From: cvs@openprivacy.orgCVS update: openprivacy/htdocs/notes
Date: Tuesday February 27, 19101 @ 22:22
Author: fen
CVSWEB Options: -------------------
Main CVSWeb: http://openprivacy.org/cgi-bin/cvsweb/cvsweb.cgi
View this module: http://openprivacy.org/cgi-bin/cvsweb/cvsweb.cgi/openprivacy/htdocs/notes
-----------------------------------
Update of /usr/local/cvs/public/openprivacy/htdocs/notes
In directory giga:/home/fen/projects/openprivacy/htdocs/notes
Modified Files:
whitepaper.shtml
Log Message:
removed background section - placed on web site
*****************************************************************
File: openprivacy/htdocs/notes/whitepaper.shtml
CVSWEB Options: -------------------
CVSWeb: Annotate this file: http://openprivacy.org/cgi-bin/cvsweb/cvsweb.cgi/openprivacy/htdocs/notes/whitepaper.shtml?annotate=1.28
CVSWeb: View this file: http://openprivacy.org/cgi-bin/cvsweb/cvsweb.cgi/openprivacy/htdocs/notes/whitepaper.shtml?rev=1.28&content-type=text/x-cvsweb-markup
CVSWeb: Diff to previous version: http://openprivacy.org/cgi-bin/cvsweb/cvsweb.cgi/openprivacy/htdocs/notes/whitepaper.shtml.diff?r1=1.28&r2=1.27
-----------------------------------
Index: openprivacy/htdocs/notes/whitepaper.shtml
This archive was generated by hypermail 2b30
: Tue Feb 27 2001 - 22:22:11 PST
diff -u openprivacy/htdocs/notes/whitepaper.shtml:1.27 openprivacy/htdocs/notes/whitepaper.shtml:1.28
--- openprivacy/htdocs/notes/whitepaper.shtml:1.27 Tue Feb 27 22:04:13 2001
+++ openprivacy/htdocs/notes/whitepaper.shtml Tue Feb 27 22:22:10 2001
@@ -9,7 +9,7 @@
</head>
<body bgcolor="#ffffff">
- <!-- $Id: whitepaper.shtml,v 1.27 2001/02/28 06:04:13 fen Exp $ -->
+ <!-- $Id: whitepaper.shtml,v 1.28 2001/02/28 06:22:10 fen Exp $ -->
<h1>OpenPrivacy - Building a Better Internet</h1>
@@ -362,168 +362,6 @@
strike a deal with B to provide her with the editorial filtering
process, saving A time and aiding B at least in reputation if not
also financially.
- </p>
- </blockquote>
- <h2><a name="overview">Background</a></h2>
- <blockquote>
- <h3>What a Profile Is (and How Profile Data Is Used)</h3>
- <p>
- When we talk of a person's <i>profile</i> we are referring to a
- store of information that may include one's name, age, gender, phone
- number, postal or electronic mail address, purchase history, web
- surfing habits, subscriptions or any of a multitude of personal
- preferences, traits and abilities. Often, a persistent cookie is
- deposited by a company's web site on one's computer or other
- personal information device so that the company can track the
- individual's behavior as they browse the company's site. More
- advanced systems, such as those used by DoubleClick, can track a
- person from site to site. The capability to accumulate and
- cross-reference this data is what supports the multi-billion dollar
- direct marketing industry.
- </p>
- <h3>Data Mining</h3>
- <p>
- Technology enables the collection and storage of vast quantities of
- data. Finding, summarizing, and creating models of the patterns,
- trends and projections from this data is what <i>data mining</i> is
- all about. It is a marriage of statistics, machine learning,
- information theory and computing that has formed a mathematical
- base for a science that has increasingly powerful tools at its
- disposal. In particular, direct marketers have created data mining
- techniques that allow them to pinpoint desired market segments for
- inclusion in their advertising and marketing campaigns.
- </p>
- <p>
- The manner in which personal (profile) information is collected
- and used today is grossly inefficient not to mention a massive
- violation of privacy. It [current standard practice] developed
- over the course of the last hundred years as capitalism matured
- and corporations grew more powerful. New, precise mechanisms
- could replace the current shotgun approach, but Industry is so far
- along the path paved by their marketers that they can't see the
- opportunity. (Remember that the marketers were originally
- beholden to Industry, but now Industry is beholden to the
- marketers.) Their fear, reinforced by the marketers, is that
- without a person's name, address, phone number and/or email address,
- they will not be able to reach the people who may be most interested
- in the products or services that they are trying to sell.
- </p>
- <blockquote>
- <p>
- [From: <i>Web-Mining: New Data Tools to Manage Web Strategy</i>
- <<a
- href="http://www.the-dma.org/cgi/dispnewsstand?article=244">http://www.the-dma.org/cgi/dispnewsstand?article=244>>
- <font color=red>This is copyrighted - paraphrase!</font>]
- </p>
- <p>
- There are three interdependent types of Web mining: usage,
- structure, and content mining.
- </p>
- <p>
- Usage mining is concerned with the discovery of site access
- patterns as logged in server access logs. The types of analysis
- range from custom reporting, usage profiling and banner ad
- targeting, to real-time recommendations and cross-sale analysis,
- to such CRM applications as customer attraction, segmentation,
- retention and Web-time value.
- </p>
- <p>
- Web structure mining refers to the application of data mining
- techniques to improve the structure and design of Web
- pages/sites. Structure mining could help assess the problem areas
- on your sites, such as major traversal paths associated with quick
- exits, preferred paths customers take to get to specific areas of
- the site, and paths that lead to sales and cross-sales. Studies
- show that people are gravitating to sites that deliver products
- and services customized to their needs.
- </p>
- <p>
- Web content mining is the mining of the content of Web pages and
- refers to the automated search, extraction and classification of
- primarily textual content information resources available
- on-line. Application areas include customer support/service,
- automated e-mail routing and reply, and knowledge management, such
- as document clustering, content categorization, and keyword
- extraction and associations.
- </p>
- </blockquote>
- <h3>Collaborative Filtering and Recommendation Systems</h3>
- <p>
- The data mining of anonymous data can have its uses, as in simple
- collaborative filtering systems. These systems collect inputs
- from many potentially anonymous people on a particular subject
- (say, what their current favorite movie is) and then average the
- results and come up with recommendations. This works with
- reasonable accuracy in a well behaved populace - that is, within a
- group that does not have shills and spoofers that may attempt to
- throw the decision one way or the other by flooding the system
- with bogus or weighted inputs.
- </p>
- <h3>Direct ("One-to-One") Marketing</h3>
- <p>
- For traditional direct marketing mechanisms to work, profile data
- must be linkable to the people that it refers to so that they may be
- reached by phone, mail (electronic or physical), banner ads or
- regional advertising campaigns. The value of such information is
- immense. One vivid example can be seen in the acquisition of
- Hotmail by Microsoft for a total of $395 million. While the
- software to create such a system was trivial, what Microsoft
- actually bought was the access to Hotmail's 10 million users.
- Another view as to the value of personal profile information can be
- seen by looking at the sales figures attributed to direct marketing:
- <blockquote>
- U.S. sales revenue attributable to direct marketing is estimated to
- reach more than $1.7 trillion in 2000. Through 2005, sales are
- estimated to grow by 9.6 percent annually to reach $2.7 trillion.
- [<i>Economic Impact: U.S. Direct Marketing Today Executive
- Summar</i> <<a
- href="http://www.the-dma.org/library/publications/libres-ecoimpact2.shtml">http://www.the-dma.org/library/publications/libres-ecoimpact2.shtml>>]
- </blockquote>
- </p>
- <p>
- Further, the Direct Marketing Association predicts that "Direct
- Marketing's E-Growth Projected to Exceed Rest Of Web In 2001" (see
- <<a
- href="http://www.the-dma.org/cgi/dispnewsstand?article=238">http://www.the-dma.org/cgi/dispnewsstand?article=238>>).
- </p>
- <h3>Privacy Concerns</h3>
- <p>
- Once all this data is collected, there are many ways that it can be
- used and disseminated by the corporations and government agencies
- that obtain it. In the simplest cases, one's profile data may be
- mined for direct mail or demographically-directed marketing
- campaigns. But it may also be used to determine health care and
- insurance premiums, credit ratings, and any of a myriad of other
- uses that the trillion dollar marketing industry may find useful.
- </p>
- <p>
- While it is reasonable that a vendor can check one's credit before
- extending same, should that vendor then be allowed to sell that
- credit information - attached to your name and address - to any and
- all takers? And as mentioned above, such trade in personal
- information is not limited to one's credit-worthiness, but also
- includes school, work and health records, purchase and travel
- history, and even sexual preferences. It is clear why many privacy
- advocates are raising alarms as we become the most tracked and
- watched society.
- </p>
- <p>
- While companies involved in e-commerce are placing "privacy policy"
- decalarations linked from their home page, they may change these
- without warning at any time <font color=red>[insert Amazon
- reference here]</font> or perhaps simply get bought, as with
- Hotmail discussed above. In Europe, there are strong laws that
- govern the collection and storage of profile data. These statutes
- allow for only the collection of data needed for a transaction
- (such as a credit report) and require that this data be destroyed
- when no longer needed. However, the United States offers no such
- protection, and as activists raise concerns, they are met with an
- industry that offers only "voluntary compliance" with no legal
- means to enforce adherance to any particular policy.
- </p>
- <p>
- [Privacy page from the DMA: <<a
- href="http://www.the-dma.org/library/privacy/index.shtml">http://www.the-dma.org/library/privacy/index.shtml>>]
</p>
</blockquote>
<h2>References</h2>