Electronic Commerce Research and Applications 12 (2013) 69–77

Contents lists available at SciVerse ScienceDirect

Electronic Commerce Research and Applications

journal homepage: www.elsevier.com/locate/ecra

Under what conditions will social commerce business models survive?

Dohoon Kim
School of Business, Kyung Hee University, Hoegi-dong 1, Dongdaemoon-gu, Seoul 130-701, South Korea

a r t i c l e i n f o a b s t r a c t

Article history: This study provides a model that captures the essential features of the social commerce business. The
Available online 28 December 2012 model focuses on the relationship between key decision issues, such as marketing expenditures and
the revenue streams that are created. As more social commerce businesses enter the marketplace, they
Keywords: are faced with fierce competition, which may lead to sharp increases in marketing and advertising expen-
Competition ditures. This type of competition may lead the industry away from its optimal development path, and at
Economic analysis worst, toward a disruption of the entire industry. Another goal of this study is to examine the possibility
Game theory
that the tragedy of commons may occur in the industry. The basic analysis presents Nash equilibrium
Marketing
Regulation
results with homogeneous and heterogeneous players. The analysis further specifies the conditions that
Social commerce the tragedy of commons can occur. I discuss the strategic implications and policy directions that may be
Social networks able to overcome the shortcomings of current business model, and help the industry to achieve more sus-
Tragedy of the commons tainable development.
Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction
Social commerce is an emerging category of e-commerce, based
on social media platforms like social network services. It allows its
users to participate in buying and selling products and services
through the platforms (Marsden 2010, Curty and Zhang 2011).
Many social commerce providers started their business by combin-
ing group-buying with selling discount coupons offered from their
partners over the Internet. These social commerce providers split
the revenues with their business partners at a predefined rate. Gro-
upon initiated this business model in 2009. It required more than a
certain number of customers to buy the coupons for future delivery
of goods or services. This has been referred to as a form of group-
buying. Social commerce services also involve daily deals and flash
deals; these service offerings are usually valid for a short period
of time.
Social commerce has grown rapidly in the past 3 years, and
more customers, business partners and investors have joined the
industry. More than 500 providers are running social commerce
businesses worldwide now (Reibstein 2011).1 In South Korea, the
associated transactions amount to more than US$200 million (Digital
Daily 2011, Lee 2011). According to Kim (2011), the sales revenue for
these kinds of businesses in the country has increased from US$45
million in 2010 to almost US$500 million in 2011. These figures sug-
E-mail address: dyohaan@khu.ac.kr
1
The statistics vary to some extent since the ways to define the social commerce
industry are different across countries. Other statistics argue that the number of social
commerce firms in the middle of 2011 amounted to 320 in the US, more than 5000 in
China, more than 300 in Japan, and 230 in Korea (Chung and Chen 2012, Kim 2011,
Lee 2011, ROA Holdings 2011).
gest that the industry has grown ten times in terms of sales revenue
and twenty times in terms of transactions over a year. By the end of
2011, more than a third of the population in South Korea subscribed
to and experienced the service (Kim 2011). Similar figures reflect
industry growth in East Asia, where the social commerce business
has gotten very popular (Chung and Chen 2012, Kim 2011, Lee
2011). Over the past few years, sales revenues have increased from
US$1.2 billion dollars to US$3.55 billion dollars in China, and from
US$8.4 billion to US$11 billion in Japan (Financial News 2011). In
addition, they have gone from US$780 billion to more than US$1 tril-
lion in the US.
The emergence of social commerce reflects the collective bar-
gaining power of end-users as the Internet has shifted the bargain-
ing power from sellers to consumers (Gu et al. 2012, Wei et al.
2011). An exemplar of this change is what social networking ser-
vices brought to the distribution channels and marketing efforts.
Thanks to this new opportunity, consumers and the younger gen-
eration who are now initiating and shaping market trends have
been exposed to more deals, discounts and new information
around their local areas (Stephen and Toubia 2009, 2010). Accord-
ingly, they have been easily attracted to the service proposals from
social commerce firms and boosted the industry in its early stage.
Some criticisms about social commerce firms are emerging now
though. For example, MacMillan (2011a, 2011c) reports on the
diminishing popularity of Groupon in the US and China. Reibstein
(2011) asserts that the fundamental weaknesses of the group-buy-
ing business model is that its success will eventually limit growth
due to a huge increase in competition that is likely to arise, as well
as a sharp decrease in its capability to deliver the original value
business proposition that was promised to consumers. Webster

1567-4223/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.elerap.2012.12.002
70 D. Kim / Electronic Commerce Research and Applications 12 (2013) 69–77
(2011) suggests that operational cost is one of the weakest links of
the group-buying business model. ‘‘. . . Groupon’s cost line will in-
crease as more staff are needed to find new merchants, expand to
new geographies or both. And according to the FT article, that is
just what is happening. In Q1 of 2011, administrative and market-
ing expenses rose from US$11 million to US$387 million against
revenues of US$270 million, creating a cumulative deficit of
US$522 million’’ (Webster 2011). Last, Wheeler (2011) also ad-
dresses problems beginning to surface in Groupon’s current busi-
ness practice that requires massive upfront investments for
customer acquisition.
Social commerce startups have also drawn skepticism due to
their unusual accounting practices, rising marketing costs and
quality assurance issues. This has made it more difficult for social
commerce firms to lure investors now. Groupon, for example,
experienced fluctuation of stock price after its initial public offer-
ing, and LivingSocial decided not to do it. However, the most crit-
ical views on the social commerce industry point out that the
industry’s growth rate is unsustainable (Reibstein 2011). Some
observers have argued that the business model has flaws and val-
uations cannot be justified in the market (Webster 2011, Wheeler
2011). The resulting instability has left customers, business part-
ners and investors disenchanted.
According to the Korea Consumer Protection Board, the number
of cases involving problems and customer complaints reached 500
in 2011 (Kim 2011). Many social commerce firms seem to have
been suffering under a large load of marketing expenses and low
average revenue per user (ARPU). This result has been predicted
since the firms’ business practices have to invest a big portion of
revenue on advertising and promotion. They also have maintained
a wide range of service offerings, but the assortment is too broad,
and the operational costs are too high, while user revenues remain
low (Kim 2011, Lee 2011).
The prosperity and adversity of the social commerce industry
has implications for other e-commerce industries. The business
model of social commerce may seem to be IT-intensive, but it
heavily relies upon other things too. Human resources and man-
power are critical and costly in the competition that occurs in this
industry. The business model needs to operate effectively across
different regions, requires negotiation and contracting with part-
ners, and needs advertising and promotion of its service offerings
to attract consumers. All of these activities require human inter-
vention. This explains why average sales per employee is lower
than in other e-commerce sectors, such as social networking ser-
vices, search engines, and business portals (Financial News 2011,
Lee 2011).
These features allow social commerce providers to lower the
initial fixed cost. On the other hand, low capital investment also
diminishes the entry barriers making it easier for late-comers to
copy the business model. The low barriers to entry make it difficult
for a marketing campaign to be effective. One possible outcome is
that the business model could end up with a bubble and the entire
industry may collapse. Social commerce firms in this new phase of
the industry should revise their value propositions and develop a
differentiated niches in comparison to online shopping malls and
open markets.
This study presents model to capture the essential features of
the social commerce business model. I will analyze the model to
see whether social commerce is sustainable, and discuss conditions
for stable evolution of the industry. First, I will focus on the rela-
tionship between a firm’s marketing effort and the revenue stream
that it produces. As more social commerce firms have joined the
industry, fierce competition has ensued, resulting in additional
sharp increases in the firm-level marketing and advertising expen-
ditures. This type of competition may lead the industry away from
its optimal development path, and at worst, result in disruption. So
one intended contribution of this study is to examine the possibil-
ity that the tragedy of commons (Greco and Floridi 2004, Hardin
1968) may occur in the industry.
In Section 2, I present a model to demonstrate the essential fea-
tures of the social commerce business process and the competitive
landscape. I will analyze the model in the next section and inves-
tigate the possibility that the tragedy of commons occurs in the
industry due to an excessive competition in market share. The
implications of my findings and analysis follow in Section 3. Sec-
tion 4 discusses the future development of the social commerce
business model to overcome its limitations, and concludes.
2. Model
Marsden (2010, p. 4) has offered a broad definition of social
commerce as follows: ‘‘Simply defined, social commerce is the fu-
sion of social media with e-commerce, or in the words of IBM, so-
cial commerce is basically the concept of word-of-mouth applied
to e-commerce. More fully, social commerce is a subset of elec-
tronic commerce that uses social media, online media that sup-
ports social interaction and user contributions, to enhance the
online purchase experience.’’ Social commerce is an emerging busi-
ness that incorporates Internet-based online media, such as social
connections, mobility, openness, and participation. Social com-
merce takes on various business formats in practice: for example,
buyer communities, group-buying, purchase sharing, social cura-
tion, social advice, co-shopping, and so on (Cavazza 2012, Marsden
and Chaney 2012). In this study, I will focus on group-buying, the
most popular format of social commerce (Chung and Chen 2012,
Lee 2011, Kim 2011).
Social commerce offers a value proposition to potential custom-
ers by allowing them to buy in groups and receive quantity discounts
(Anand and Aron 2003, Jing and Xie 2011). Merchants or suppliers,
who act as business partners with social commerce firms, also gain
benefits from selling a high volume of goods or services to a group
through a single social commerce channel, instead rather than sell-
ing on a one by one item or service basis. Further, suppliers can use
social commerce firms as a marketing channel to access new cus-
tomers and increase sales. The key to the social commerce business
model lies in deep discounts and the pooling of consumers willing-
ness-to-pay, and then turning that potential to an effective real de-
mand. Although business practices vary by region (Chung and Chen
2012, ROA Holdings 2011), my study focuses on the pooling of po-
tential demand and the generation of revenue.
Potential investors have expressed uncertainty about whether
the business model is sustainable, and seem less interested than
before (MacMillan 2011a, Wheeler 2011). The weakest link of the
business model comes from its simplicity. It is simple enough to
be copied without large initial investment, and it is difficult to pre-
vent potential competitors from entering. Competition has been
intensifying as a result.
Social commerce firms also appear to forced to spend more
money to acquire customers due to the intensifying competition.
In 2010, Groupon spent US$4.95 per new customer, but in 2011
it had to spend US$5.30, an increase of about 7% (MacMillan
2011a). This increase is worrisome to potential investors since it
may signal that that it is getting more costly to acquire and retain
customers.
This problem arises in the presence of lower available inventory
or service capacity. When thus is the case, the willingness of sup-
pliers to offer deep discounts will decline, and price-sensitive
shoppers may switch to other social commerce firms that offer a
better deal. In the long run, competition among social commerce
firms will drive down the discount rate to the minimum required
customer service configuration threshold.
 D. Kim / Electronic Commerce Research and Applications 12 (2013) 69–77 71

Note: Although the real business practice in the social commerce industry varies from country to country, the basic
features and the key elements remain the same: in particular, social commerce as the group buying model based on
social media. Figure 1 depicts the conceptual frame of the social commerce business model in this study. Three key
elements are social commerce providers, users and partner suppliers. Arrows in solid thick line represent business
transactions. Arrows in broken thick line show the money flow. The social commerce providers connect users with
suppliers and collect service fees (commission) from suppliers.

Fig. 1. Social commerce business model: stakeholders, structure and process.

I will formulate a social commerce business model that incorpo-
rates the strengths and inherent weaknesses, and examine the pos-
sibility of transaction and market failures. I also will focus on key
decision issues such as marketing effort and service configurations
that are offered to customers. (See Fig. 1.)
Due to the competition, however, the revenue sharing rate be-
tween the supplier and the group-buying intermediary is likely
to be standard across most social commerce firms. For example,
the rate in South Korea was about 20% several years ago, according
to Kim (2011). With a fixed rate, my model allows a social com-
merce firm to leverage its minimum required number of items to
be sold as a competitive tool. I further assume that the discount
rates associated with different service configurations are already
reflected in this customer service configuration threshold level.
Suppose there is a set of N social commerce firms, where k is an
individual social commerce firm. N may also denote the set of so-
cial commerce firms if it is clearly distinguishable in the context,
N = {1, . . . , n}. I define the following modeling notation:
ek = marketing efforts of social commerce firm k;
tk = customer service configuration threshold (hereafter thresh-
old) set by social commerce firms (a reference point represent-
ing a service configuration including a discount package and a
minimum number of customers in order for the service offering
to be effective)2;
dk = social commerce firm k’s tolerance capability to maintain
positive cash flows in choosing the threshold level (the maxi-
mum level of threshold that social commerce firms k endures);
and,
E = total marketing efforts currently under way in the industry
P
ðE ¼ k2N ek Þ.
My model of social commerce firm business activities follows.
First, a firm issues coupons that can be used to purchase products
or services from a partner supplier at a constant discount rate.
However, those coupons are effective only when the amount of
coupons sold exceeds a minimum required number of users, or a
threshold tk set by the corresponding firm k. The revenue of the
firm k will be proportional to the effective demand that the social
2
As I explained earlier, social commerce business models in some countries do not
include tk and dk as their key decision elements. I will provide a business model in
Section 2 which excludes these from the decision elements.
commerce firm faces. Usually, the revenue function of firm k can be
represented by rk (tk, ek, E), where
@rk @r k @ 2 rk
> 0; < 0; and < 0: ð1Þ
@ek @Ek @t 2k
For example, I may employ r k ðtk ; ek ; EÞ ¼ eEk  t k  ðdk  tk Þ, where dk is
the maximum threshold that firm k uses; firm k will loses money if
it sets tk beyond dk.
Now, some additional explanation about the conditions in Eq.
(1) is in order. First, the sales revenue based on the number of deals
concluded by firm k will be proportional to its relative marketing
expenses. This feature reflects the current situation with low entry
barriers and brand recognition directly related to market share.
This yields the first inequality in Eq. (1). The marketing efforts of
other players negatively affect the corresponding revenue rk, as
the second inequality suggests.
The threshold positively and negatively affects the sales reve-
nues of firm k. The bigger tk is, the larger the profit margin firm k
will expect. On the other hand, the probability of transaction fail-
ure increases as tk rises. A transaction failure occurs when demand
falls short of the threshold needed to permit sales to be completed.
Social commerce firms should compensate for transaction failures
according to some predefined service-level agreement that is
reached with the partner suppliers, recognizing losses in their rev-
enue streams, and losses from the customers’ viewpoint too.
According to a survey conducted by the Korea Chamber of Com-
merce and Industry (2011) in 2011, more than 50% of the com-
plaints from social commerce firm customers come from
transaction failures, where there are quantity shortages and qual-
ity degradation due to excessive sales of coupons. Firms are
responsible for their transaction failures, and they should compen-
sate customers for breaches of service agreements, which reduces
their revenues. Thus, an increase of the threshold tk will enhance
the revenue at first, but it will also increase the possibility of trans-
action failures, thereby reducing revenues in the end. I model the
effect of tk on revenues as a concave function, which requires the
third inequality in Eq. (1).
Finally, I need to net the costs of effort of social commerce firm
k, which is assumed to be proportional to the amount of effort ek:
that is, ck  ek. Note that ck involves both pecuniary and non-pecu-
niary unit costs incurred in the course of operations pertaining to
marketing. Thus, it can be thought of as all the ex ante burden
when firm k implements one unit of marketing action. Also, ck
should not be confused with marketing expenses ek, which
72 D. Kim / Electronic Commerce Research and Applications 12 (2013) 69–77
Note: The relative marketing efforts εk’s are uniquely determined if the pair of constants (ζ12, ζ13) belong to the feas-
ible region specified above. For example, for a given tuple (ζ12, ζ13) = (1, 0.5), which belongs to the feasible region
above, the relative marketing efforts in a Nash equilibrium, ε1*, ε2*, ε3* are determined as , , .
 2
di c
Fig. 2. fij  dj
 cj parameter region where
i
represents ex post values paid for marketing-related activities.
There will be no costs associated with the decision for tk since
the decision is a matter of deliberation and does not incur pecuni-
ary costs. Thus, the profit of firm k is:
e 
k
pk ¼ r k  ck  ek ¼ tk   ðdk  t k Þ  ck  ek ðtk 6 dk Þ
E where Ek ¼ i–k ei . The matrix is negative definite at the points sat-
P
and total industry profit is defined as k2N pk .
3. Analysis
My analysis first presents the Nash equilibria of the model.
Assuming that heterogeneous social commerce firms may employ
different strategies, the following proposition shows that there are
infinitely many solutions, as in the Nash bargaining game (Fuden-
berg and Tirole 1991), for a best response individual marketing ef-
 e
fort ek. However, the relative marketing effort ek  Ej can be
uniquely determined. Let k denote an arbitrary firm among N social
d2 c
commerce firms, k 2 {1, . . . , N}. Also define the constants fij  di2  cji
e j
for all i < j and ej  Ej for all j in N. The latter represents the relative
strength of marketing effort from social commerce firm j. This
leads to Proposition 1:
Proposition 1 (The Best Response Threshold and Marketing Effort
Proposition). Given that N  1 firms except firm k have decided their
optimal best response threshold tj and marketing effort ej
(j = 1, . . . , k  1, k + 1, . . . , N), firm k’s optimal t k and ek (both positive)
are: t k ¼ d2k and ek as a solution to the following linear equation
system, 1  fij ¼ fij  ei þ ej ; 8i < j in N. Then, ek is determined by
P  ek with a suitable proportional constant P.
Proof of Proposition 1. First, one can show that t k satisfies the
first-order necessary condition. The linear equation system for the
ek’s (k = 1, . . . , N) comes from a set of first-order necessary condi-
tions. It is possible to derive a closed form solution for ek by utiliz-
ing the matrix structure of the linear system and employing
Cramer’s rule (Chiang 2006, Fudenberg and Tirole 1991). The de-
tailed procedure is omitted here, but an example is demonstrated
below. (See Fig. 2.)
Once the ek ’s are identified, next is to construct the values of ek
by multiplying the corresponding ek by the constant P. Although
the system of simultaneous equations for the ek’s has infinitely
many solutions, thanks to the linearity of the equation system
for ek’s, ek is unique up to scalar multiplication.
ej ’s are feasible (i.e., 0 6 ej 6 1).
To check out the second-order sufficient conditions, I construct
the Hessian matrix H as follows3:
" #
 2EEk tEk4ðdk tk Þ Ek ðdk 2t k Þ
H¼ E2
;
dk 2t k
E
 2eEk
ð2Þ P
isfying the first-order necessary conditions if both t k and ek are po-
sitive as I have assumed in the proposition. h
For example, when N = 3 and (f12, f13) lies in the shaded area in
the figure below, the ek ’s are: e1 ¼ 1þf 12 þf13
1þf12 þf13
; e2 ¼ 1f12 þf13
1þf12 þf13
, and
e3 ¼ 1þf12 f13
1þf12 þf13
.
According to the Best Response Threshold and Marketing Effort
Proposition (P1), the optimal threshold is proportional to the capa-
bility that the corresponding firm can exert in the market. The rel-
ative marketing effort of firm k; ek (and ek too) increases as fkj
(j – k) increases, but decreases as fjk (j – k) increases. Thus, more
marketing effort by firm k is expected if its relative capability
(ddkj ; j–k) increases or its relative marketing cost (cckj ; j–k) de-
creases. However, the former will have a stronger effect on ek than
the latter since fij is proportional to square of the relative capabil-
ity. Subsequently, the critical competitive edge can be gained from
enhancing the capability that a social commerce firm has to main-
tain a positive cash flow against low margins.
If all firms have the same capability and cost structure, a sym-
metric Nash equilibrium can be found based on the following prop-
osition.4 As for homogenous social commerce firms, it is not
necessary to distinguish the player strategies ek’s since they will be
the same. Thus, I employ e, instead of ek, as the representative
player’s strategy without specifying a player’s index.
The symmetric case with homogeneous firms may fit two stages
of the industry lifecycle. The first is the very early stage of the
industry, when a small number of similar-size companies consti-
tute the industry. Another one is the mature stage of the lifecycle,
3
The Hessian matrix is the square matrix of 2nd order partial derivatives of a
function with multiple variables. It is frequently used by examining the sign of the
matrix to prove that the current solution under investigation is really a local
extremum (Chiang 2006). Here, I need to employ the Hessian matrix since the model
deals with two types of decision variables: tk and ek.
4
Nash equilibrium in a symmetric game permits changing the identities of the
players without changing the payoff to the strategies (Fudenberg and Tirole 1991).
Here, the game becomes symmetric since all the social commerce firms are now
assumed to be identical in terms of the capability and the cost structure.
 D. Kim / Electronic Commerce Research and Applications 12 (2013) 69–77
Fig. 3. A sample graph of E (c = 1 and d = 10; thus, b
E ¼ 25).
where many small and medium-sized firms with low value of dk
are forced out of the market, and a small number of big with sim-
ilar properties survive. This discussion leads to Proposition 2:
Proposition 2 (The Symmetric Player Strategies Proposition). In the
symmetric strategies case, where ck = c and dk = d for all k = 1, . . . , N, a
symmetric Nash equilibrium is determined via tk ¼ t ¼ 2d and
d2
ek ¼ e ¼ N1
N2
 4c for all k = 1, . . . , N.
Proof of Proposition 2. For tk here, the same reasoning as in
Proposition 1 is applicable. By assuming symmetric strategies, I
can construct a system of linear equations for the ek’s directly from
2 8ca
the set of first-order necessary conditions, c  ðE Þ2 ¼ d4  Ek , for all
2 d2 ðN1Þ
k. The last equation reduces to c  N  e ¼ 4 
since E⁄ = N  e⁄
   
and Ek ¼ ðN  1Þ  e . Thus, I get tk and ek as above, from which
the second-order sufficient conditions are satisfied too. h
In the case of symmetric strategies, the optimal level of the
threshold t will not depend on the number of social commerce
firms in the industry. On the other hand, it is interesting to look
into the combined effort or total expenditure from all social com-
merce firms, E ¼ N  e ¼ N1
2
 d4c, which depends on how many
N 
there are. The combined effort increases with N dE

> 0 , but the
dN
2 
rate of growth is diminishing in NðddNE2 < 0Þ. Furthermore, E con-
verges to a number as N goes to infinity: limN!1 E ¼ d4c  b
2
E. Thus,

E is a concave function of N, which converges to E. (See Fig. 3.) b
Although each social commerce firm may exert less marketing
effort when there is a large number of firms (see e in Proposition
2), the addition of new social commerce firms will swamps this ef-
fect, leading to increases in total marketing efforts (E) in the mar-
ket. If we assume that the revenue function reflects market
demand, then there will be the possibility of overexploitation of
customers. Collectively, social commerce firms will exert market-
ing effort far beyond the point that boosts potential market de-
mand to its maximum level. This resembles the typical situation
of the tragedy of the commons: this sort of negative externality
is at the heart of the problem (Alroy 2001, Hardin 1968, Greco
and Floridi 2004). When a social commerce firm advertises, it does
not take into account the negative effect that its action might have
on the revenue streams of other social commerce firms.
To examine this possibility in more detail, I first define the
industry performance measure W() as a function of the total mar-
keting expenses E and the average customer service configuration
threshold t:
WðE; tÞ ¼ CBðE; tÞ  PBðE; tÞ ð3Þ
where CB and PB stand for customer benefits and producer benefits.
Under symmetric strategies, 
d ¼ d and t ¼ t, the latter is simply the
sum of all the profits from the social commerce firms:
73
X
PBðE; tÞ  pk ¼ t  ðd  tÞ  c  E:
k2N
CB is supposed to have a linear and additive relationship with E and t:
that is, CBðE; tÞ ¼ a  E  b  t, where a and b are positive coefficients.
This notion of CB is natural since the scale of demand for social com-
merce services is more likely to increase with greater total marketing
effort. In addition, the degree of customer benefits will grow as the
average customer service configuration threshold level falls.
However, CB and PB are not measured by the same standard,
and they cannot be combined in a simple weighted sum. The liter-
ature on cost-benefit analysis and multi-criteria decision-making
suggests to employ a multiplicative form in this case (Alroy
2001, Sirakoulis and Karafyllidis 2012). It is also possible to incor-
porate additional weights to adjust the balance between CB and PB.
We did not apply such weights in Eq. (3) though, since mypurpose
is not to quantify or estimate the exact amount of the industry per-
formance, but to examine qualitative behaviors of the system. As a
result, my industry performance measure is the product of CB and
PB, and the expression for W(–) in Eq. (3) is:
WðE; tÞ ¼ ft  ðd  tÞ  c  Eg  ða  E  b  tÞ: ð4Þ
With the industry performance measure defined in Eq. (4), I next of-
fer a third proposition to explain how socially-optimal values of E0
and t0 can be determined.
Proposition 3 (The Industry Performance Proposition). Assume that
the firms employ symmetric strategies p asffiffiffiin Proposition 2, and that the
following condition holds dc > ab  ð1 þ 3Þ. Then, the total marketing
effort E0 and the average threshold t0 maximize industry performance
as defined in Eq. (4) will be t0 ¼ dc and E0 ¼ dðadþ2cbÞ .
Proof. First, the first-order conditions are satisfied with t0 and
E0, if a  d > c  b (in particular, for t0), which is also satisfied by
the condition above. To check out the second-order conditions,
we constructed the Hessian matrix,
 
2c  a c  b þ a  ðd  2tÞ
H¼
c  b þ a  ðd  2tÞ b  ð4t  dÞ  2a  E
. This is negative defi-
nite at t0 and E0, when (a  d  c  b) > 3(c  b)2. This is equivalent to
2
d b2  2  2  2
c
 a > 3  ab , or dc  2  ab  dc  2  ab > 0. Since dc is positive,
this inequality is satisfied if the condition in the proposition
holds. h
Note that t0 = t; so at least for the threshold, the socially-opti-
mal level and the optimal level of individual choice coincide. This
permits us to predict that social commerce firms will manage their
threshold levels at the socially-optimal level.
This desirable feature may not be sustained when we consider to-
tal marketing effort. Further, a ramification of the tragedy of the
commons shows the possibility of a phase transition, where the rela-
tionship between dc and ab specifies the sharp boundary of the phase
transition. We have already seen that a relationship between these
two terms presents the conditions in the Industry Performance
Proposition (P3). These conditions hold when dc is far larger than ab.
The next proposition goes further and provides another relationship
in a somewhat different format between these two terms, and this
relation is critical in triggering the tragedy of the commons:
Proposition 4. (The Trigger for the Tragedy of the Commons Proposi-
tion). Assume the symmetric situation as in Proposition 2. Now, consider
the following two cases that are mutually exclusive and complete.
 Case A. When dc > 2  ab, then there is a positive critical value T
such that the tragedy of commons occurs (E > E0) if the number
of social commerce firms exceeds this critical number (N P T). T
2ad
is larger than 1 and determined by T ¼ ad2cb .
74 D. Kim / Electronic Commerce Research and Applications 12 (2013) 69–77
Note: Here, = 1 is fixed, but this ratio does not affect the overall pattern of partitions.
Fig. 4. Parameter areas where the tragedy of commons can be avoided.
 Case B. When dc 6 2  ab, then for any number of firms N, the total
marketing effort falls short of the socially-optimal level
(E 6 E0).
Proof. E > E0 if and only if (a  d + 2c  b)  N < 2a  d  N  2a  d.
This expression can be rearranged into (2c  b  a  d)  N < 2a  d.
We have two cases to explore. The condition in Case A corresponds
to the situation where the left-hand side is negative, while the con-
dition in Case B guarantees that the left-hand side is non-negative.
Thus, in Case B the inequality E > E0 cannot hold unless N is nega-
tive, which is impossible. In Case A, E > E0 holds for N > ad2cb
2ad
 T.
Further, the denominator of T is always bigger than the numerator
under the condition in Case A, which guarantees that T > 1. h
The results of the Trigger for the Tragedy of the Commons Prop-
osition (P4) imply that one cannot expect that the social commerce
industry will be sustainable unless the condition in Case B comes
true. It depends on the number of social commerce firms whether
the industry in Case B survives. A limited number of social com-
merce firms may thrive only if the size of the industry is main-
tained at less than T.
Fig. 4 illustrates some partitions of the c–d parameter space
along the critical boundaries for a phase transition. The shaded re-
gions represent the area where the tragedy of the commons can be
avoided (the condition in P4) and the socially-optimal level can be
attained (the condition in P3). Note that the shaded regions occupy
a relatively small portion of the entire parameter space. Further,
even a small increase in the number of social commerce firms
(from 3 to 6 in Fig. 4) may significantly reduce the size of the
shaded region. Therefore, Case B together with the limited oppor-
tunity of N < T in Case A of Proposition 4 will be rarely observed in
practice, and the tragedy of the commons seems inevitable in most
situations.
a , we can see that T is larger than 2 and
By rearranging T into a22cb
d lows: T
dT d2 T
converges to 2 as d becomes large. Since dd
< 0 and dd2
> 0, T is
diminishing but will slowly converge to 2 as d goes infinity. How-
ever, T shows a different behavior when q  ab changes. Again, by
rearranging terms in T, we can get another expression for
2d
T ¼ d2cq , and dT
> 0 and d2 T
> 0 when d > 2c. Subsequently, T is
dq dq2
 
close to 2 when a is far larger than b ab  0 , and very rapidly in-
b d
creases to infinity as a approaches to 2c
ð> 1Þ. This behavior puts
an upper bound on the relative size between a and b: b cannot
be larger than d2ca. As a result, T appears more sensitive to ab than
to d as shown in Fig. 5. (See Fig. 5 and compare the scales in the
axes of figures.)
Since t0 = t and they do not depend on the number of social
commerce firms that use symmetric strategies, we can view the
performance structure from a different angle by defining two para-
metric functions based on my model: H ¼ HðtÞ  b  t and
J ¼ JðtÞ  t  ð
d  tÞ. Note that at a symmetric equilibrium, both

H() and J() are constant functions: specifically, H ¼  b2d and
d2

J ¼ 4 for both the social and individually-optimal levels (t0 and
t). Accordingly, the performance measure in Eq. (4) can be viewed
as if it were a function of E only:
c ¼ ða  E þ HÞ  ðJ  c  EÞ:
W ð5Þ
This characterization of system performance is similar to a well-
known system performance measure in ecology (Alroy 2001, Hardin
1968). One can interpret H and J as a location parameter and an eco-
logical capacity. The latter J is proportional to the average capability
and the former H, together with J, determines the generic perfor-
mance without marketing effort; that is, J  H (<0) corresponds to
the performance level when E = 0. From Eq. (5), we know that both
solutions to W c ¼ 0 in terms of E are positive, and Wc is maximized
at bE ¼ aJcH
2ac
.
There are two forces at work in Eq. (5). First, for a given poten-
tial market size with fixed H and J, more marketing effort by social
commerce firms means additional revenues: this is the first term
a  E + H in Eq. (5). In fact, in the early stage of the industry, excess
marketing expenditures contributed to the rapid growth of the en-
tire market for social commerce services (Anderson et al. 2012). In
South Korea, for example, the social commerce business has grown
into a US$1 billion industry over the last several years, and many
experts agree that the massive marketing activities have raised
consumer awareness. However, due to the fierce competition,
more marketing effort has resulted in a smaller population to tar-
get in the next period: the second term (J  c  E) in Eq. (5). The
overall effect of these two forces is reflected in the multiplicative
system performance measure.
The corollary below provides more streamlined expressions of
the conditions pertaining to the tragedy of the commons when
the tk’s are identical and fixed at some t such as t0 = t for some pol-
icy reasons, and ek is the only effective strategy for firm k.
Corollary (Alternative Conditions for the Tragedy of the Commons to
Occur). Let us assume that the tk’s are fixed at t and the
performance measure is given as in Eq. (5). With symmetric
strategies as in Proposition 2, each social commerce firm sets its
optimal marketing effort at tðdtÞ
c  N1
N2
, thereby, E ¼ tðdtÞ
c  N1
N .
There are two cases.
 Case A. When c  b < a  (d  t), then there is a positive critical
b such that the tragedy of the commons occurs ðE P b
value T EÞ
if the number of social commerce firms exceeds this critical
b Þ. Tb is larger than one and determined as fol-
number ðN P T
b  2aðdtÞ .
aðdtÞcb
 Case B. When c  b P a  (d  t), then for any number of firms,
the total marketing effort will fall short of the socially-optimal
level ðE < b
EÞ.
The proof follows procedures similar those for Propositions 3
and 4, and so it is omitted.
b > 1 in Case A of Corollary 5. This suggests that the
Note that T
tragedy of the commons is not inevitable, even in Case A when
 D. Kim / Electronic Commerce Research and Applications 12 (2013) 69–77
Fig. 5. Changes of critical value (T) (c is fixed at 1; thus, d > 2 and q < 0.5).
N<T b . Unfortunately though, T
b is likely to always be larger than 2,
and not small in most normal situations.
4. Conclusions
My analysis work reveals the weak points in the group-buying
business model associated with the social commerce industry. I
will review the main results and suggest some practical implica-
tions. I will focus on the value proposition that social commerce
firms should deliver to the market: they should seek a niche that
is different from online shopping malls and open markets. Some
social commerce firms have revised their prospectuses to clarify
how they are differentiating themselves, but they have not yet
proved to investors that their businesses can achieve sustainable
growth (Webster 2011, Wheeler 2011). I also will present some
guidelines and policy directions to address these difficulties.
First, my basic analysis under the symmetric competition
assumption argues that there will be a significant possibility of
the tragedy of commons. Although there is a gap between practice
and my assumption of homogeneous firms with symmetric strate-
gies, leading firms in the industry nevertheless seem to actually be
implementing fairly similar strategies. For example, the top three
social commerce firms in South Korea constitute almost 90% in
market share in terms of sales revenue, and their marketing ex-
penses in 2010 to 2011 ranged from US$5 million to US$5.5 million
(Baek 2011). Further, their service configurations and consumer
participation threshold levels, controlling for category of commod-
ity or service, also look remarkably similar (Kim 2011). Thus, a
symmetric model appears to be a good approximation of the cur-
rent situation.
I also have claimed that, under a certain conditions, the social
commerce industry is likely to experience a tragedy of commons
and face disruption. My findings suggest that the critical number
of firms for the tragedy of the commons to occur is quite small.
A small number of large social commerce firms in the industry
may lead the market toward problems. The arguments that I have
made suggest that the explosive growth of the social commerce
business may become too great to allow an effective business mod-
el to continue to exist.
Prior research offers insights on how to deal with the potential
for excessive competition and diversify social commerce business
models so that the original intention of social commerce can be
recovered. Liang et al. (2011), Stephen and Toubia (2009, 2010),
and Westland (2010) have demonstrated the potential of social
business, but increasing marketing expenses need to be addressed
to remedy future problems. Since the relationships between key
75
ratios in my analysis, dc and ab, play an important role in determining
whether the outcome will be a market failure, further analysis will
be useful to understand the mechanics a little better. For example,
if the unit costs of marketing are much larger than the average
capabilities of firms with similar ratios, then the opportunity to
avoid market failure will improve.
Thus, social commerce firms need to be aware of the real costs
of marketing in their business and industry. The current levels of
marketing expenses are too high because the firms are spending
on activities that are beyond the point of social optimality.5 This
distortion in the marketing-related efforts of social commerce firms
needs to be addressed. If these firms were forced to abide by the
rules and practices that normally are implemented among ex-
change-listed companies, their costs would increase further.
My comments suggest a policy direction to overcome the short-
comings of the current business model, and to encourage the
industry to develop sustainable business models for the future. A
set of regulations may be required to curb the unnecessary adver-
tisements and promotions to reduce overall marketing expendi-
tures. In fact, excessive competition in the social commerce
industry is becoming the focus of monitoring and regulation in
many countries. For example, the UK Office of Fair Trading investi-
gated Groupon’s British unit after receiving a complaint from the
country’s advertising watchdog (Larson 2011). The Korean Fair
Trade Committee also has shown serious concerns about social
commerce firms’ ability to comply with proper advertising codes
(Digital Daily 2011).
There are a number of different activities that tend to increase
the marketing costs of social commerce firms.6 The costs of running
a customer care center and operating a help desk tend to be sub-
sumed into overall marketing costs. Some major social commerce
firms in South Korea are starting to operate logistics centers and
implement refund policy for unused coupons too. Institutional and
legal devices to establish and enhance trust between social com-
merce firms and customers are another example of such measures.
5
Since only one social commerce provider, Groupon, has gone public so far, it is not
easy to accurately identify the financial status of most of the others. Many experts
look at the firms’ reports on their marketing activities and financial justification with
suspicion (Financial News 2011, MacMillan 2011a, Webster 2011). This is also
partially due to unconventional accounting practices in the industry. I argue that the
current marketing cost is underestimated, and the actual figure should go up
significantly if all the true marketing efforts are included.
6
To avoid confusion, I stress again that it has a positive connotation in my context
to increase the marketing costs. My point here is that the current marketing efforts
focus on promotions and advertising and overlook necessary efforts that should be
regarded as marketing activities.
76 D. Kim / Electronic Commerce Research and Applications 12 (2013) 69–77
These may include quality assurance through careful selection of
partners and service opportunities, self-inspection systems for con-
firmation of service specifications and other agreements, a legal obli-
gation that forces social commerce firms to subscribe to insurance
for damage compensation, and so on.
All of these measures should have been employed in the early
stages of the social commerce industry. For example, Groupon
introduced a deal evaluation system to build trust with its custom-
ers and partners. After deploying this system, Groupon’s expenses
decreased due to a sharp increase in other kinds of marketing costs
(Patel 2011). Similar trends have been observed in South Korea
after the Korean Fair Trade Committee strongly encouraged that
social commerce firms should establish standardized refund poli-
cies, together with damage compensation plans. Actual marketing
expenses diminished and over-issuance of free coupons also de-
clined after institutional measures were introduced to raise the
real costs of marketing.
Last but not least, we seek for technological innovations to
strengthen the potential merits of the social commerce business
model. One of my results implies that it is necessary to reduce
the gap between a firm’s capability index and its service configura-
tion threshold to avoid the tragedy of the commons, or at least
diminish its likelihood. For example, the new launch of location-
based services for smart phone users may narrow this gap since
the service will provide more information to consumers, and in-
crease the chance of successful service offerings. Some players such
as Groupon Mobile in the US and TicketMonster Now in South Kor-
ea are already moving in this direction (Baek 2011). These firms
sell real-time mobile coupons to smart phone users and provide
customized service packages targeted to specific consumer seg-
ments. Such activities differentiate these social commerce firms
from online shopping malls and other e-commerce sites, as dis-
cussed by Heidemanna et al. (2012). When 4G networks and inter-
active mobile communications proliferate, location-based services
will be one of the most promising areas for social commerce. The
technology will give the firms more capacity to configure service
offerings with a wide variety of negotiation and coordination capa-
bilities to establish partnerships with local firms. One of the most
promising directions along this line of development will be inte-
grating location-based, personalized social commerce services
with mobile payment systems using near field communication.
In conclusion, my analysis suggests that social commerce firms
should return to their original value proposition, where the mar-
keting activities focus more on word-of-mouth diffusion through
social networks than through aggressive multi-media advertising
campaigns. We believe that this will be the best way to strike a bal-
ance between the ‘‘social’’ and ‘‘commerce’’ aspects of the social
commerce business model (Knowledge at Wharton 2006, Kumar
and Rajan 2012a, 2012b; Urstadt 2008). Although social commerce
firms have engaged in massive advertising campaigns to generate
public awareness and interest in the early stage of the industry’s
lifecycle, it is time to return for them to emphasize the social as-
pects of their business models now.7
References
Alroy, J. A multispecies overkill simulation of the end-Pleistocene mega faunal mass
extinction. Science, 292, 2001, 1893–1896.
Anand, K. S., and Aron, R. Group buying on the web: a comparison of price-discovery
mechanisms. Management Science, 49, 11, 2003, 1546–1562.
Anderson, M., Sims, J., Price, J., and Brusa, J. Turning ‘‘like’’ to ‘‘buy’’: social media
emerges as a commerce channel. White paper, Booz & Company, New York, NY,
7
Polyvore, Shopkick, Swipely and others in the US are trying new business models
different from group-buying (Anderson et al. 2012, Kumar and Rajan 2012a, 2012b;
MacMillan 2011b). Their service offerings emphasize the social aspect in their value
proposition based on local commerce information.
January 20, 2012. Available at www.booz.com/media/uploads/BaC-
Turning_Like_to_Buy.pdf.
Baek, B. S. Ticket Monster and Coupang, head-to-head competition for the industry’s
number one position (in Korean). Economy Today, November 25, 2011.
Available at www.eto.co.kr/news/outview.asp?Code=20111125145743203&
ts=133239.
Cavazza, F. The six pillars of social commerce. Forbes Magazine, February 1, 2012.
Available at www.forbes.com/sites/fredcavazza/2012/02/01/the-six-pillars-of-
social-commerce/.
Chiang, A. C. Fundamental Methods of Mathematical Economics, 4th edition. McGraw-
Hill/Irwin, New York, NY, 2006.
Chung, W., and Chen, L. Group-buying e-commerce in China. IT Pro, 2012, 24–30.
Curty, R. G., and Zhang, P. Social commerce: looking back and forward. Proceedings
of the American Society for Information Science and Technology, 48, 1, 2011, 1–10.
Digital Daily. Special report on brightness and darkness of social commerce in Korea
(in Korean). April 6–20, 2011. Available at www.ddaily.co.kr/news/
news_view.php?uid=76416.
Financial News. Special report on social commerce (in Korean), December 18, 2011.
Available at www.fnnews.com/
view?ra=Sent0901m_View&corp=fnnews&arcid=09224
94751&cDateYear=2011&cDateMonth=12&cDateDay=18.
Fudenberg, D., and Tirole, J. Game Theory. MIT Press, Cambridge, MA, 1991.
Gu, B., Park, J., and Konana, P. The impact of external word-of-mouth sources on
retailer sales of high-involvement products. Information Systems Research, 23, 1,
2012, 182–196.
Greco, G. M., and Floridi, L. The tragedy of the digital commons. Ethics and
Information Technology, 6, 2004, 73–81.
Hardin, G. The tragedy of the commons. Science, 162, 1968, 1243–1248.
Heidemanna, J., Kliera, M., and Probstb, F. Online social networks: a survey of a
global phenomenon. Computer Networks, 56, 18, 2012, 3866–3878.
Jing, X., and Xie, J. Group buying: a new mechanism for selling through social
interactions. Management Science, 57, 8, 2011, 1354–1372.
Korea Chamber of Commerce and Industry. A consumer satisfaction survey on social
commerce services (in Korean). Research report. Seoul, South Korea, March 8,
2011. Available at www.korcham.net/EconNews/KcciReport/CRE01102R.asp?
m_chamcd=A001&m_dataid=20110308001&m_page=1& m_query=TITLE&m_
queryText=%BC%D2%BC%C8%C4%BF%B8%D3%BD%BA.
Kim, Y. H. Social commerce: current market situations and policy issues. Korea
Information Society Development Institute Issues Report, 23, 11, 2011, 41–63 (in
Korean).
Knowledge at Wharton. Dot-com bubble, part II? Why it’s so hard to value social
networking sites? October 4, 2006. Available at knowledge.wharton.upenn.edu/
article.cfm?articleid=1570.
Kumar, V., and Rajan, B. Social coupons as a marketing strategy: a multifaceted
perspective. Journal of the Academy of Marketing Science, 40, 1, 2012, 120–136.
Kumar, V., and Rajan, B. The perils of social coupon campaigns. MIT Sloan
Management Review, Summer 2012, 15–16.
Larson, E. Groupon advertising practices being probed in U.K. on promotions,
savings. Bloomberg, December 2, 2011. Available at www.bloomberg.com/
news/2011-12-02/groupon-advertising-practices-are-being-probed-by-u-k-
antitrust-regulator.html.
Lee, E. M. Global market survey on social commerce. Korea Information Society
Development Institute Issues Report, 23, 3, 2011, 36–44 (in Korean).
Liang, T. P., Ho, Y. T., Li, Y. W., and Turban, E. What drives social commerce: the role
of social support and relationship quality. International Journal of Electronic
Commerce, 16, 2, 2011, 69–90.
MacMillan, D. Groupon’s stumbles may force it to pare back size of IPO. Bloomberg
Online (October 3, 2011a). Available at www.bloomberg.com/news/2011-10-
03/groupon-s-stumbles-seen-paring-back-size-of-ipo-as-investor-interest-
wanes.html.
MacMillan, D. LivingSocial aims to be different from Groupon. Business Week
Online (September 22, 2011b). Available at www.businessweek.com/magazine/
livingsocial-aims-to-be-different-from-groupon-09222011.html.
MacMillan, D. Groupon China venture said to fire workers for poor performance.
Bloomberg Online (August 24, 2011c). Available at www.bloomberg.com/news/
2011-08-23/groupon-china-joint-venture-said-to-fire-workers-for-poor-
performance.html.
Marsden, P. Social Commerce: Monetizing Social Media. Syzygy, Hamburg, Germany,
2010.
Marsden, P., and Chaney, P. The Social Commerce Handbook: 20 Secrets for Turning
Social Media into Social Sales. McGraw-Hill, New York, NY, 2012.
Patel, K. Groupon marketing spending works almost too well. Ad Age Digital
(November 12, 2011). Available at adage.com/article/digital/groupon-
marketing-spending-works/230777.
Reibstein, D. How sustainable is Groupon’s business model? Knowledge at Wharton
(May 25, 2011). Available at knowledge.wharton.upenn.edu/article.
cfm?articleid=2784.
ROA Holdings. The rapidly expanding social commerce market of South Korea and
Japan. Research report (February 21, 2011). Available at global.roaholdings.
com/report/research_view.html?type=country&num=143.
Sirakoulis, G. C., and Karafyllidis, I. G. Cooperation in a power-aware embedded-
system changing environment: public goods games with variable multiplication
factors. IEEE Transactions on Systems, Man, and Cybernetics – Part A, 42, 3, 2012,
596–603.
Stephen, A. T., and Toubia, O. Explaining the power-law degree distribution in a
social commerce network. Social Networks, 31, 2009, 262–270.
 D. Kim / Electronic Commerce Research and Applications 12 (2013) 69–77
Stephen, A. T., and Toubia, O. Deriving value from social commerce networks.
Journal of Marketing Research, 47, 2, 2010, 215–228.
Urstadt, B. Social networking is not a business. MIT Technology Review (July/August
2008). Available at www.technologyreview.com/business/20922/.
Webster, K. Groupon’s business model: bubble or the real deal? Pymnts.com
(September 19, 2011). Available at pymnts.com/commentary/pymnts-voice/
groupon-s-business-model-bubble-or-the-real-deal/.
Wei, Y., Straub, D. W., and Poddar, A. The power of many: an assessment of
managing Internet group purchasing. Journal of Electronic Commerce Research,
12, 1, 2011, 19–43.
77
Westland, J. C. Critical mass and willingness to pay for social networks. Electronic
Commerce Research and Applications, 9, 1, 2010, 6–19.
Wheeler, R. Groupon gone wrong! Harvard business fellow’s warning to investors
and entrepreneurs. Pymnts.com, August 23, 2011. Available at pymnts.com/
briefingroom/shopping-and-social-buying/social-shopping-and-social-buying/
groupon-gone-wrong-a-warning-to-investors/.