Increasing Trust Towards eCommerce
Privacy Enhancing Technologies Against Price Discrimination
Christos Makris
1
, Konstantinos Patikas
1
and Yannis C. Stamatiou
2,3
1
Dept. of Computer Enginering and Informatics, University of Patras, Rio, Patras, 26504, Greece
2
Dept. of Business Administration, University of Patras, Rio, Patras, 26504, Greece
3
Computer Technology Institute & Press – “Diophantus”, N. Kazantzaki, University of Patras, Rio, Patras, 26504, Greece
Keywords: Price Discrimination, Privacy Enhancing Technologies, Privacy ABCs, eCommerce,
Personally Identifiable Information (PII).
Abstract: Price discrimination is a recently introduced practice in the domain of eCommerce. It is manifested by the
appearance of different prices when the same product is browsed by different prospective buyers, based on
their profiles. Thus, for instance, the price of an item may increase at the instance it is browsed by a user
coming from a rich neighbourhood or has performed a series of purchases of expensive objects in the past.
Price discrimination can lead to decrease of profits and loss of clientele, in the long run, as well as decrease
of people’s trust towards eCommerce. In this paper, we propose the deployment of Privacy Enhancing
Technologies in order to handle users’ personal information. These technologies empower users to have
command over their own privacy by allowing them to reveal only what is absolutely necessary (minimal
disclosure principle), or what they agree to reveal, in order to use a service avoiding any Personally
Identifiable Information (PII). Thus, eCommerce services that employ such technologies for handling their
clients’ personal data can attract more loyal clients, increase their popularity while, at the same time, suffer
from minimal client data and company image loss in case of a massive customer data theft attacks.
1 INTRODUCTION
Price discrimination, in the context of privacy, refers
to the display of different prices (usually higher) for
the same product to different users depending on their
profile elements such as, for example, area of
residence as well as purchasing behaviour and
history.
Price discrimination is neither a new term nor
describes, necessarily, a negative phenomenon in
general. Its usual content in economy is that the price
of a single item may change in three distinct ways or
price discrimination degrees: 1) [First Degree]
Charging different price for every item consumed, 2)
[Second Degree] Charging different prices for
different quantities, and 3) [Third Degree] Charging
different prices to different consumers for the same
item. The third degree price discrimination is the most
usual one and it is this type to which we refer to in
this paper.
Price discrimination is a complex phenomenon
with multiple facets and plays a central role in
shaping prices in economy and enterprise competition
environments (Armstrong, 2006). However, in the
context of our paper, we will study price
discrimination in connection to user privacy, why it
is not beneficial to commercial organizations and how
it can be overcome by suitable privacy enhancing
mechanisms.
Price discrimination, although it appears as
something beneficial the for eCommerce vendors
since it may lead to profit increase, it nevertheless
poses a severe danger for the future of eCommerce
due to the decrease of people’s trust. Beyond the
moral issue of violating the right of users to be offered
the correct price, price discrimination hampers the
widespread acceptance of eCommerce services
because of fear of users that they can be the victims
of such discrimination.
Work performed in the context of price
discrimination within the eCommerce domain, which
is related to users’ privacy, has been carried out with
respect, mainly, to identify it rather than prevent it
(see Mikians et al. 2012 and Mikians et al. 2013).
Since price discrimination is, essentially, based on
gathering information about users, our view is that by
Makris, C., Patikas, K. and Stamatiou, Y.
Increasing Trust Towards eCommerce: - Privacy Enhancing Technologies Against Price Discrimination.
In Proceedings of the 12th International Conference on Web Information Systems and Technologies (WEBIST 2016) - Volume 1, pages 25-31
ISBN: 978-989-758-186-1
Copyright
c
2016 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
25
deploying technologies that protect users’ privacy
towards eCommerce sites can eliminate or, at least,
limit the price discrimination phenomenon. These
technologies, termed Privacy Enhancing
Technologies, or PETs for short (see, e.g., Huberman,
et al. 1999), give the power to users to reveal about
themselves only what is absolutely necessary in order
to use a service avoiding, thus, to reveal any
Personally Identifiable Information (PII)
(Narayanan, et al. 2010). In other words, contrary to
customary user authentication methods such as the
X509 certificates, which reveal all identity
information towards a service, PETs allow users to
reveal only a subset of their identity elements. Thus,
for instance, they may choose not to reveal an element
which they suspect or know that can subject them to
price discrimination, such as their job or place of
residence. Moreover, PETs allow users to present
themselves with unlinkable, but certified,
pseudonyms which are unlinkable and, thus, avoid
profiling which is a powerful means for imposing
price discrimination.
Our position, as we explain in this paper for the
first time, to the best of our knowledge, is that PETs
can boost trust towards eCommerce services in
various ways. It may appear, from a superficial
consideration, that eCommerce players would be
reluctant to adopt PETs because they will prevent
them from imposing price discrimination on their
clients. However, we believe that this reluctance can
be easily overcome since PETs can bring to
eCommerce players the following advantages:
1. Establishment of trust and confidence of
users in eCommerce, since they will know
that their behaviour is neither monitored nor
profiled and, thus, price discrimination is not
possible.
2. Increased profits precisely because of the
increased trust of users towards eCommerce
services that deploy PETs. In other words
users, based on the functionality of PETs,
will be convinced that no price
discrimination can take place and will tend
to make more purchases.
3. Creation of a trust environment among
sellers and buyers in the eCommerce
ecosystem, which is expected to develop,
further, eCommerce and increase its wider
acceptance.
Based on the above, in this paper we present our
approach on how PETs can be deployed towards the
reduction of price discrimination using a recently
released PET technology called Privacy-ABCs.
Privacy-ABCs was the outcome of the research
project ABC4Trust (for a summary see Sabouri, at al.
2012 and for all the details of the project see the
ABC4Trust project book Rannenberg, et al. 2015).
2 PRIVACY-ABCs
Many electronic applications and services require
authentication of participants in order to verify their
eligibility to use them. Most authentication methods
however, such as password based on X.509 based,
have the negative consequence that the service
actually knows who the user is since all the elements
of the user’s identity are revealed. If we exclude
critical applications, such as eBanking for instance, in
which full user authentication is mandatory, in most
other online commercial applications it is not
necessary for users to reveal their full identity or some
of their identity elements. If they do, this leads to
privacy problems, such as profiling, linking of online
activities to a user, and history of purchases which in
turn gives rise, by some eCommerce sites, to the
phenomenon of price discrimination, based on what
is known (to the sites) about the users (see Acquisti
and Varian 2005).
The main Privacy Enhancing Technologies
(PETs) available until the completion of the
ABC4Trust project in February 2015 were IBM’s
Idemix (Camenisch and Lysyanskaya, 2001,
Camenisch, and Lysyanskaya, 2004, and Camenisch
and Groß 2012) and Microsoft’s U-prove (Brands,
2000, and Brands, et al. 2007). The ABC4Trust
project unified, transparently, these two technologies,
adding more features along the way, and developed a
new technology called Privacy-ABCs which
supports, in an interoperable way, both the two pre-
existing technologies.
Figure 1: Entities and interactions diagram.
This section provides a brief description of the
Privacy-ABCs elements and functionalities which we
User
Issuer Revocation Authority
Verifier
Inspector
credential issuance
token presentation
revocation info retrieval
revocation info retrieval,
credential revocation
credential revocation
token inspection
WEBIST 2016 - 12th International Conference on Web Information Systems and Technologies
26
will exploit in order to guard users against price
discrimination. Figure 1 gives an overview of the
different Privacy-ABCs entities and their interactions
(see Sabouri, et al. 2012).
We start our description of the basic privacy
architecture entities by the Issuer. The Issuer
constructs and provides credentials (e.g. on a smart
card) containing identity attributes to the User. Upon
request by the User, the Issuer generates the
credentials by executing a credentials issuance
protocol and, then, provides them to the User. The
information that is contained in the credentials comes
from either the User herself or the Issuer, if the Issuer
already has this information. The credentials are
signed by the Issuer who, in this way, attests for their
correctness. For example, an Issuer could be attesting
attributes of the User to the university for the attribute
of students status, to the bar association for the
attribute of being an advocate, and to the trade
register for a company’s commercial status.
The User is the service user who obtains her
credentials from the Issuer through th issuance
protocol. These credentials enable her to provide
proof of the required identity attributes towards a
Verifier (e.g. an eCommerce site).
The Verifier receives a presentation token from
the User which process that the User possesses certain
identity attributes. The Verifier (e.g. an eCommerce
site) provides some kind of service to the User for
whom it requires to know certain identity elements
which, in turn, requires the User to either reveal or,
simply, prove possession of these elements.
The Inspector, which is an optional entity, reveals
the identity or other encrypted attributes of a User,
thereby lifting her anonymity, upon a legal request.
Legality is judged against previously defined criteria
(e.g. fraud, service misuse etc.), clearly stated to the
users when they first obtain credentials from the
Issuer. Due to privacy subtleties involved in the
Inspector entity, an Inspector building block should
not be, by default, a part of a Privacy-ABCs platform
but it should be included only after thorough
consideration and if it is absolutely necessary. The
issue of the Inspector has attracted many discussions
which are beyond the scope of this paper since they
involve, mostly, legal issues and depends on the
specific mandates of personal data protection
legislation (please see Bieker, et al. 2015 for a
thorough discussion on these issues).
The Revocation entity is responsible for revoking
issued User credentials. If revoked, the credentials
cannot be used for producing valid proofs about
credentials (i.e. presentation tokens). The Revocation
entity is, again, an optional component of a Privacy-
ABCs system. Often, the entity offering the
Revocation service is the same as that offering the
Issuer service, since the Issuer can be assumed to
have the most up-to-date and complete information
about users.
In an actual deployment, some of the above
entities may, actually, be implemented by the same
entity or be split among many different entities. For
example, an Issuer can, at the same time, have the role
of Revocation Authority and/or Inspector, or an
Issuer could later also be the Verifier of tokens
derived from credentials that it issued itself.
In Figure 2 we can see the generic Privacy-ABCs
architecture based on the entities/services described
above. The focus in this description is on generality
rather than attention to the peculiarities of a specific
use case, one of which we are going to examine next
in the context of price discrimination.
Figure 2: A generic Privacy-ABCs architecture.
3 PRIVACY-ABCs AGAINST
PRICE DISCRIMINATION
In this section we explain how Privacy-ABCs can be
exploited in order to protect eConmmerce site users
from price discrimination. To this end, we describe
the roles of the identities that we discussed in the
previous section. All implementations can be based
on the Privacy-ABCs reference implementation
available at Github (please visit
https://github.com/p2abcengine/p2abcengine).
Our first entity, the Issuer, can be a major
eCommerce player that agrees to assume the
responsibility of issuing credentials to the Users. The
Issuer, alternatively, can be a governmental agency
(e.g. Ministry of Commerce) that assumes this role in
order to create an eCommerce environment without
price discrimination for users. The Issuer is
implemented by the modules in the Issuer block in
Increasing Trust Towards eCommerce: - Privacy Enhancing Technologies Against Price Discrimination
27
Figure 2. The credential issued to users includes basic
information (attributes) about a user’s identity and it
looks like the following:
We can see that this is the usual information found
in a customary eIdentity card or certificate (e.g.
X.509) plus, perhaps, some other information useful
in an eCommerce environment such as, for instance,
whether the user has a credit card or not (last attribute
in the credential figure).
The major differences from the customary
identification schemes lie in the following two
features:
The user is in power to reveal only the
credential attributes that she desires to reveal
(or agrees to reveal according to the
eCommerce site policy).
The user can prove a property about an
attribute, without revealing it. For instance,
the user can prove that her age is above 18
without revealing its exact value.
Verifiers, i.e. eCommerce sites, who wish to
operate in an eCommerce environment against price
discrimination, should develop their eCommerce
platform based on the Privacy-ABCs reference
implementation. The Verifier contains the modules
shown in the Verifier block in Figure 2.
Users should install the modules shown in the
User block in Figure 2 on the devices through which
they visit eCommerce sites. The installation is very
easy and convenient while it induces minimal
computation overhead on the device (see Benenson,
et al. 2013 and Benenson, et al. 2014 for more on the
user satisfaction survey we conducted within the
context of the ABC4Trust project).
We consider the Revocation and Inspector entities
unnecessary for an eCommerce environment against
price discrimination and, thus, we do not include a
role for them.
4 AN EXPERIMENTAL
ECOMMERCE SITE BASED ON
PRIVACY-ABCs
The ABC4Trust project has made the Privacy-ABCs
code and support modules available through the
Github repository. They are freely offered for
prospective developers of any application that
requires privacy preserving authentication methods.
The application we present in this section was
built, relatively easily, using the Githtub Privacy-
ABCs modules. It is a prototype eCommerce
application, named Grails, for selling online a variety
of products. It is based on the Grails Hotel
Reservation demo application offered by the
ABC4Trust project in Github in order to demonstrate
the easiness with which one can build a complete
eCommerce application based on the modules
available from the project.
The application uses the privacy features of
Privacy-ABCs in order to enhance users’ trust that
they are not subjected to price discrimination while
they are making their purchases. The additional
feature the is not often found in usual eCommerce
sites and which is based on the properties of Privacy-
ABCs is that it allows duty free shopping to byers
who can prove that they have a valid passport issues
in another country. The passport owner simply proves
the possession of the passport as well as the country
of issuance, keeping all other personal information
secret.
The system is composed of the following
subsystems and modules:
A Grails application carrying the webpage of
the online vendor, acting as the Verifier of
buyers’ passport and credit card credentials.
Throughout the verification process, the
buyer’s anonymity is preserved due to the
properties of Privacy-ABCs.
A Grails application with the webpage of a
bank which issues credit card credentials to
users, acting as an Issuer.
A Grails application with the webpage of a
governmental agency which issues passport
credentials, acting as a second Issuer.
A Revocation service for revoking credit card
credentials due to credit card theft or
expiration.
A User service for managing the user’s
credentials, the communication with the user’s
smartcard which contains the credentials and
performing the presentations of credentials.
credID (Revocable):
User identification number:
First Name:
Last Name:
Address:
City:
Country:
Place of birth:
Birth date:
Profession:
Phone number
:
WEBIST 2016 - 12th International Conference on Web Information Systems and Technologies
28
A user UI (User Interface) service, providing
a GUI (Graphical User Interface) for the User
service.
After booting the virtual machine containing all
the eCommerce application modules described above
(in a simulated eCommerce environment) one simply
logs on with the password “abc4trust”. All the
necessary system services for for revocation
authority, vendor verifier, passport issuer, credit card
issuer, user service and UI service, are activated
automatically.
We can see the available credential issuance
controllers below as they appear on the user’s screen:
The objective of the application is to allow users
to buy a product from an eCommerce site without
subjecting them to price discrimination. The site does
not require from users to show their names or other
details such as country of origin or place of residence,
which could, possibly, allow the eCommerce site to
infer some information that could lead to price
discrimination against the user. The users only have
to prove that they possess a valid credit card and a
passport showing the appropriate credentials,
uncovering only the country of password issuance.
A user wishing to use the Privacy-ABCs features
with this eCommerce site should start by obtaining a
passport credential from the Issuer. The Issuer
maintains a database of people eligible for a passport
credential (possibly by contacting a governmental
agency or authority). In order to be able to get a
credential issued, a user has two options. Either you
can use a preexisting credential, or you must create a
new person, attach a passport to this person and
finally attach a credential to the passport.
After creating a new user identity, we create a new
passport and we attach it to the person. Furthermore,
we have the option of creating a credential for the
passport.
Finally, we will enable issuance of credentials for
the newly created passport. As mentioned earlier, it is
also possible to use a pre-existing credential. At this
point, we should write down the values for cpr (ID
number), passport number and issuance key, as they
are the required values for issuance.
Now you are ready for the actual issuance. We
can start the issuance by filling in the fields with the
values from above and clicking “Issue”.
The process of issuing a credit card credential is
similar to issuing a passport credential. We use the
above controllers.
Increasing Trust Towards eCommerce: - Privacy Enhancing Technologies Against Price Discrimination
29
As with the passport issuance, again it is possible
to either create your own user identity or use a
preexisting one if it exists. The process of creating a
new user identity for the credit card credential is
identical as the one described above for the passport.
Having obtained the password and credit card
credentials, the user can then proceed to the
eCommerce site to perform, anonymously and
without being price discriminated, her purchases.
Also, the sample site we created allows the vendor
to add products and edit their characteristics.
The user is able to search using product search criteria
in order to locate the desired product.
Once the product is found, the user submits the
purchase order by clicking on the “Submit” button.
Then an Privacy-ABCs presentation token is send
over to the vendor to certify that the user has a valid
password and valid credit card credential.
The vendor site verifies the token and the
webpage displays the “Verification was successful”
message to the user, signifying acceptance of the
order at the shown price. What is important is that
throughout the purchase process, nothing is known
about the user except that she has a valid passport and
a valid credit card. After the process is complete with
the shown product price, the user can enter a second
interaction protocol with the vendor whereby she
reveals her real identity in order to receive the
goods.
5 CONCLUSIONS
Price discrimination is not a new phenomenon in
commerce. Moreover, it does not necessarily have
negative implications for the consumer or the seller.
In recent years, however, price discrimination
appears to cross the border of consumers’ privacy and
exploit personal information in order to adjust prices
offered to consumers for maximizing profit.
In this paper we argued that a new Privacy
Preserving Technology, Privacy ABCs, can be
exploited towards creating an eCommerce ecosystem
in which price discrimination is not possible due to
the preservation of consumers’ privacy, while they
are still able to prove facts about themselves which
are necessary in order to perform a purchase.
Our point of view is that price discrimination,
although it appears beneficial to eCommerce sites, it
destroys the trust relationship between sellers and
buyers which can lead, in the long run, to financial
losses due to loss of customers or, even worse,
financial losses due to convictions because of leaks of
personal data of the customers.
Of course, there are counterarguments to our point
of view. The first one is that eCommerce players aim
at building close relationship with their customers in
order to serve them best. To this end, they need to
know as much about them as possible in order to
match their needs and, even, reward their loyalty. For
instance, returning customers may be offered some
special products or services best suited for them,
based on their purchase history or other personal
information that has been provided by them. The
other counterargument is that most people bake
online purchases from special applications
downloaded to their mobile devices (either phones or
tablets) which ask of the users to fully identify
themselves in order to proceed to make purchases.
Thus, there appears that giving personal information
is either beneficial to the buyer (first
counterargument) or even mandatory (second
counterargument). With respect to the first argument,
we should stress the fact that we are not against
customer loyalty programs or benefits given to them
based on their personal information and purchase
history. It is the price discrimination in increasing
prices based on such information which we attempt to
reduce using PETs, at least for customers who
perform purchases from a variety of eCommerce sites
and are not much interested in rewards so as to
provide personal information or allow tracking of
WEBIST 2016 - 12th International Conference on Web Information Systems and Technologies
30
their buying behaviour. With respect to the second
argument, the dilemma of the buyer, when faced with
the application’s request to provide identifying
information, is the following: “Should I proceed with
the purchase, giving the requested personal
information or to decline and leave the site?”. We
believe that our approach can offer a compromising
solution for both the eCommerce site and the buyer:
the eCommerce site can ask for information related
to, e.g., the buyer’s preferences, desires etc. or ask of
the buyer to prove something about herself such as,
for instance, that her age is at least 18. Such
information may still be used by the site in order to
provide tailored services to the user, without
breaching the user’s privacy with the risk of the buyer
being subjected to price discrimination.
We believe that the “price or no price
discrimination” question is not easy to settle. Our
paper can at least be a starting point for further
discussions. Technically, PETs can help reduce the
price discrimination phenomenon and boost buyers’
trust towards eCommerce vendors. Of course, it is
also necessary to create a suitable policy framework
in which it is clear what information is necessary for
what types of services and convince all stakeholders
to respect it (vendors and byers alike). The
technology for protecting privacy is here but what
remains is to assess its usefulness and applicability in
the price discrimination phenomenon.
REFERENCES
Acquisti, A.,Varian, H.R., 2005. Conditioning Prices on
Purchase History. Marketing Science. Vol. 24, Issue 3.
Armstrong, M. 2006. Recent Developments in the
Economics of Price Discrimination. In Proc. Advances
in Economics and Econometrics: Theory and
Applications, Ninth World Congress, Cambridge
University Press.
Benenson, Z., Krontiris, I., Liagkou, V., Rannenberg, K.,
Schopf, A., Schröder, D., Stamatiou, Y. C., 2013.
Understanding and Using Anonymous Credentials. In
Proc. 9th Symposium on Usable Privacy and Security
(SOUPS 2013).
Benenson, Z., Girard, A., Krontiris, I., Liagkou, V.
Rannenberg, K. Stamatiou, Y. C., 2014. User
Acceptance of Privacy-ABCs: An Exploratory Study.
In Proc. of the Second International Conference on
Human Aspects of Information Security, Privacy, and
Trust (2
nd
HCI International 2014), pp. 375-386,
Springer-Verlag.
Bieker, F., Hansen, M., Mikkelsen, G.L., Obersteller, H.,
2015. ABC4Trust Workshop on Core Features of
Privacy-ABCs, Practical Use, and Legal Issues.
Privacy and Identity Management for the Future
Internet in the Age of Globalisation, Volume 457 of the
series IFIP Advances in Information and
Communication Technology, pp. 253 – 266, Springer-
Verlag.
Brands, S., 2000. Rethinking Public Key Infrastructures
and Digital Certificates; Building in Privacy. First
Edition,. The MIT Press.
Brands, S., Demuynck, L., Decker, B. De., 2007. A
Practical System for Globally Revoking the Unlinkable
Pseudonyms of Unknown Users. In Proc. 12
th
Australasian Conference on Information Security and
Privacy (ACISP 2007), pp. 400-415, Springer-Verlag.
Camenisch, J., Lysyanskaya, A., 2001. Efficient non-
transferable anonymous multi-show credential system
with optional anonymity revocation. In Proc.
International Conference on the Theory and
Application of Cryptographic Techniques
(EUROCRYPT 2001), LNCS, pp. 93-118. Springer-
Verlag.
Camenisch, J., Lysyanskaya, A., 2004. Signature Schemes
and Anonymous Credentials from Bilinear Maps. In
Proc. 24
th
International Conference on Cryptology
(CRYPTO 2004), pp. 56-72, Springer-Verlag.
Camenisch, J., Groß, T., 2012. Efficient attributes for
anonymous credentials. ACM Transactions on
Information and System Security (TISSEC), Vol. 15,
Issue 1, Article No. 4, ACM Press.
Huberman, B. A., Franklin, M., Hogg, T. 1999. Enhancing
privacy and trust in electronic communities. In Proc. 1
st
ACM Conference on Electronic Commerce, pp. 78–86,
ACM Press.
Mikians, J., Gyarmati, L., Erramilli, V., Laoutaris, N. 2012.
Detecting price and search discrimination on the
internet. In Proc. 11th ACM Workshop on Hot Topics
in Networks, pp. 79 – 84, ACM Press.
Mikians, J., Gyarmati, L., Erramilli, V., Laoutaris, N.,
2013. Crowd-assisted Search for Price Discrimination
in E-Commerce: First Results. In Proc. 9th
International Conference on merging Networking
Experiments and Technologies (CoNEXT 2013). ACM
Press.
Narayanan, A., Shmatikov, V., 2010. Myths and fallacies
of “personally identifiable information”.
Communications of the ACM, 53 (6): 24, ACM Press.
Rannenberg. K., Camenisch, J., Sabouri, A, eds., 2015.
Attribute-Based Credentials for Trust: Identity in the
Information Society. Springer-Verlag.
Sabouri, A., Krontiris, I., Rannenberg. K. (2012). Attribute-
Based Credentials for Trust (ABC4Trust). In Proc. 9th
International Conference, on Trust, Privacy, and
Security in Digital Business (TrustBus). Springer-
Verlag.
Increasing Trust Towards eCommerce: - Privacy Enhancing Technologies Against Price Discrimination
31