Privacy and Data Mining

When the 2003 version of the Outlook e-mail program came out, it would by default not display graphics in HTML e-mails where they were to be downloaded from the Internet once you opened the mail. The message that displayed in place of the graphic was “To help protect your privacy, Outlook prevented automatic download of this picture from the Internet”. What did this strange message mean? How would downloading a picture damage my privacy?

There are two issues here. The first is that there are some pictures you may not want to see and which could potentially offend. This is, however, a not very common worry.

The second, and more important, issue is that once you download the pictures, the sender of the e-mail knows you have opened the mail. Thus your anonymity is compromised. They may not know who you are, but they do know you are someone who has at least bothered to open the e-mail before deleting it. By downloading the picture you provide information to the sender that the e-mail address is live and someone is looking at the mail, even if only for a moment.

And of course if you click on any of the links in the mail and go to the sender’s landing pages, you provide the sender with more information about yourself and your viewing habits. The marketer can then start to build a profile about you. They can use cookies to track what you do on their sites over time. If you purchase anything, the amount of information they have about you starts to increase rapidly.

This brings us to the subject of data mining. I’m using data mining in the popular sense as being any activity that collects and analyzes data for marketing purposes. Professionals know that data mining is more properly defined as automated methods of discovering hitherto unknown patterns in any set of data. But the popular media have pretty well equated it with an invasion of privacy.

There are two broad categories of actions that marketers can take with data collected about individuals. First, they can target the person directly with offers that might interest the individual and prompt them to purchase goods or services. Second, the individual can be grouped together with others who seem to have similar characteristics in order to determine how that group behaves, leading to marketing that can better target that group or similar groups.

The first scenario, personal targeting, is the thing that consumers often find intrusive and sometimes just plain “spooky”. They don’t want their mail boxes or e-mail in-boxes filled with offers for things that they would normally not consider purchasing. They have a low tolerance to ads that say, “you might also be interested in...” or “other people who bought this product also bought ...”. Those kinds of suggestions must be used sparingly or the consumer is turned off.

Some consumers consider the use of their spending habits for market analysis as an invasion of privacy as well. But usually they are not even aware this is going on. Most people use grocery store discount cards with hardly a thought as to what is being analyzed about their shopping habits – they just like getting the discounts that are available. The use of the data gathered from such loyalty card programs has been limited, primarily due to the fact that marketers aren’t sure how far they can go towards targeting consumers with specific ads without offending them and driving their business away.

The controversy over data mining is just one more symptom of our conflict between identity and privacy in this digital age. We want to stay anonymous to feel safe, but we also have a need to be identified properly when we want to transact our business, on-line or off-line.

I’ll be talking further about the subject of identity and privacy at my presentation at Enterprise Data World later this month in San Francisco. My presentation will be at 8:00 a.m. on March 17. For more information on the conference, see:
http://edw2010.wilshireconferences.com/