The Linked Data Value Chain:
A Lightweight Model for Business Engineers
Atif Latif, Anwar Us Saeed
(Graz University of Technology, Austria
{atif.latif,anwar.ussaeed}@student.tugraz.at)
Patrick Hoefler
(Know-Center Graz, Austria
phoefler@know-center.at)
Alexander Stocker, Claudia Wagner
(Joanneum Research, Graz, Austria
{alexander.stocker,claudia.wagner}@joanneum.at)
Abstract: Linked Data is as essential for the Semantic Web as hypertext has been
for the Web. For this reason, the W3C community project Linking Open Data has
been facilitating the transformation of publicly available, open data into Linked Data
since 2007. As of 2009, the vast majority of Linked Data is still generated by research
communities and institutions. For a successful corporate uptake, we deem it important
to have a strong conceptual groundwork, providing the foundation for the development
of business cases revolving around the adoption of Linked Data. We therefore present
the Linked Data Value Chain, a model that conceptualizes the current Linked Data
sphere. The Linked Data Value Chain helps to identify and categorize potential pitfalls
which have to be considered by business engineers. We demonstrate this process within
a concrete case study involving the BBC.
Key Words: Linked Data, Linking Open Data, Value Chain, Business Case, Business
Models
Category: H.m, L.1.4, M.0, M.4
1 Introduction
For several years now, the Semantic Web [Berners-Lee et al. 2001] has been of
great interest to the international research community. As a subtopic, the concept
of Linked Data has gained much attention in the recent months.
Linked Data is based on four simple rules [Berners-Lee 2006]:
1. Use URIs as names for things
2. Use HTTP URIs so that people (and machines) can look up those names
(see also [Sauermann et al. 2008])
3. When someone looks up a URI, provide useful information
4. Include links to other URIs so that they can discover more things
Proceedings of I-KNOW ’09 and I-SEMANTICS ’09
2-4 September 2009, Graz, Austria
568

To seed the Semantic Web with Linked Data and to promote its adop-
tion, the W3C community project Linking Open Data1 was founded in 2007
[Bizer et al. 2007]. The project helps to solve the causality dilemma (chicken-
egg problem) between Semantic Web content and Semantic Web applications
by providing RDF2 data sets from existing open data repositories. To enable
intelligent applications that generate a valuable output for the end user, a crit-
ical amount of high-quality interlinked datasets across different domains is a
crucial precondition, as shown by [Jaffri et al. 2008] and [Raimond et al. 2008].
The vision of the scientific Linked Data community can therefore be described
as follows: First, facilitate the generation of semantically enriched Linked Data,
and as a result, semantic applications will be built on top of this data.
Linked Data incorporates a lot of potential for enterprises [Servant 2008].
However, there is a significant difference between the aims of a scientific com-
munity and the demands and requirements of enterprises, such as revenue flow
and generated value. Furthermore, every successful commercial adoption requires
the discussion of inherent technical, social and business risks connected to the
Semantic Web and Linked Data.
We propose that limited commercial Semantic Web adoption is, among other
reasons, caused by the lack of conceptual work supporting the development of
business cases and the identification of associated risks. Our publication is moti-
vated by these factors and intends to start a discussion which moves the Semantic
Web and Linked Data closer to businesses.
In section 2 we present the Linked Data Value Chain, a model of the Linked
Data life cycle along with participating entities and involved roles and types
of data. In section 3, we apply the Linked Data Value Chain to an existing
business case from the BBC and use the aforementioned model to highlight
potential pitfalls. We conclude our results and present an outlook to potential
future research in section 4.
2 The Linked Data Value Chain
As a prerequisite for the development of successful business cases in the emerging
context of Linked Data, three concepts have to be introduced first: Participat-
ing Entities, their assigned Linked Data Roles and processed Types of Data, as
depicted in Fig. 1.
Our contribution tries to support business engineers with the process of as-
signing Linked Data Roles to Entities, modelling interactions and responsibilities
of Linked Data Roles, and transforming data from Raw Data to Linked Data
and Human-Readable Data, thereby increasing its value along the way.
1http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/
LinkingOpenData
2http://www.w3.org/RDF/
A. Latif, P. Hoefler, A. Stocker, A. Us Saeed, C. ...
569

Figure 1: The Linked Data Value Chain
We propose the Linked Data Value Chain as a lightweight model, which
builds upon the concepts of Linked Data and Value Chains and makes the in-
terdependencies of entities, roles and different types of data – as the output of
the value creation process – explicit.
The value chain as introduced by [Porter 1985] is a concept from the business
domain: In a nutshell, a value chain is a chain of activities producing outputs,
each activity increasing the value of its particular output, finally shaping a highly
valuable end product. In the case of Linked Data with respect to business cases,
Human-Readable Data is the most valuable output for the targeted End User.
2.1 Participating Entities & Linked Data Roles
In the context of Linked Data, participating entities – both corporate and non-
corporate, e.g. persons, enterprises, associations, and research institutes – can
occupy one ore more of the following roles:
– A Raw Data Provider is a role that provides any kind of data in any non-
RDF format.
– A Linked Data Provider is a role that provides any kind of data in a machine-
readable Linked Data format. Such data is currently provided through deref-
erenceable URIs, a SPARQL endpoint or an RDF dump.
A. Latif, P. Hoefler, A. Stocker, A. Us Saeed, C. ...
570

– A Linked Data Application Provider is a role that processes Linked Data
within an application and generates human-readable output for human end
users.
– An End User is a human, consuming a human-readable presentation of
Linked Data. He or she does not directly get in touch with Linked Data,
and typically does not even want to.
The Linked Data Value Chain allows a flexible assignment of roles to entities:
In most cases, one entity just occupies one role, but it may – in extreme cases –
also occupy all roles at once. For example, one enterprise could own the role of a
Data Provider, a Linked Data Provider, and a Linked Data Application Provider
all at the same time. The Linked Data Value Chain also supports multiple sources
of data: A Linked Data Provider may acquire Raw Data from more than one
Raw Data Provider simultaneously, and will usually provide Linked Data to
more than one Linked Data Application Provider.
2.2 Types of Data
Three different types of data can be identified within the Linked Data Value
Chain:
– Raw Data is any kind of data (structured or unstructured) that has not
yet been converted into Linked Data. Such data usually has some structure,
but generally less structure than Linked Data, and is in most cases also not
universally identifiable.
– Linked Data is data in a RDF format that uses dereferenceable HTTP URIs
to identify resources and is linked with other RDF data. This data can be
generated by the Linked Data Provider itself, or data provided by a Raw
Data Provider can be ”RDFized”. Linked Data is intended to be consumed
and processed by machines only.
– Human-Readable Data is any kind of data which is intended, arranged and
formatted for consumption by humans. Consuming this data generates value
for the human end user, which is crucial to the success of any Linked Data
business case.
2.3 Interaction between Roles and Data
Entities can act in different roles. These roles are closely connected through
three types of data, which they provide and/or consume: A Raw Data Provider
provides Raw Data as the input for a Linked Data Provider, who turns it into
Linked Data, increasing its value by semantically enriching it. Linked Data in
A. Latif, P. Hoefler, A. Stocker, A. Us Saeed, C. ...
571

turn serves as the basis for a Linked Data Application Provider, who generates
Human-Readable Data as the most valuable output for a human End User.
Each combination of roles and entities as well as every transformation step
of data holds inherent risks, some of which will be presented in the next section
along with a concrete case study. Knowledge about such risks is crucial for the
development of successful business cases. If not considered properly, these risks
may become pitfalls to business success.
We identified two main areas where pitfalls may arise, grouping them into
Role-Related Pitfalls and Data-Related Pitfalls. In a nutshell, Role-Related Pit-
falls are either related to individual roles or to the interaction of different roles.
Data-Related Pitfalls are either related to the data itself or the data transfor-
mation process. We will explain selected pitfalls emerging in the following BBC
case study.
3 Applying the Linked Data Value Chain
3.1 BBC Case Study
As the idea of Linked Data is still young, there are not many appealing Web
interfaces for human end users yet. One enterprise that is on the cutting edge,
both in regard to deployed Semantic Web technologies and the end-user inter-
face, is the BBC3. Furthermore, the BBC is a pioneer when it comes to adopting
Linked Data within a business case. Their system utilizes Linked Data tech-
nologies to interconnect distributed micro-sites within the BBC network, e.g.
BBC News4 and BBC Music5, and reuses external data from DBpedia and Mu-
sicBrainz [Kobilarov et al. 2009]. By doing so, the BBC generates additional
value for the human end users, while allowing them to immediately consume
Linked Data Roles Participating Entities
Raw Data Provider
BBC
Wikipedia
Linked Data Provider
BBC
MusicBrainz
DBpedia
Linked Data Application Provider BBC
Table 1: Linked Data Roles and Participating Entities in BBC case study
3http://www.bbc.co.uk/
4http://news.bbc.co.uk/
5http://www.bbc.co.uk/music/
A. Latif, P. Hoefler, A. Stocker, A. Us Saeed, C. ...
572

contextually and semantically relevant content from third party sites as well as
interconnected BBC sites.
We apply the Linked Data Value Chain to examine role assignments along
with their interactions as well as the data transformation processes within the
BBC business case. As summarized in Table 1, BBC acts as Raw Data Provider
and Linked Data Provider for their own data as well as Linked Data Applica-
tion Provider for all data, including external data from Wikipedia (Raw Data
Provider) via DBpedia (Linked Data Provider) and from MusicBrainz (Linked
Data Provider).
3.2 Discussion of Potential Pitfalls
In the BBC case study, BBC micro-sites utilize data from DBpedia as an impor-
tant input for the business case. Unfortunately, transforming Raw Data (from
Wikipedia) to Linked Data (via DBpedia) is a very time-consuming and, at
best, semi-automated effort, which is currently undertaken by a team of re-
searchers [Auer et al. 2007]. Linked Data generated this way is therefore hardly
ever complete, correct and up-to-data [Jaffri et al. 2008]. There are no service
level agreements or similar contracts between BBC and DBpedia or DBpedia
and Wikipedia securing all these issues. If not considered well, such performance
and data quality risks may become pitfalls.
Second, BBC end users may want to edit content on BBC Websites which
is provided by DBpedia / Wikipedia. Unfortunately, BBC does not provide an
automated feedback loop leading back to the Linked or Raw Data Providers, in
this case DBpedia and Wikipedia. Such feedback loops are neither implemented
nor, most of the time, conceptualized yet. Currently, users are requested by BBC
to directly edit the respective articles in Wikipedia. Unfortunately, a synchro-
nization of data between Wikipedia and DBpedia will take a very long time,
depending on Wikipedia’s data dumping and DBpedia’s transformation inter-
vals, possibly annoying a human end user, who certainly is not interested in such
technological issues, but wants to see his or her changes promptly, if not in real
time.
Third, BBC provides related links to third party sites. Such a procedure is a
pitfall to successful commercialization, because users may leave the site of the
Linked Application Provider (BBC). Reusing data should be based on widgets
and embedded content, thereby making users stay on the site longer, but still
having the benefits from consuming third party contents.
Fourth, users need information about the provenance of Linked Data in or-
der to be able to assess whether the displayed data comes from a trustworthy
provider. Therefore, Linked Data Application Providers should state clearly from
which Linked Data Providers and Raw Data Providers they present data. Com-
pleteness, correctness and actuality of data strongly depends on the involved
A. Latif, P. Hoefler, A. Stocker, A. Us Saeed, C. ...
573

Raw Data Provider, the underlying algorithms and techniques which convert
Raw Data into Linked Data as well as on the quality of service provided by
Linked Data Provider.
Fifth, the BBC is currently integrating content from only a few selected
Linked Data Providers (DBpedia, MusicBrainz), thereby not utilizing the full
potential of the Linked Data sphere. On the other hand, the BBC retains firmer
control over the displayed content and also avoids problems that may occur
when displaying large amounts of semantic data from different data sources in
a human-readable way [Heath 2008].
Sixth, the BBC has chosen the data which they retrieve from Linked Data
Providers in respect to its provenance. BBC Music takes their introductory texts
for bands from Wikipedia, whose content is inherently unstructered. Discogra-
phies and related links are taken from MusicBrainz, which is a well-structured
music database by design. By not yet using e.g. content extracted from Wikipedia
info boxes, BBC avoids the problems related to the transformation process of
less-structered data into Linked Data.
4 Conclusion and Outlook
Our contribution is dedicated to facilitate the commercial uptake of the Seman-
tic Web vision. We have presented the Linked Data Value Chain as a lightweight
model for business engineers to support the conceptualization of successful busi-
ness cases. Thereby, we identified three main concepts: Different Entities acting
in different Roles, both consuming and providing different Types of Data.
We demonstrated that the assignment of roles to entities, the combination
and involvement of roles, the data selected as well as the data transformation
process hold inherent business risks. We exemplarily applied the Linked Data
Value Chain within a concrete case study from BBC to showcase selected busi-
ness risks.
Our future research will deal with a more detailed classification scheme for
pitfalls arising when businesses strive to adopt the Semantic Web, and we will
elaborate our concept of the Linked Data Value Chain.
Acknowledgement
This contribution is partly funded by the Know-Center and the Higher Education
Commission of Pakistan.
The Know-Center is funded within the Austrian COMET program – Com-
petence Centers for Excellent Technologies – under the auspices of the Austrian
Federal Ministry of Transport, Innovation and Technology, the Austrian Federal
Ministry of Economy, Family and Youth and the State of Styria. COMET is
managed by the Austrian Research Promotion Agency FFG.
A. Latif, P. Hoefler, A. Stocker, A. Us Saeed, C. ...
574