BULKER
A Mediator System Grounded on Social Networks for Online Trading of Batches of
Products
Martín López-Nores, Yolanda Blanco-Fernández, José J. Pazos-Arias and Manuela I. Martín-Vicente
Department of Telematics Engineering, University of Vigo, Vigo, Spain
Keywords:
E-commerce, Social Networks, Newsfeed.
Abstract:
We present a system that mediates between consumers and providers for the trading of batches of products
on the Internet, leaning on social networks as a vehicle for the distribution of publicity and the formation of
groups of interest. The system enables new means for providers to reach groups of users potentially interested
in their products, reducing the burden of specifying the characteristics of the target audience of each campaign.
As regards the consumers, they are offered a convenient and reliable mechanism to purchase products at a
lower price, with the additional advantage of a new form of targeted, trust-based advertising. In terms of
exploitation, a clear business model arises, based on obtaining a profit margin from the discounts enjoyed by
those who decide to participate in the purchase of a batch of products, as well as on the optimization of the
distribution logistics in the case of material assets.
1 INTRODUCTION
The Internet has become a mainstream medium to de-
liver offers about many different types of products
to a great number of users (potential customers) who
navigate the web as a common practice in their daily
lives. Among the most advanced initiatives to exploit
this capability, we can highlight tools like Google Ad-
Words and Google AdSense, which allow to buy and
sell advertising space in the context of the results ob-
tained in response to the users’ searches in the web.
On the one hand, AdWords allows the advertisers to
classify their offerings as per a set of keywords and,
thereafter, pay for the ads to appear next to the re-
sults given by Google to a search that includes some
of those terms. On the other hand, AdSense provides
a mechanism for web site owners to profit from their
spaces by perceiving a fee in exchange for displaying
advertisements.
The implantation of tools like the aforementioned
ones soon revealed a number of limitations that ham-
per their goals to a significant degree (Research and
Markets, 2007). These limitations have to do with
(i) the difficulty to accurately define the keywords
linked to each advertisement and (ii) the imposibility
of displaying the ads out of the context of the users’
searches. This way, we havewitnessed a shift to a new
paradigm of social publicity, promoted by the con-
solidation of the so-called Web 2.0 and characteristic
communication media as blogs and, especially, social
networks. These technologies enable a shared inter-
active space that bears much potential for the users,
inasmuch as they can get to know the opinion of other
individuals about the products in offer, to read mes-
sages written by experts (in principle) not biased by
commercial interests, to gather information about the
reputation of a given online shop, etc. All of that,
with no need to explicitly declare interest in a given
product of brand through a web search.
Web 2.0 technologies enjoy great popular-
ity nowadays and their influence has grown
rapidly, to the point that studies carried out by
PriceMinister.es (PriceMinister, 2009) or Econsul-
tancy.com (Econsultancy, 2010) in 2009 and 2010,
respectively, place blogs and social networks as
the second most influential sources of information
among users when it comes to making up their minds
about purchasing a given product, reaching 35%
from the 8% estimated in 2008. Only the direct
opinion of relatives and friends has greater influence.
Encouraged by such studies, new solutions like
Facebook Ads or AdLemons have come into scene,
exploiting social networks and blogs, respectively, as
advertising platforms:
• On the one hand, Facebook Ads allows the adver-
tisers to manually create advertising campaigns
552
López-Nores M., Blanco-Fernández Y., Pazos-Arias J. and Martín-Vicente M..
BULKER - A Mediator System Grounded on Social Networks for Online Trading of Batches of Products.
DOI: 10.5220/0003934505520557
In Proceedings of the 8th International Conference on Web Information Systems and Technologies (WEBIST-2012), pages 552-557
ISBN: 978-989-8565-08-2
Copyright
c
2012 SCITEPRESS (Science and Technology Publications, Lda.)
that will be visible within the context of the Face-
book social network. This platform bases its busi-
ness model around two main features: (i) the
ability to segment the publicity by characterizing
the target audiences as per the demographic in-
formation stored in the users’ profiles (e.g. res-
idence, age, gender or marital status) and (ii)
the viral propagation of the products or brands
by the users’ themselves, grounded on the news-
feed mechanism. Because of newsfeed, every
time a user expresses an opinion about an adver-
tised product, it is automatically communicated to
his/her contacts in the social network. The adver-
tisers pay for each newsfeed message that carries
their products or brands, while benefitting from an
advertising model based on trust, in which their
campaigns are more likely to catch on among the
users for going along with the opinions of some
people they know.
• On the other hand, AdLemons comes as a platform
devised specifically to manage publicity in blogs.
The idea is to enhance the targeting of advertising
by taking advantage of (i) the usually high level of
specialization of the contents appearing in blogs,
and (ii) the consequent formation of communities
of interest around them. The owners of the blogs
benefits from the fees paid by the advertisers in
exchange for offering space for brands and prod-
ucts in their pages.
In spite of their unnegligible advantages, the cur-
rent models of social publicity do not fully exploit the
potential of targeted advertising. In the case of Face-
book Ads, for example, the advertisers can merely
identify target audiences by their demographic data,
obviating such relevant factors as the preferences and
needs that could well be inferred from the users’ ac-
tivities in the social network. It is clear to us that,
having knowledge about the users beyond their per-
sonal data, it would be possible to reach the people
potentially most interested in each type of product in
a more elegant and powerful manner.
Our proposal in this paper has been inspired by
a less sophisticated approach to gather users with a
purchase in mind over the Internet. For many years,
people have resorted to online forums in order to get
in touch with others who could be interested in buy-
ing a number of units of a given product, in order to
get a better price per unit from some provider than
proceeding invidivually. This approach has not been
addressed in e-commerce research for two main rea-
sons:
• First, forums merely provide a rendezvous to
gather a number of purchasers, with neither a
way to ensure that there will be providers in the
position to satisfy their demands, nor to assure
the payment from each one who had in princi-
ple agreed to participate in the batch purchase. In
the absence of such guarantees, it usually happens
that there seems to be a sufficient number of users
to make a purchase, but many of them disappear
when it is finally time to transfer money. This
way, forums turn out to be an unreliable medium
for both consumers and providers.
• Secondly, it is clear that forums leave just too
much work in the hands of the users, since they
have to take the initiative to get together and
manage the whole process of batch purchase (in-
cluding the search for and communication with
providers), interacting actively without any kind
of assistance. Indeed, a similar shortcoming can
be found in AdLemons and Facebook Ads, inas-
much as the advertisers have to define manually
their campaigns by identifying the most suitable
audience segments for each kind of product.
We present a system that addresses the aforemen-
tioned limitations by acting as a mediator between
users and providers for the trading of batches of prod-
ucts over the Internet, taking advantage of the poten-
tial of social networks as a vehicle of trusted public-
ity and a means to gain customer loyalty. The sys-
tem, called Bulker, gathers knowledge about the pref-
erences and needs of the users through the different
social networks in which they participate, in order to
identify the products that best match their interests
and to assist in the formation of groups. It also acts
to locate providers able to offer batches of products
in advantageous conditions for the users, as well as to
control the propagation of publicity through networks
of contacts. Finally, the system includes secure means
of payment in order to guarantee the commitment to
purchase from all the users interested in the batches
in offer.
The overall design and the main features of the
Bulker are described in Section 2. Next, Section 3
presents our plan of development and deployment. Fi-
nally, Section 4 provides a summary of conclusions,
including an outline of the business model envisaged
for commercial explotation.
2 OVERALL DESIGN AND
FUNCTIONALITIES
The Bulker appears as a mediator between consumers
and providers for the trading of batches of products
over the Internet, with a double objective:
• To Search for Customers. A provider registers
BULKER-AMediatorSystemGroundedonSocialNetworksforOnlineTradingofBatchesofProducts
553
one offer involving a batch of a given product, in-
dicating a target price subject to the selling of a
certain number of units. The goal is to find a suf-
ficient group of users interested in the purchase,
each one committing to paying for one or several
units.
• To Search for Providers. Having identified a
group of users interested in purchasing a given
product, the goal is to find the provider that offers
the best price for a batch of it.
To fulfill these goals, the system relies internally
on a database that stores information about registered
providers and their offers, and about registered users
and the groups of interest they form. In addition, it
leans on social networks as the basic mechanism for
the distribution of offers. Figure 1 shows a diagram of
all the actors involved in the operation of the Bulker,
whose interactions are described in the following sub-
sections:
2.1 The System towards Product
Providers
Product providers interact directly with the database
of the system to register and, thereafter, to perform
some of the following actions through dedicated web
pages with business-to-business (B2B) orientation:
• To introduce a new offer (see Figure 2 for a snap-
shot of the interfaces), identifying the product in
question and indicating the number of units that
make up the batch to be sold, the price to be paid
in case of achieving a sufficient number of pur-
chasers, and the period of time during which the
offer is valid.
• To consult the own offers or to browse those of
other providers in an attempt to improvetheir con-
ditions (by reducing the target price, the required
number of purchasers and/or the validity dead-
line).
• To modify the own offers. It is possible to modify
an offer in any way while no user has commit-
ted to it (we shall explain what this means next);
otherwise, it is only possible to improve the orig-
inal conditions. Such modifications can be mo-
tivated by warnings sent to the provider by the
Bulker system itself, for example, because it has
been able to gather a greater number of purchasers
than required, or because it has gathered that num-
ber much sooner than the deadline.
• To remove own offers at any time, leaving the
users who might have committed to them unat-
tended and, thereby, assuming a risk of gaining
bad reputation.
• To browse a list of groups of interest created by
the users asking for products for which there are
no registered offers in the system, just in case it
were possible to introduce one. The system itself
can send warnings to the providers in case it de-
tects overlappings between their areas of activity
or their catalog and the products requested by the
users.
2.2 The System towards the Internet
Users
The Bulker stores in its database information about
the users, derived from their actions in social net-
works. Those actions are conveniently recorded
by software applications named client applications,
which take either the form of (i) classical applications
of an operating system, or (ii) web applications resid-
ing in the development platforms provided by some
social networks. In the first case, the client applica-
tion itself can access several social networks in behalf
of the users; in the second, each client application can
work only with data from the corresponding social
network. The users can access the system’s function-
alities in a transparent manner, regardless of whether
they use a single client application or several ones.
The system gathers knowledge about a registered
user in two ways: (i) explicitly, by means of forms
in which the user introduces personal data and top-
ics or products of his/her interest, and (ii) implicitly,
by processing the information stored in his/her pro-
files in the social networks, as far as the permissions
granted by the user allow the client application to go.
This knowledge is processed in order to identify the
most suitable users to warn about the registration of
new offers (some details will be given in Section 2.4).
The users will also be able to take the initiative to
browse the list of available offers (using querying
forms provided by the client applications) and to cre-
ate or browse groups of interest created by others in
order to purchase batches of products for which there
are no registered offers yet.
When showing the user the details of an offer, the
client applications will provide three possibilities:
1. To commit to the offer, which implies providing
the system with the data needed to make the pay-
ment if the requested number of purchasers is fi-
nally achieved. As a core feature of the proposal,
this decision triggers the dissemination mecha-
nisms of the social networks (newsfeed and other
variants available) to automatically communicate
it to the user’s contacts, in the shape of announce-
ments with the following text:
WEBIST2012-8thInternationalConferenceonWebInformationSystemsandTechnologies
554
Figure 1: Diagram of the Bulker system.
“<USER> has committed to
participate in the purchase
of a batch of <PRODUCT>
for <PRICE> euros per
unit. Having gathered
<NUMBER> purchasers thus
far, <NUMBER> more are needed
before <DEADLINE> to get the
batch. Press here for more
information.”
2. To propagate the offer to contacts who may be in-
terested, who will receive warnings with the fol-
lowing text:
“<USER> believes you could be
interested in participating
in the purchase of a batch of
<PRODUCT> for <PRICE> euros
per unit. Having gathered
<NUMBER> purchasers thus
far, <NUMBER> more are needed
before <DEADLINE> to get the
batch. Press here for more
information.”
3. To ignore the offer.
In the first two cases, the user will have left an-
nouncements in the social network that will serve to
disseminate the registered offers among other users,
registered or not in the system. If a non-registered
user decides to take part in an offer, the client appli-
cation faces him/her with the interfaces needed to reg-
ister, to proceed normally thereafter with the commu-
nication of payment data and the dissemination of the
decision.
In any case, the interactions of registered users
with the offer announcements are recorded in the sys-
tem’s database, so as to ensure the traceability of the
propagation. This information lets the system to aug-
ment the knowledge available about the users’ pref-
erences and to define metrics of their respective lev-
els of activity and influence. It is possible to con-
sider offering additional discounts, for example, to
users who have participated in more batch purchases
or who have intervened in the propagation of offers
that ended up gathering a sufficient number of pur-
chasers.
2.3 About Payment Gateways and
Financial Entities
In order to reinforce the reliability of the system
against improper use (remember the comments about
forums in Section 1), the commitment of a user to
an offer requires him/her to guarantee the payment
of the price indicated in it. To this aim, the user
has to provide the system with the data needed to
charge that amount to his/her bank accounts, credit
cards or any other means offered by the financial en-
tities that ultimately manage his/her money. The exe-
cution of the payment will not be realized unless the
system gathers, at least, the number of purchasers re-
quested by the provider within the indicated deadline
BULKER-AMediatorSystemGroundedonSocialNetworksforOnlineTradingofBatchesofProducts
555
Figure 2: Introducing a new offer.
—obviously, the amount to be charged will be that in-
dicated in the offer at the moment, which could have
been revised downwards since some users committed
to them. Both the communication of data and the ex-
ecution of the payment will be securely handled by
payment gateways.
2.4 About the System's Administrators
and its Internal Operation
The system’s administrators can use a set of inter-
faces to manage the information stored in the database
about providers, offers, users and groups of interest.
Among other options, it is possible to supervise the
selection of the most suitable users to notify about
new offers, whose identification is based on both the
knowledge available about them and estimations of
the influence they might have among their contacts in
the social networks. To this aim, the system considers
internally such aspects as the following ones:
• Characteristics of the registered offers (product,
target number of purchasers, price and period of
validity), to be contrasted with the users’ demo-
graphic information and preferences, as well as
with information about their previous purchases.
• Information about the users who have checked the
details of the offers (either to end up committing
to, propagating or ignoring them) and metrics de-
rived from the traceability of their propagation.
• Information about the groups of interest created
by the users asking for products for which there
are no registered offers in the system (what prod-
ucts they refer to, how many users are involved,
etc).
• Metrics of social influence over the networks of
contacts of the users, including those presented
in (Goh et al., 2003; Mason et al., 2007): degree,
centrality, betweenness, etc.
The Bulker can automatically identify cases in
which there exist two or more offers of batches of
the same product but none has gathered a sufficient
number of purchasers, even though it would be possi-
ble to fulfill the requirements of one by redistributing
some of the purchasers. The administrators can de-
cide whether to proceed with the redistribution and
how to do it (in principle, favouring the best prices
and the users who have participated in more offers
or get more of their contacts to do so before). The
system also provides mechanism to detect synergies
among different groups of interest or between groups
of interest and valid offers (e.g. for referring to very
similar products). This logic is driven by a varia-
tion of the semantic reasoning mechanisms of the rec-
ommender system presented in (López-Nores et al.,
2010), which deals with several ontologies that char-
acterize and interrelate user preferences and products.
2.5 About Distribution Services
Inasmuch as the Bulker system works with batches of
products, whenever these are material assets it is pos-
sible to optimize the distribution logistics as per the
purchasers’ addresses, in order to reduce the postage
WEBIST2012-8thInternationalConferenceonWebInformationSystemsandTechnologies
556
and packing expenses. The system can interact with
several providers of distribution services (again, using
dedicated web pages with B2B orientation) in quest
for the best options in each case, considering possibil-
ities of aggregatingseveral units (from the same batch
or from different ones) along certain routes while
bearing in mind questions of delivery times.
3 DEVELOPMENT AND
DEPLOYMENT PLAN
The Bulker system is being implemented as part of
the undergraduate thesis projects of several students
from the Telecommunication Engineering School of
the University of Vigo, providing client applications
for the Facebook and LinkedIn social networks. The
first operativeversion of the application is expected to
be released by the second half of 2012, and we have
reached agreements with online providers of different
types of products (especially electronic devices, furni-
ture and sports equipment) as well as with supermar-
ket retailers working in our region. This deployment
will serve to fine tune the internal processing of the
system —especially the semantic reasoning mecha-
nisms, which have been borrowed from a slightly dif-
ferent domain of application.
4 CONCLUSIONS
We have presented the main ideas behind a system de-
signed to facilitate the trading of batches of products
over the Internet, in advantageous conditions for both
users and providers. The system enables a clear busi-
ness model based on obtaining a profit margin from
the discounts enjoyed by those who decide to partici-
pate in the purchase of a batch of products, as well as
on the optimization of the distribution logistics in the
case of material assets.
One of the main advantages with regard to pre-
vious solutions has to do with ensuring proper use
of a rendezvous point by having each interested user
commit to paying for the requested units of a prod-
uct if it is possible to gather a sufficient number of
purchasers during the period of validity of an offer.
Besides, the system harnesses the power of social net-
works to enable a trust-based and non-invasive means
of distributing publicity, which is the cornerstone of
its twofold operation as a searcher of purchasers and
a searcher of providers. The providers are freed from
the task of characterizing the target audiences of each
campaign, thanks to the implementation of reasoning
mechanisms to match users’ demographic data and
preferences with the details of each available offer.
ACKNOWLEDGEMENTS
This work has been partially funded by the Ministe-
rio de Educación y Ciencia (Gobierno de España) re-
search project TIN2010-20797 (partly financed with
FEDER funds) and by the Consellería de Educación
e Ordenación Universitaria (Xunta de Galicia) incen-
tives file CN 2011/023 (partly financed with FEDER
funds).
REFERENCES
Econsultancy (2010). Value of social media report. http://
econsultancy.com/.
Goh, K., Kahng, O., and Kim, D. (2003). Betweenness-
centrality correlation in social networks. Physical Re-
view, 67.
López-Nores, M., Pazos-Arias, J. J., García-Duque,
J., Blanco-Fernández, Y., Martín-Vicente, M. I.,
Fernández-Vilas, A., Ramos-Cabrer, M., and Gil-
Solla, A. (2010). MiSPOT: Dynamic product place-
ment for digital TV through MPEG-4 processing and
semantic reasoning. Knowledge and Information Sys-
tems, 22(1):101–128.
Mason, W. A., Conrey, F. R., and Smith, E. R. (2007). Sit-
uating social influence processes: Dynamic, multidi-
rectional flows of influence within social networks.
Personality and Social Psychology Review, 11:279–
300.
PriceMinister (2009). III estudio sobre comercio elec-
trónico en España (3rd study of electronic commerce
in spain). http://www.priceminister.es/.
Research and Markets (2007). Social advertising:
Analysing the benefits and challenges. http://
www.researchandmarkets.com/reports/c74774.
BULKER-AMediatorSystemGroundedonSocialNetworksforOnlineTradingofBatchesofProducts
557
