The realization of Semantic Web based E-Commerce and its impact on Business, Consumers and the Economy Authors: Univ.-Prof. Dr. Schahram Dustdar dustdar@infosys.tuwien.ac.at Technical University of Vienna Univ.-Prof. Dr. Dieter Fensel dieter.fensel@deri.org Digital Enterprise Research Institute - Galway / Innsbruck / Seoul / Stanford Markus Linder ml@smart-infosys.com Smart Information Systems - Vienna Univ.-Prof. Dr. Heinrich Otruba heinrich.otruba@cec.eu.int European Commission - DG Information Society and Media Brussels Mag. Tassilo Pellegrini t.pellegrini@semantic-web.at Semantic Web School Vienna Bakk. Martin Schliefnig ms@smart-infosys.com Smart Information Systems Vienna This article intends to summarize the outcome of the event Semantic Technologies - The Future of E-Business, which was organized by Wirtschaftskammer Österreich (WKÖ), Oesterreichische Computergesellschaft (OCG) and Austriapro and took place in Vienna on April 26th 2006. The article gives an overview about the state of the art of the topic and an outlook to the near future.
Summary Semantic Web Technologies enable machines to interpret data published in a machine-interpretable form on the web. At the present time, only human beings are able to understand the product information published online. However, semantic web technologies will give digital assistants and agents the ability to search the web for products that correspond best to the specific needs of a certain user. While consumers today have to rely on the limited number of offers available on centralized e-commerce portals when looking for products, future applications will be able to provide users with a search process based on product attributes, which will include all products published in this form on the Internet. Furthermore, in a next step semantic web services will enable digital assistants to handle business processes like selling and buying or even negotiations automatically. The semantic web initiative is being promoted by the World Wide Web Consortium (W3C). Key principles of the Semantic Web include RDF and OWL, which are W3C standardized languages used to express information in a machine interpretable form. Using these languages, manufacturers can publish their product and service descriptions in a standardized form based on product- and service ontologies. An ontology can be seen as an enhanced standardized taxonomy which is used to describe or model aspects of reality. One of the application areas of ontologies is product- and service description made by business companies and published on their web space parallel to their normal HTML Websites. Similarly, it allows vendors to describe their offers in a machine interpretable form by using offer ontologies. Consequently, all semantic engines are able to find, retrieve and interpret products and offer information for their users. As these engines are able to understand this structured information, they are able to compare the products and offers based on specific attributes for the individual user. Today, consumers can make use of centralized e-commerce platforms to compare prices of various dealers for a certain product. Only a few of these platforms allow consumers to compare products of different manufacturers based on product features. Now Semantic Web Technologies provide the opportunity to include all offers which are published in a machine-interpretable form online into a product- Page 2
attribute based search process. This will increase the market transparency and maximise the consumers chances to find the best and the cheapest offer. From the perspective of the vendor companies, semantic web technologies in e- commerce offer huge opportunities while simultaneously exposing companies to certain risks. The new technology will, for example, offer manufacturers and vendors a major opportunity to target niche customers in a more efficient way because digital assistants can evaluate and choose the offers that fit the customer's requirements best. However, the rising transparency will make it increasingly difficult for weak or moderate companies to be successful in the battle of competitiveness. In the past, several technological and organizational innovations leading to a better aggregation of offer and demand on the markets could be witnessed. Market innovations like the introduction of stock exchanges or the World Wide Web have given the market participants a better overview of the current market situation. The last step towards a higher degree of efficiency in the global market place was the introduction of centralized e-commerce platforms like ebay.com, which have became increasingly popular. The next generation market will allow customers to get the complete picture of supply and demand and will enable them to take effectively every single offer on the market under consideration with the help of digital assistants. These technological changes will lay the foundation for a more efficient market which is a key variable towards increased growth and employment. Page 3
Semantic Web Technologies in E-Commerce The popularity of Electronic Commerce is experiencing rapid growth and the importance of the Internet as a sales channel for suppliers and dealers is irresistibly rising. A search for a product or product offers is the starting point for most e- commerce transactions. However, the quality of the current search engines is far from perfect. Search engines, the most popular point of entry into the product purchasing process, deliver links to pages in which the searched phrases appear, but these links are often not helpful from a customer's perspective. As an example, a consumer searching on Google for: " 'mobile phone' camera email download ringtones", will normally end up with a list of links to websites offering ringtones, in addition to some links to websites of mobile phone manufacturers or vendors. However, the search will most certainly not return a list containing all mobile phones with a camera, email function and the ability to download ringtones. On the other hand, if the consumer decides to make use of an e-commerce portal like Ebay to search for a mobile phone, he would end up with a list of thousands of mobile phones. Nevertheless, the search result will not clearly distinguish between the phones that offer the desired features and those of no interest to the consumer. Hence, internet users would be able to find a large number of offers by a large number of dealers by following the above-mentioned approaches, but would probably have to browse many differently structured web pages to find the product providing the needed features and the best possible offer. Today consumers drown in an incredible amount of unstructured information about products and offers. The common search engines and e-commerce portals do not have the ability to facilitate this challenge for the users. Semantic web technologies give digital assistants the ability to search the web for the product that corresponds best to the specific needs of a certain user. Different from a human being a digital assistant is able to analyze thousands or even millions of products and offers by comparing their attributes according to the needs of the user. While consumers today have to rely on the limited number of offers available on centralized e-commerce portals when looking for products, the next generation engines will be able to provide users with a product attribute based search process, Page 4
which will include all products published in this form on the Internet. Furthermore, in a next step, semantic web services will even enable digital assistants to handle business processes like negotiations or placing orders completely automatically. How the system is going to work The semantic web initiative was started by Sir Tim Berners Lee who is the famous inventor of the World Wide Web and today's director of the World Wide Web Consortium W3C, the planet's leading standardisation and development organization for Internet technologies. From the very beginning, the Semantic Web was part of his general vision of the World Wide Web (Berners-Lee, 2000). Since then, researchers and Internet pioneers have been developing the technological groundwork to make this vision come true. The technical principle of the Semantic Web is rather simple. It is based on the consensus that every object or term of the real world is to be referred to by a Unified Resource Identifier (URI). Other key principles of the Semantic Web include RDF and OWL, which are W3C standardized languages used to express information in a machine interpretable form. Manufacturers and vendors of goods and services will be able to publish their product- and service descriptions in this standardized form based on product- and service ontologies. Technically ontology can be seen as an enhanced standardized taxonomy, which is used to describe or model aspects of reality. These ontologies will enable companies to publish product- and service descriptions in a machine interpretable way on their own website, parallel to their normal HTML-Websites. Simultaneously, it allows vendors to describe their offers in a machine interpretable form by using offer ontologies. Consequently, all semantic search engines are able to find, retrieve and interpret those products and offers for their users. For the reasoning part, description logics can be used to describe general and specific rules to conclude new knowledge from the information of these websites. As already mentioned, the general idea behind the Semantic Web is to bring the World Wide Web from its syntactic to a semantic level. This means that it should not only be possible to publish information for humans on websites but also to enrich this Page 5
information with machine interpretable meaning. Today, it is possible to implement many complex internet solutions that help us make everyday work easier. Google makes it possible to find the information we need quite well and centralized e- commerce platforms, like for example Amazon or Ebay, enable us to buy nearly everything we need over the internet. Nevertheless, all these applications operate on a syntactic level. For example, Google does not understand the meaning of the information we are searching for because it is only able to compare some words or sentences. Likewise, Amazon or Ebay have no idea of the meaning of the products they sell. At this point, the potential of the Semantic Web becomes important. When utilising the Next Generation Web, it is possible to describe the meaning of things and information that are published on websites jumping from the syntactic to the semantic level of description. For example, this would enable Google to understand complex search questions and it would therefore not return results you are not searching for. Similarly, Amazon and Ebay would be able to understand the meaning and therefore also the functions of the products, which would enable them to present substitutes as alternatives. One of the biggest advantages of this new approach is the possibility to use the information that is published on the websites in a direct way without the step of converting or inserting the information in relational databases. Therefore, the information is always up to date and web applications are able to work with this information directly. This makes the collecting of and searching for information easier, faster and more powerful and the applications more flexible. To structure information databases are no longer necessary, as a general description of real time objects will provide the end-user with a list of offers directly relating to the desired product or services. These general descriptions of real time objects are called ontologies and describe the properties of the objects and relations to other objects in a way that comes closer to natural language descriptions. A second big advantage of the semantic level is the possibility of easy and flexible reasoning mechanisms created by the use of ontologies, which provide the search assistants with knowledge about the meaning of a described object and its properties and relations. To practice automatic reasoning a knowledge base and an inference system are necessary. In addition, the inference system includes a set of rules to Page 6
produce new knowledge from the knowledge base. It must therefore be able to generate semantically correct propositions by using syntactic algorithms, as artificial intelligence only constitutes a simulation of human actions and reasoning. Furthermore, the more semantic power included in the knowledge base, the stronger the inference system, which in turn will produce new knowledge by reasoning. At the event Semantic Web Technologies The future of E-Business Smart Information Systems presented a test application which put the semantic technology principles previously outlined into practice. The test application was developed for the mobile phone market, and was supported by mobile phone producers like Nokia or Samsung that assisted the demonstration by publishing machine-interpretable RDF files on their website. These files contain a description of the mobile phones currently on offer utilising specific product ontology for mobile phones including information, such as, size, weight, supported frequencies, available storage size and operating system. The RDF files were created with the help of an application, which allows the user to simply specify the attributes of the products in a table. Subsequently, the application translates this information into a machine-interpretable file, which in turn is used as a unique identifier of the respective mobile phone. A files used for the purpose of the demonstration can be found at the following URI: http://www.nokia.at/owl/nokia_6230i.owl. Similarly, the mobile phone dealers published their mobile phone offers in a machine interpretable form on their respective websites parallel to the normal HTML offer. Their offers were described in an offer ontology similar to the ontologies used by the mobile phone producers specifying information such as price and the expected delivery time. Clearly, the offer ontology utilised by the mobile phone dealers is independent of the product areas, as the product itself is described on the web page of the producer, and was created based on a simple subset of the international ebxml standard. In the near future web shops and Enterprise Resource Planning Systems will be able to translate the contained product offers into RDF automatically without any further involvement of the dealers. As demonstrated by the test application, the published RDF files can be retrieved and interpreted by any semantic search engine or digital assistant, - in this case the Page 7
so-called Smart Assistant. The Smart Assistant allows the user to perform an attribute-based search for mobile phones and mobile phone offers published in RDF online. Additionally, the Smart Assistant performs reasoning processes based on the enhanced information base and allows the user to find a phone that fits his or her requirements. How the consumers will be affected by semantics in e-commerce Up until now, consumers have had to utilise centralized e-commerce platforms to compare prices of various dealers for a certain product. However, those platforms can only suggest offers, which have been published there previously by suppliers or dealers. Consequently, offers on such virtual market places will always be limited to a certain percentage of all offers available on the market, as a supplier cannot know all of these platforms, not to mention, offer his products on every single platform. Semantic search engines retrieve their product offers directly from the supplier's web pages and can therefore include all offers published in a machine-interpretable form on the web. Hence, they can potentially provide consumers with the whole picture of the market, as they will not only feature suppliers willing to pay fees to centralized e- commerce platforms, but every single offer published on the web. Only a few e-commerce platforms allow consumers to do a feature based product comparison between different manufacturers and their range of models. Semantic Web technologies, on the other hand, enable search assistants to analyze the attributes of all products on the market and to compare them against each other. Consequently, consumers will be able to state specifically what they want. Therefore, to come back to the example mentioned above, the consumer is now able to specify that he is looking for a mobile phone that supports e-mail and downloading of ring tones, and has a camera. A semantic search engine will only provide the user with product models, which correspond to the expressed needs. Hence, the consumers get the opportunity to consider the maximum amount of possible products allowing them to find the solutions, which fit their individual requirements best. Furthermore, the possibility to perform reasoning on the Semantic Web data, will allow search engines to display the aggregated information in a much more user-friendly way. Accordingly, the strongly increased transparency of the market will maximise the Page 8
chance of finding the best and cheapest offer. In the future, semantic web services will enable search assistants to order products for a user at a dealer s web shop. This means that users do not have to get accustomed to the user interface of yet another platform. They just have to register once on the platform of the semantic search assistant and this assistant can then interact with the dealers websites on behalf of the consumer, and order for example a certain mobile phone. The impact of Semantic Web based E-commerce on businesses Dealers and suppliers only have to publish their product information and offers once on their own website and all search engines will be able to interpret this information. From the vendor's perspective, semantic web technologies in e-commerce have both positive and negative effects. A huge opportunity is that vendors will have a much better and more direct access to the consumers. Currently, many niche markets are too costly to target by means of conventional marketing. Likewise, many small and medium companies fail because they are unable to market and distribute their products successfully. Semantic Web based e-commerce will allow these companies to simply describe their products and their specific attributes on their own web page allowing them to be automatically considered in thousands of consumer based search processes with a maximum chance to be found if the offered product fits the needs of a customer. Consequently, companies are able to target niche customers in a more efficient way, because of tuned customer specifications and the digital assistants being able to evaluate which offers fit those requirements best. Semantic Web Technologies allow users to specify exactly what they want and to find products, which come up to those requirements. Semantic Web Technologies will thereby foster direct contact between suppliers and consumers. This is expected to lead to a further tightening of the value chain. Conversely, the rising transparency will make it increasingly difficult for weak or moderate companies to survive the battle of competitiveness. Page 9
Organisational factors driving the adoption of semantic technologies From an organisational point of view, the development and adoption of semantic technologies can be deduced from three major trends: 1) Due to increased competition the demand for dynamic product and services bundles is rising. 2) The acceleration of business transactions favours temporal overlapping of organisational structures based on ad hoc teams. 3) Customers and users are increasingly integrating and compiling their relevant data and information in self-service on behalf of smart dialogue systems and search engines, which themselves are a technological precondition for the realisation of the above-mentioned trends. Self-service, dynamic product bundles and ad hoc teams are strongly intertwined phenomena, which entail each other. As a result of the overall acceleration and dynamisation of business processes, we can observe an increasing degree in organisational complexity. More than in the past, workers are mutually embedded within multiple contexts and roles, which are themselves characterised by a high functional interdependence. Semantic technologies support the user to cope with these varying contexts by contextualising the users' tasks within the overall organisational framework. Hence, semantic technologies are helping to reduce organisational complexity and at the same time open up opportunities for new products and business models. Users are being actively supported in information retrieval, management and compilation. Powerful analytical tools based on semantic technologies enable producers and manufacturers to make use of their vast amount of transaction-generated data for the purpose of demand tracking, personalization, sales and marketing. Furthermore, the various possibilities of automation through wrapper technologies and semantic web services promise to increase the speed and quality of information retrieval processes dramatically by simultaneously lowering costs herein. Page 10
Nevertheless, beside these promising opportunities, which are already being capitalised in diverse fields of application and branches, it is necessary to keep in mind that semantic technologies are powerful instruments in helping to leverage the knowledge society with all its positive and negative results. Thus, semantic technologies have to be applied in a non-trivial and balanced manner, especially when it comes to issues like privacy, property rights and freedom of speech. Macroeconomic impacts on the economy due to improved aggregation of offer and demand In the past, several technological or organizational innovations leading to a better aggregation of offer and demand in the market place could be witnessed. Market innovations like the introduction of stock exchanges or the World Wide Web gave the market participants a better overview of the current market situation. The most recent step toward a higher degree of efficiency in the global market place was the introduction of centralized e-commerce platforms like ebay.com, which have become increasingly popular. The next generation market based on semantic web technologies will allow customers to capture the complete picture of the relevant market. In many areas, the importance of advertising is expected to diminish, as companies will be better able to communicate their strengths over the structured description of the product features. In particular, this is of high importance for small and medium companies as these market participants cannot rely on heavy advertisement due to restricted financial capabilities. The improved situation for niche-suppliers is expected to enable them to make a higher gross-margin as customers are willing to pay a higher price for a product, which better fits their needs. Simultaneously, the stronger market segmentation is expected to lead to a higher level of employment in these companies as economies of scale will be achieved to a lower extent. Consumers will be able to quickly find the products, which fit their specific needs. This will therefore lead to a reduction of mispurchases, and as a consequence, to a Page 11
higher disposable income for more relevant investments. Important effects are also expected with regards to the export business, as exporters frequently have difficulties communicating their products to potential customers abroad. Semantic Web technologies in e-commerce have the potential to lead to a significant shift from offer-to demand-driven markets. A more efficient market is a key variable towards growth and employment in the European Union. The success of the implementation of these new models of making business on the WWW crucially depends on several other factors, including broadband penetration and usage, as well as on net neutrality. Both factors mentioned are related to the organisation of electronic communication networks and services markets. The creation of functioning markets for certain commodities and services using more advanced technologies, such as the Semantic Web, will be supported by a very broad uptake of broadband access on fixed or wireless network platforms and free access to these markets. One could state this in a more provocative way by saying that successful e-commerce needs competitive electronic communication network and services markets. Competition in information technology supported end-user markets can only be successful, i.e. deliver results for the market parties, if a sufficient percentage of the population has access to these markets and can do business on these platforms. This clearly indicates the crucial role of broadband access and net neutrality. Regulation can significantly support the development of e- commerce by providing a competitive environment for markets for electronic communication services and networks. This includes creating and maintaining competition on access markets as well as preserving the ability of consumers to use all resources available on the WWW without interference by access providers, who sometimes seem to prefer to act as gatekeepers to certain resources. Outlook The technological groundwork for a broad realization of Semantic Web Technologies in E-Commerce has been implemented. In the near future two big questions will have to be solved to facilitate a rapid expansion of this technology in innovation driven markets. Page 12
The first challenge to solve is the process of creating standardized ontologies for the different product areas. One of the in the short-term most promising approaches is the transformation of existing international standards like eclass, EDIFACT, ebxml into Ontologies. (Hepp 2006/1-2) These standards allow the description of properties of products and services. Nevertheless, experiences show that the process of defining industry standards for product areas in international standardization committees is too slow and not flexible enough to keep pace with the developments. Therefore, the need to develop new approaches should be encouraged by allowing real time online development of such standards by companies, experts and interested consumers. At the present time, a research project is planned tackling this challenge. A second problem is that anyone can publish information about anything in the semantic web. Mechanisms have to be put into practice guaranteeing authenticity of authors. Here organizations like chambers of commerce play a crucial role as they could offer directories where their members can publish links to their trustworthy machine interpretable information. Page 13
Bibliography Baader, Franz et al. (2003): The Description Logic Handbook. Cambridge University Press Berners-Lee, Tim. (2000): Weaving the Web: The Original Design and Ultimate Destiny of the World Wide Web by Its Inventor. New York: HarperCollins Berners-Lee, Tim, James Hendler and Ora Lassila (2001): The Semantic Web A new form of Web content that is meaningful to computers will unleash a revolution of new possibilities. In: Scientific American, May 2001 issue Fensel, Dieter et al. (2003): Spinning the Semantic Web Cambridge, etc.: The MIT Press Golbeck, Jennifer (2003): Personalizing Applications through Integration of Inferred Trust Values in Semantic Web-Based Social Networks. University of Maryland, College Park Hepp, Martin (2003): Güterklassifikation als semantisches Standardisierungsproblem. Deutscher Universitäts-Verlag, Wiesbaden 2003. [Zugl. Diss., Universität Würzburg, 2003] Hepp, Martin (2006/1): Products and Services Ontologies: A Methodology for Deriving OWL Ontologies from Industrial Categorization Standards, Int'l Journal on Semantic Web & Information Systems (IJSWIS), Vol. 2, No. 1, pp. 72-99, January-March 2006. Hepp, Martin (2006/2): The True Complexity of Product Representation in the Semantic Web Accepted for the 14th European Conference on Information System (ECIS 2006), June 12-14, 2006, Gothenburg, Sweden. Pellegrini, Tassilo; Blumauer, Andreas (2006). Semantic Web. Wege zur vernetzten Wissensgesellschaft. Berlin: Springer Verlag Schnurr, Hans-Peter and Andreas Nierlich (2005): Semantische Suche. Scharer-Druck, Karlsruhe W3C Semantic Web http://www.w3.org/2001/sw/ Page 14