Extracting Market Feedback from Usenet Traffic
Traditional sources of market information are limited in various ways. No single instrument is able to combine all the desired properties of any methodology for gathering customer information. Instead, several instruments have to be combined in a system which is carefully tailored to the specific management or marketing problem. This paper illustrates one such system.The fast paced electronics industry requires acceptance and adoption of immature product offerings. The internet enables a culture where professionals help other professionals through the use of electronic media in order to overcome the problems and pitfalls associated with immature products. Utilizing a real-world example, data from a usenet group will be analyzed to answer the following questions:
- can marketing data be derived by scanning usenet traffic?
- what is the relevance of the derived marketing data?
- what are possible applications for the derived marketing data?
The usenet traffic is analyzed by matching the text of the contributions against dictionaries representing individual vendors. Several approaches are taken to remove bias by performing a multi-dimensional analysis.
The results show that with surprisingly simple methods, ad-hoc data can be obtained which very closely match published marketing data. Possible applications include the following:
- observation of newsgroups to derive marketing data, directly or indirectly
- interaction with newsgroups, e.g. to test the response to announcements
- influence newsgroups to leverage the effect of opinion leaders
In summary, this thesis explores all aspects of a business problem: outlining a conceptual solution, providing a first implementation, validating the results, and identifying future improvements. As such, this thesis provides a practical approach while still maintaining a link to academic thinking.