Prof. Dr. Heiko Beier
Diversity sells: How semantic meta data can revive sales within classical book trading infrastructures
With the rise of e-Commerce platforms like Amazon, iTunes and others book trading has changed completely. Convenience of online book shopping by far outrules that from traditional shopping in book shops, with spending much time for searching books and lining up at the cash desk in particular in peak seasons.
But it is more than traditional backwards thinking to state: the process of book buying has lost much of its charme. Current sales-supporting technologies like search and recommendations based on click behavior add a strong bias on few, most dominant books. "The winner takes it all" paradigm is heavily subsidized by online book selling platforms. It spoiles diversity, richness of content and finally revenue.
This presentation makes a proposal to revive classical book trading structures, where experienced staff in local book stores help people find not only what they have directly been looking for, but assist in personally finding something exciting to read or getting the right book for a friend with a certain interest or cultural background.
Current practices in online book shopping like "looking into the book" are just a short-sighted one-to-one transfer from analog to digital experience. They offer no equivalent to the experience of "in-store" book browsing and by far do not exploit the power of digital technologies.
Meta-data play a crucial role for users to be able to deep-dive along various story-lines, quickly exploring the richness of hundreds of thousands of books in a manner unparalleled in real world book stores. This sets various requirements for meta data standards including
- extending existing meta data schemes like ONIX in book trading by rich semantic meta data
- Integration of standards like SKOS to represent book traders' knowledge about content and storylines
- standards for mapping semantic concepts between different subject domains and expressing intercultural differences
- standards for editorial campaign-driven control of search and recommendation