Personally Identifiable Data Can Be Made From “Anonymous” Data
Online Marketers Who Collect Visitor Data Are Potential Targets of Regulation in 2010.
Law and health care practices are required to protect personally identifiable information (PII). However, in many cases they are encouraged to circulate so-called anonymous data. It turns out the distinction between the anonymous and the personally identifiable isn’t all that real.
Latanya Sweeney, a Carnegie Mellon University computer science professor, took anonymous data from medical records and used it to identify real patients.
She took what seemed like anonymous data from a health record: Mr. X lives in zip code 02138 and was born July 31, 1945. And by using publicly available databases, she found that only one one person in the US matched those attributes: former governor William Weld.
In a very real sense, “anonymized” or “merely demographic” information about people may be neither. Web sites that ask “anonymous” users for seemingly trivial information (such a zip code or birth date) can use the same techniques to create a unique record and connect them back to extensive consumer databases.
If you are a marketer using “anonymous” web data, or matching web leads to demographic databases, then your data may be regulated in the next few years.
Over the next year, I believe we’ll see increasing regulatory interest from the FTC and other bodies. When Sears paid users to download an application that tracked their site use, the FTC intervened. They are regulating what gifts bloggers can take, and what information ad networks can share for behavioral targeting.
The future of digital experiences will be built by strategists who grasp the full array of emerging business, social, and technical models. Specialties in user experience, branding, application design, and data science are laying the foundation for richer user experiences and business models breakthrough products and revenue based marketing.
2 Responses to "Personally Identifiable Data Can Be Made From “Anonymous” Data"
May 21, 2010
Your article is good, I like it very much!
Your point is very positive.
June 15, 2015
[…] of residence and what type of car they own. In fact there’s better than 80% chance of personally identifying people through three pieces of non-personal data (zip code, birthdate, and gender). As Facebook, Apple, and others collect your specifics, they can […]