of all, the publication of Schreiner and Hess (2013a), which can be seen as the
staring point of the dissertation, successfully applied the van Westerndorp's Price
Sensitivity Meter (PSM) (1976) to investigate the willingness of Google Plus and Fa-
cebook users to pay for privacy-enhancing additional features. The research aim
was to statistically examine whether a so-called Freemium model, an omnipresent
pricing structure in online businesses that consists of a free and a premium version,
could be applied to the two social networks. Thereby, the paper serves as a perfect
example of accurate data collection through online surveys and application of a rec-
ognised and advanced technique to estimate willingness to pay. Despite the recog-
nition of the well-outlined and integrated PSM method, it must be emphasised that
the chosen approach is considered to be lacking incentive compatibility by state of
the art literature (Kim et al., 2012, Miller and Hofstetter, 2009). Beyond the lack of
theoretical grounding, the PSM mechanism is rather
difficult to explain and to illus-
trate in graphs. Another weakness of this paper is the relatively small sample size of
160 participants and non-representativeness concerning age and educational back-
ground (primarily students), which further limits the possibility to draw general impli-
cations. Those weaknesses, especially in regard to the used mechanism, directly
impact the dissertation as they provide profound insight into the major difficulties in
identifying consumer's monetary preferences, as they usually don't want to disclose
it (Wertenbroch and Skiera, 2002).
Secondly,
the paper of
Bauer et al. (2012), which
explores the perceived value of
actively as well as passively uploaded information on the social networking platform
Facebook, is of utmost importance. In the course of their study, the authors tested
several variations of the Becker-DeGroot-Marschak mechanism on a large sample
of 1045 participants in form of an online survey-based experiment. However, a seri-
ous weakness of the study is the distorting influence of external stimuli (e.g. free
10
iPhones), which had a significant correlation with the stated WTP (Bauer et al.,
2012). In other words, the WTP was fluctuating depending on whether a low- or
high-value external incentive was used.
Moreover, the study is completely lacking a
control group in order to make statistical adjustments
2
retrospectively. As a result,
the authors regression model to identify driving factors for participants WTP only has
a "small explanatory value of 14.5%" and consequently "couldn't explain one's pro-
file valuation" (Bauer et al., 2012, p. 8)
In the further evaluation, the
study outcomes are
mainly transferred to the context of
online privacy by the authors, but are also dependent on the perceived value of the
overall online profile (WTA of the profile subtracted by the social surplus), which is
one of the research questions of my dissertation.
Interrelation under the assumptions of rationality
3
(own illustration calculated accord-
ing to (Neus, 2013)):
2.2
Characteristics of Network Industries
In order to appropriately analyse the value of an Instagram account in later steps,
the social network Instagram has to be set into the right context. This guarantees
correct assessment of the special environmental circumstances in a network indus-
try, where the focus shifts from valuing an entity instead of valuing its connection to
other entities.
Figure 3: Characteristics of Network Industries applied to Instagram (own illustra-
tion)
2
Concerning the distorting influence of external stimuli on the stated WTP
3
Formula demonstrates the connection between the study of Bauer et al. and my research
question.
11
Figure 3 analyses the social network with respect to the dominant theoretical
frameworks of network industries by using four characteristics that distinguish those
industries from ordinary markets for commodities or stocks (Gabel, 1994, Majumdar
and Venkataraman, 1998, Shy, 2001). The first and eponymous factors in such in-
dustries are direct and indirect network effects, which may also be regarded as posi-
tive externalities (Akerlof and Kranton, 2000, Katz and Shapiro, 1986). In the case of
Instagram, direct networks (within one side) are present between users of the plat-
form. Moreover, they can also be seen as the driving force that enabled Instagram
to quickly establish a critical mass of users (tipping point), where the process of
gaining new users becomes self-sustaining, with no marketing costs or other incen-
tives needed (Markus, 1987, Valente, 1995). A large and growing body of literature
label this unique process as the "bandwagon effect" (Farrell and Saloner, 1985). In-
direct network effects on the other hand, are considered to be present at a minor ex-
tent at the moment, but may rise with the gradual implementation of sponsored con-
tent in 2014 and 2015 (Instagram, 2014). However, the key problem is that it may be
hard to distinguish in reality whether a network effect is truly direct or indirect since
an exact demarcation between the same and a different entity or side is impossible
4
.
The second characteristic focuses on (technical) complementarity to other products
or services, which often requires a compatible standard resulting from a negative
cross elasticity in demand (Kuraitis, 2009). In the case of Instagram, the application
requires a smartphone or tablet with Android or iOS software to operate. Limited ac-
count management functions or non-interactive Web 1.0 viewers are available via
PC.
Moreover, network industries usually face immense economies of scale as the set
up costs (fixed costs) of a platform or software are assumed to be relatively high
(Shy, 2001). Due to the fairly low distribution and maintenance costs (variable costs)
in the case of electronic copies, firms in network industries have the opportunity to
achieve a fast degression of fixed cost, especially after reaching a critical mass
(Koshal, 1972). Instagram is no exception in regards to economies of scale and has
been able to significantly reduce the cost per new user since the launch of the plat-
form (TechCrunch, 2010).
Lastly, switching costs and lock-in effects, which occur if members of a platform try
to switch to a similar service, are also present on Instagram (Klemperer, 1987,
Kuraitis, 2009, Shy, 2001). They play a distinct role in the estimation of the value of
an account, as users may value their account higher as the rebuilding of an account
4
E.g. are business profiles on the same or on a different side in relation to a normal ac-
count?
12
may come with high or prohibitive costs. Shapiro and Varian (2013) provide a holis-
tic framework that further classifies various sources of lock in effects. In the case of
Instagram contractual-, information-, search- and especially data conversion costs
should be emphasised at this point. Shapiro and Varian (2013) didn't consider the
huge potential losses of connections in the network through switching (e.g. follow-
ers, followings, comments and likes), which might exceed the pure data transfer
costs by far.
All in all, the investigated pivotal characteristics of network industries can easily be
transferred to the social network Instagram. Hence, the above mentioned, underly-
ing theoretical models have to be taken into account in the methodology of the em-
pirical research. Questions have therefore been specifically designed to not only
survey the value of an account but also to reflect direct network effects that enable
the emergence of it. However, it still has to be emphasised that Shy's (2001) frame-
work of characteristics only reflects a small piece of digital reality and strategically
marginalises mixed forms of industries.
2.3
The Social Network Instagram
2.3.1 Historic
Development
After describing Instagram in the right business context, this section now targets the
social network itself from a microeconomic point of view, which enables a better un-
derstanding of the following chapters.
Instagram was developed by Kevin Systrom und Mike Krieger after working on their
HTML5-based check-in software burbn, which combines elements of Foursquare
and Mafia Wars (Systrom, 2010). The extended version of burbn, which allowed us-
ers to share, comment and like pictures was initially released on Apple's App Store
on October 6
th
2010 (Desreumaux, 2014). At the same time, the project was re-
named to Instagram, which is a portmanteau of "telegram" and "instant camera" and
according to the developers "better captured what you were doing" in the application
(Systrom, 2010). Their funding in the development phase from the 5
th
of March 2010
onwards, raised a total amount of 500.000 USD, including early, well-known inter-
net-investors such as Andreessen Horowitz and Twitter co-founder Jack Dorsey
(Siegler, 2010). From a computer science point of view mainly open source solu-
tions were used such as Ubuntu Linux, nginx-webserver, Django and PostgreSQL.
Similar to Foursquare and Dropbox, Instagram is hosted in an Amazon EC2 cloud,
which helped to reduce infrastructure and maintenance costs through economies of
scale (Bains, 2014, Armbrust et al., 2010).
13
In early 2011, the development team decided to integrate a "hashtag" function,
which is a method for content based filtering for microblogging services and social
networks. Although, hashtags were mainly used inside the boundaries of twitter
5
at
that point in time, they rapidly set out to be a useful tool to find photos with a specific
theme or content on Instagram. On the 3
rd
of April 2012 Instagram launched their
services for Android based operating systems, which drastically increased global
dissemination at that time (Thomas, 2012). A version for personal computers is not
in the interest of the company and many key features are still only available via mo-
bile devices.
The same month, Facebook announced to take over the 13 employee strong Insta-
gram team for the price of one billion US dollars, which was equal to 1,592 billion
British pounds in 2012, in cash and stock and ranks at the highest takeover price in
the industry (Figure 4) (Facebook, 2012). After the deal was closed and approved by
the Federal Trade Commission in the U.S. in September 2012, Facebook decided to
keep the company independently managed. However, Facebook has a decisive in-
fluence on the company in relation to the monetisation of its business model and
launch of advertising services (Instagram, 2015a).
Figure 4: Comparison of takeover prices within the industry (own illustration)
2.3.2 Distinct
Features
As established before, Instagram successfully convinced consumers since its launch
on October 6
th
2010 to adopt their new social networking platform. The creators
Kevin Systrom and Mike Krieger base their success on several key components.
One component was the strict implementation of a differentiation strategy, which
aimed to separate the network from their main rival at that time, Facebook. This was
mainly achieved through limiting the potential photo size to a square shape, similar
to Kodak Instamatic and Polaroid images, which improved clarity, simplicity and mo-
bile accessibility (standard scripts of transfer and substantially lower traffic of data).
5
Hashtags originated in twitter in 2007
14
Moreover, Instagram enables instant photo/video upload and editing by using 24
unique and patented filters. For the first time an online service made professional
photo adjustments options available for smartphones and tablets, which are other-
wise only available via desktop computers (e.g. Dodge and Burn technique). Like-
wise, the network also served as a useful add on for Twitter, which already featured
54 million active users in Q4 2010 (Statista, 2015c). This is mainly due to the fact
that Twitter did not feature an appropriate option to upload or rather edit photos from
mobile devices before sharing. For this reason, Instagram mainly served as an in-
termediate between mobile devices and Twitter in the start up phase of the network
by substantially lowering transfer costs of consumers (Manning et al., 1995). This
opportunity of convincing Twitter users to sign up for Instagram was strategically
used by the creators to overcome potential problems in the decisive early diffusion
process of the network (Rogers, 2010, Kwak et al., 2010, Valente, 1995, Valente,
1996). Consequently, from the start up of the network, photos could be shared par-
allel on other social networks with a single click (Instagram, 2015c). However, Twit-
ter realised the on-going cannibalisation and outmatch of Instagram in terms of
growth rates and advised it's high-profile users (e.g. celebrities) to stop sharing In-
stagram links and photos on Twitter (Fiegerman, 2015).
Further, the service is still available for free, making full use of the bandwagon effect
(increasing preference for a commodity due to direct network effects) and network
industry theory concerning consumer's adoption of new technologies. As a conse-
quence, the network did not have any significant source of income until early 2015,
when their advertising program was launched in the US (eMarketer, 2015). This
program is estimated to generate 0.59 billion in 2015, 1.48 billion in 2016 and 2.81
billion USD in 2017, due to new advertisement products for capital-intensive as well
as extensive advertisers, including the "Custom Audience" feature, which aims to
reduce scattering losses (Acar et al., 2015, eMarketer, 2015, Paul et al., 2015). Oth-
er e-commerce pricing strategies to further increase sales are still feasible and may
work alongside Instagram's advertising program (Lee, 2001, Turban et al., 2015).
Similar to the video game market, a free-to-play pricing strategy, based on the
Freemium software model, can be considered as suitable, as the incentive to join
the platform remains high (Hanner and Zarnekow, 2015, Alha et al., 2014). A possi-
ble implementation of this pricing strategy for Instagram will be discussed in Section
2.5.2 alongside with a feasibility analysis. Moreover, the outcome of this study is ex-
pected to directly influence the applicability of a Freemium model.
In contrast to free-to-play pricing, pay-to-play in general is not a suitable option for
social networks, aiming for a high amount of users in order to establish direct net-
work effects (Nojima, 2007). A large body of literature investigated the negative cor-
15
relation between high upfront fixed costs and speed of adoption (Barron and Torero,
2015, Ma et al., 2015, Shriver, 2015). Similarly, a part of existing users may not be
willing to pay for a service, which was previously for free.
2.4
User based advertising on Instagram
Instagram is increasingly described as a microblogging platform, which enables up-
load of user-generated content in smaller amount or file size than a normal blog
(Culloty, Leadbeater, Zappavigna and Zhao). The uploaded content itself are usually
named microposts and consist of short sentences, individual images, or video links
(Aichner and Jacob, 2015). As the similarity in the name microblog and blog would
already suggest, the associated revenue models are largely congruent, with the ex-
ception for revenue stemming from platform unique advertisement like Google Ad-
sense, banner advertisements or pop-ups (Lee et al., 2015).
Users or microbloggers on Instagram generally have the opportunity to receive two
types of capital-forming payments from advertising companies. One is actual mone-
tary compensation in advance in exchange for favourable mentions of the respective
brand in hashtags and pictures in line with the company´s marketing communication
strategies among the users target group. In the second option users "only" receive
free items or gifts from their clients in return for their microblogs. It is often, however,
a mixture of both forms of compensation. Increasingly free products are also sent to
Instagram users with a big audience without prior contractual agreement on adver-
tisement, in the hopeful prospect that the recipient likes the respective products and
thus discreetly advertises over a shared picture. Bloggers and companies alike em-
bed Instagram as an integral part of their multichannel marketing campaigns
amongst platforms like Twitter and Facebook for either self-promotion or to increase
product sales or to enhance their brand reputation as an approachable and interac-
tive brand.
To determine the actual value and price of this form of picture based advertising and
to define the business foundation of digital social influence it is necessary to sys-
tematically examine the underlying variables of the account. It is to assume that
number of followers, as user who can share, like and comment on the picture, and
thus determine the outreach of a profile, is the most critical determinant in this case.
Still further interaction indicators and ratios, need to be taken into account to estab-
lish the value more accurately and determine profiles that may boast a large follow-
ing, but score low on interaction (e.g. over bought, inactive followers).
16
2.5
Significance of the study for professional practice
In general, four different stakeholders are involved in the social network, whereby
three of them (Instagram users, Instagram and insurance companies) could directly
benefit from the results of the study and the outcomes of the three research objec-
tives. The fourth stakeholder, companies that seek advertising space and sponsor-
ing partners on the platform are not targeted by this study. Though, important find-
ings may also be derived from this empirical research as firms may achieve a better
understanding of their contractual partners, especially their demographics. The ad-
vertising business model of companies is, however, widely regarded as analogous
to the ones offered on Facebook.
2.5.1 For Instagram Users
As outlined in Section 2.4, ordinary people enjoy the progressively decreasing start
up costs for building up a business using Instagram. However, the price of their
main source of revenue is still hard to assess for individuals. While some Instagram
users demand only a couple of pounds, others sell their advertising space in the
form of sponsored posts for thousands of pounds per photo. For instance, Danielle
Bernstein of the Blog "We Wore What " states to earn from 5,000 USD (£3,255) to
15,000 USD (£9,767) per sponsored post on her Instagram Account
"@weworewhat", monetising her audience of around 1 million followers (Kapadia,
2015). This uncertainty concerning the pricing of brand partnerships may be based
on three key factors.
Firstly, information about pricing of other Instagram users is scarcely available
online and outside the two partners of a market transaction. Likewise, it may not be
in the interest of one transaction partner to provide full information to the other party
(Ballwieser et al., 2012). This form of market failure, falls in line with classical micro-
economic and macroeconomic theory and results in a large amount of uncertainty in
the market for sponsored posts concerning pricing as well as the quality of the ser-
vice (Ottum and Moore, 1997). As a result of asymmetrically distributed information,
the supply side of a market transaction faces an incentive to offer low quality spon-
sored posts. However, rational advertising firms take this into consideration and pre-
fer sponsored posts with moderate or low quality due to risk aversion. In line with
Akerlof's theory of 1970, this ultimately leads to an equilibrium where mainly low
quality services are traded in the market (assumptions of the theory are reviewed in
Appendix 1). Namely, distribution platforms such as "buysellshoutouts.com" and
igshoutouts.com are affected and offer advertising space on inactive accounts.
However, critics of the model stem from fact that buyers can frequently seek ways to
17
prove the quality (e.g. by pre-testing, online user reviews or indicators for account
interactiveness).
Secondly, there is no consensus on which key figures of an Instagram account are
pivotal for its value, as no study previously investigated this topic. However, be-
cause the visible and predictable indicators of accounts are finite, they can be
shortlisted and tested for correlation with the overall value. Hence, it enables Insta-
gram to improve the relevant components, which increase the attractiveness of the
profile for sponsored posts and brand partnerships.
Lastly, many individuals, especially beginners, in the online advertising business are
not aware that the sale of advertising space also leads to variable costs for the indi-
vidual user. Every single post of sponsored content may reduce the number of fol-
lowers and thus one of the key figures of an account. This may on the one hand be
due to a deviation in content of sponsored posts, no longer matching the interest of
the person following or on the other hand due to increased post frequency, now an-
noying their followers. According to observation and empirical documentation the
loss in the amount of followers per sponsored post varies between 0.01% and 0.8%
(Appendix 2). With the help of the overall account value (WTA) and the correlation in
relation to the number of followers, the optimal price could be determined at which
sponsored posts should be offered (Figure 5). However, users may aim for a price
that exceeds the monetary compensation, due to volatility effects that requires a risk
premium
6
(RP) (Cunningham et al., 2005). Two examples, including the special
scenario, where the user increases the expected number of followers through selling
sponsored posts (e.g through increased reputation) are outlined in the Appendix 3.
Moreover, they show that, irrespective of the assumptions, a universally valid result
with the correct mathematical sign is generated by the formula.
Figure 5: Formula to determine the optimal price for sponsored content (own illustra-
tion)
6
Assumption: Instagram Users are risk-averse E(u(w)) < u(E(w))
18
All in all, advertising on Instagram should be viewed as a method of monetisation or
gradual sale of the respective audience. In order to determine the optimal price (p*)
for sponsored content, it is crucial to know the overall value of the profile (WTA) and
to the correlation with the number of followers (r
NF;WTA
). Both variables are part of
this work and had direct impact on the selection of the three research objectives.
Moreover, all variables in the formula are easy to obtain or to appreciate, and thus it
offers an easy way the approximate the optimal price, using the overall account val-
ue.
2.5.2 For Instagram in Terms of a Freemium Model
As mentioned in Section 2.3.2, a Freemium model could complement Instagram's
current pricing strategy and could provide additional features for a particular group
of users. Thereby, Freemium is a portmanteau and describes a combination of a
free (eng. Free) and a paid premium offering, which is widely used by software firms
and service providers in the Web 2.0 era (Bekkelund, 2011). For instance, in-
application purchases of Freemium applications resemble about 92% of revenues
on Apple's App Store (Schoger, 2013). They key concept was first characterised by
venture capitalist Wilson on his personal online blog in 2006, as a hybrid business
model, which offers basic function for free and offers users the possibility to sub-
scribe a premium version (Wilson, 2006). In the course of this, the standard version
does not usually contribute to the revenue of the company (Wilson, 2006, Lee et al.,
2013). However, as already emphasised, the launch of a premium version is not ex-
pected to hinder the parallel expansion of Instagram's ad program.
It is of vital importance to assess, which differentiation strategy is the most appropri-
ate in terms of expected revenue and applicability for the social network. Puyol
(2010) identified three independent, object related possibilities (by quantity, feature,
or distribution) that enable premium services to increase the derived utility of a sub-
set of users. Particularly, a premium version that enhances privacy and security of
user profiles in the shape of a value-added service, could be a promising approach,
as various studies find a existing market among users (Bauer et al., 2012, Schreiner
and Hess, 2013b, Schreiner and Hess, 2015). The practical implementation could
then be composed in the form of back-up solutions, geographically secured pass-
words and extended privacy settings. According to Puyol (2010) this form of differ-
entiation would be assigned by feature. Similarly, other authors describe the busi-
ness model as a performance-related price differentiation with self-selection (Simon
and Fassnacht, 2008). Thereby, the allocation to the premium version is not manda-
tory for certain individuals (Skiera, 1999). Rather, every individual assigns himself or
19
herself, based on his individual preference, to a service (free or premium version)
that maximises the respective personal utility (Skiera, 1999).
Recent evidence indicates three basic prerequisites for successful implementation
of a Freemium model, which shall be analysed in the following for Instagram
(Bekkelund, 2011).
In the first place, the company should hold the ability to offer its service at low varia-
ble costs, at least on the free version (Bekkelund, 2011). In the case of Instagram as
a business in the e-commerce sector, high fixed costs and low variable costs,
amounting to one USD per additional user can be assumed (Leber, 2012, Stähler,
2002). Furthermore, a large, established audience is necessary for advertisement
for the premium version, which is present in the case of Instagram (300m users).
Lastly, a sufficiently large proportion of the established user base should have a will-
ingness to pay for the premium service (Anderson, 2009, Bekkelund, 2011). As a
basis for this, user must assign a value to their own profile, which can be protected
by enhanced security mechanisms and privacy settings. This value shall empirically
be approximated in this study by using the willingness to accept to sell the respec-
tive profile (RQ2).
2.5.3 For
Insurances
A potentially high perceived and actual value (RQ2) could open up new target
groups for insurance companies. Especially, since some users generate significant
revenues through Instagram as a part time job or sometimes even as a profession
(analogous section 2.4). At the same time, users increasingly consider who may
have access to personal data and how data is used (Dinev and Hart, 2006,
Taddicken, 2014, Tucker, 2014). This form of privacy concern, can be subdivided,
according to Smith et al. (1996) into four dimensions ("Collection, Errors, Secondary
Use and Unauthorised Access to Information"). In particular degree, Errors and
Secondary Use shall be emphasised, as they may lead to deletion of the profile and
thus of the potential revenue stream.
The survey in this thesis is therefore designed to collect the main figures that could
help insurances to enlarge their on-going expansion in the e-commerce sector by
covering Instagram accounts (Gordon et al., 2003, Vaidyanathan and Devaraj,
2003). According to current information (August 2015) no English or German insur-
ance company provides aforethought insurance for Instagram accounts, although a
significant amount of users earn money using them. After an initial offering of ac-
count insurances, the insured sum could be based on key performance indicators, a
20
desired sum could (for example, the WTA) or be based on the perpetuity of account
revenues (Dufresne, 1990).
In summary, as a result of the existing gap in the market the expansion of digital
business models, it is only a matter of time until account values have to be covered
by insurance companies. Hence, the author and a team of two students are cooper-
ating with the entrepreneurship centre at the LMU Munich for further assistance and
development of a business plan, following on from this dissertation.
2.6 Conceptual
Framework
2.6.1 Willingness to Accept
In the context of quantifying the personal value of an account, the distinction be-
tween willingness to pay (WTP) and willingness to accept (WTA) is of great signifi-
cance for the outcome and is consequently highlighted in the conceptual framework
of the study (Horowitz and McConnell, 2002). Thereby, the WTA corresponds to the
monetary equivalent at which an individual is indifferent between keeping a good
and abandoning it (Zeiler and Plott, 2004). On the other hand, WTP, is usually de-
scribed as the maximum amount, an individual is willing to pay in order to transfer a
good into his own possession. Research findings regarding the amount of the WTA
have remained unchanged for the past 15 years and are aptly described by John K.
Horowitz et al. (2000, p. 1): "Previous authors have shown that WTA is usually sub-
stantially larger than WTP, and almost all have remarked that the WTP/WTA ratio is
much higher than their economic intuition would predict." The difference between
the two metrics is partially explained by the endowment effect
7
, a psychological ef-
fect that leads to higher perceived value of goods in own possession (Knetsch et al.,
2001, Kahneman et al., 1991, Zeiler and Plott, 2004). Other researchers, especially
in the online context, use Akerlof's (1970) assumptions regarding quality uncertain-
ties in order to explain the discrepancy (Horowitz and McConnell, 2000, Neus, 2007,
Pae, 2005). The concept refers to the uncertainty of the decision maker about the
value of the commodity. Decision makers are assumed to generally avoid having to
make a decision under (complete) uncertainty (Neus, 2007, Davis and Reilly, 2012,
Okada, 2010). Therefore, decision makers strictly prefer to delay the purchase situa-
tion until more information about the commodity is available or expertise is gained.
In the case of a decision maker not being able to defer the moment of decision to a
later moment in time, he will subsequently demand for an adequate risk premium (R
f
> 0)
to compensate for the presence of adverse information. However, it must be
7
Sometimes described as divestiture aversion
21
emphasised that this explanation can only account for imperfect markets, as it relies
on the existence of asymmetric information.
As uncertainty is expected to play a significant role when it comes to assessing the
value of an Instagram account, the used mechanism will be adjusted to mitigate a
possible discrepancy (Georgantzis and Navarro-Martínez, 2010, Isik, 2004, Zhao
and Kling, 2001). Current literature in the field, such as Bauer et al. (2012) often
failed to consider this significant difference and consequently received distorted re-
sults.
22
Details
- Pages
- Type of Edition
- Erstausgabe
- Publication Year
- 2016
- ISBN (PDF)
- 9783960675716
- ISBN (Softcover)
- 9783960670711
- File size
- 15.4 MB
- Language
- English
- Institution / College
- Manchester Metropolitan University Business School
- Publication date
- 2016 (August)
- Grade
- 0,8
- Keywords
- WTA Instagram Freemium model Becker-DeGroot-Marschak Mechanism Social network Empiric study WTP Willingness to pay Vickrey Auction
