Keeping abreast of technology-driven business model evolution: a dynamic patent analysis approach

Abstract

All firms that depend on technology for their competitive positions recognise that keeping abreast of technology-driven business model evolution is of vital importance. However, previous studies cannot offer a concrete way of profiling trends owing to the lack of quantitative data and systematic processes. We propose a dynamic patent analysis that can identify complex relationships among business method patents and visualise trends in technology-driven business model evolution. At the heart of the suggested approach is morphological analysis (MA) for structuring different types of business models at a technological attribute level and modified formal concept analysis (FCA) for investigating technological changes in business models over time. A case study of business method patents concerning electronic shopping is presented to show the feasibility of the proposed approach. We believe our method can promote consensus-building on up-to-date trends in technology-driven business model evolution, serving as a starting point for a more general model.

Keywords:

• technology-driven business model evolution
• dynamic patent analysis
• business method patents
• text mining
• morphological analysis (MA)
• formal concept analysis (FCA)

Acknowledgements

This work is supported by the National Research Foundation (NRF) of Korea Grant funded by the Korean Government (MEST) (NRF-2011-357-D00266).

Notes

The TF-IDF index is a numerical statistic which reflects how important a word is to a document in a corpus. This measure increases proportionally to the number of times a word appears in the document, but is offset by the frequency of the word in the corpus to adjust the importance of words that are more common than others.

The Salton index is derived by dividing the frequency of co-occurrence of two words by frequency of occurrence of each word. A higher value indicates a closer relationship between the pair of words. This measure has an advantage over the Pearson correlation in that the similarity is insensitive to the number of zeros since the cosine value is not based on the mean of the distribution.