PublicationsPrevious | item14 | NextFrom: Dave Jones dave@mineit.com Subject: Web Mining and Privacy Date: Fri, 7 Apr 2000 16:28:36 +0100 Web Mining For Personalization, Without Privacy Problems This article is a response to a KDNuggets poll that asked the question "Are you concerned about web mining compromising your privacy?" (see www.kdnuggets.com/polls/2000/privacy-2000-03-30.htm)
PrivacyEvery person on this earth has rights - freedom of speech and movement, the right to breathe fresh air every day and also the right to their own personal privacy. The rise of the Internet and its associated technologies (including the ubiquitous cookie) has sparked extensive debate on aspects of personal privacy on the web.These technologies allow companies to analyse movements users make on the web and to adapt their sites for each individual visitor. This may give rise to an image of a `Big Brother' being able to trace exactly what you as a user are doing when you are online, knowing exactly what you have looked at and where. Currently there is some confusion, apprehension and a general lack of awareness about web mining and how it relates to privacy- this article will attempt to explain them.
Web MiningWeb mining is the application of data mining technologies to log files and associated data from a particular web site in order to discover knowledge on browser and buyer behaviour on that site. Web mining allows companies to understand how their site is used and by whom, in order to adapt the site to better serve the visitors needs and ultimately generate a higher return on investment from the site.However, more importantly from a users' perspective, Web mining also enhances the privacy of the individual as all analysis is done on web log data that is largely anonymous. Indeed, Berry and Linoff remark that "identifying the consumer on the web is perhaps harder than identifying consumers in a material world." This is because of the potential anonymity of the visitor in the data that is available.
Web Privacy MechanismsThe web has various mechanisms designed to enhance the individuals right to privacy.Most web browsers have the ability to recognize when a cookie is being sent to it and react accordingly. The individual has the choice as to whether to accept the cookie or not. In Germany, it is illegal for a vendor to sell (or make available for download) a browser with `cookie checking' turned off. Individuals hold the right to remain anonymous unless they explicitly decide otherwise. That is, whilst web mining can usually identify a visitor as coming from a particular IP address or having a certain cookie id, they remain anonymous until the user provides some form of personal information, such as an email address. Even then web mining will only recognize the visitor by their "virtual presence" not by any physical means. Recent press has cited several examples of alleged abuse of personal privacy. However, if a company is seen to abuse its customers' data and, by implication, its customer's rights then it may lose its customers. In many ways, the individuals hold all the aces. Unless they want to start trading them - but would visitors trade personal information? It would seem so. In a recent survey by the Personalization Consortium, 81% of the respondents said that, in exchange for a personalized experience online, they would provide their address, 95% their email address and 96% their name.
Group AnalysisMost opponents of web mining seem to be concerned with the idea that it is focusing on what individuals are doing. This can be incorrect. Much more information can be gleaned from looking at group behaviour than that of the individual. It may often be more insightful to know that 80% of your visitors to the Bordeaux red wine pages came from 10% of your banner ads than to know that John Smith bought a bottle of 1983 Chardonnay.Once you have identified the groups that visit your site, you must then put individual visitors into a group in order to allow you to personalize their browsing experience on the site. Indeed, Locke writes that personalization can be used "in such a way that they benefit from each other's accumulated experience, knowledge, interests, inclinations and tastes. When this approach is taken," Locke continues, "privacy rarely becomes a concern."
ConclusionFrom one point of view it can be seen that any analysis of visitor data is a breach of personal privacy, but how realistic is that when you consider how many ways the individual can shield their identity online: a user can refuse to give information, move to a different site, even not go online! From another point of view, the benefits of being recognized online far outweigh the negative aspects: personalized service, relevant information and offers, a one to one relationship with the company.Also, consider the methods by which web mining utilizes its findings - namely personalization. Alternative methods of using online data are based around the collection of email addresses, which are then used by companies to bombard people with the companies' latest product offerings. The pervasive versus invasive methods are proving to be much more successful. So if a company provides a personalized site that delivers exactly what its visitors want, then it is seen to be a site that delivers and meets visitors' needs. But they must understand what the visitors want? Bruce Kasanoff claims that "Enlightened companies remember information *for* customers, not *about* them." Is it so wrong to use web mining to understand the needs of the visitors and to adapt the site accordingly?? Ask the happy visitors and see what they say! However, in reality most companies actually do NOTHING with the data they have! There will always be people unwilling to accept the benefits that web mining can bring, because it is a new technology and new technologies always carry certain negative energy. However, if embraced, Web Mining can deliver sites that reflect the interests and needs of the individual visitor - and all of this can be done without them losing their privacy. David Jones VP, Business Development MINEit Software Email: Dave@MINEit.com Web: http://www.mineit.com Previous | item14 | Next |
Copyright © 2000 KDnuggets. Subscribe to KDnuggets News!