How Apple wants to track you

Flickr user hyku

Flickr user hyku
COMMENTARY Apple (AAPL), Google (GOOG), Microsoft (MSFT), Twitter, and others regularly claim how important user privacy is to them and often point fingers at one another over alleged privacy violations. But sometimes you get a peak under the marketing machine's hood for a sense of whether "anonymous" data is really all that anonymous. And that's what Apple patent application 20120042262, recently made public, offers.

Called Population Segmentation Based on Behavioral Patterns and filed in August 2010, the application outlines at least one way that Apple pulls together everything -- previous behavior shown by your account, device information, content that you request, and more -- to decide what messages you should receive by putting you into the most appropriate marketing segments.

Privacy "fixes" will do nothing for consumers
Is Google ignoring Internet privacy?
Google wants to track calls related to online ads

Although Apple is explicitly in the content business through providing music, e-books, video, and more, the "potential targeted content" that the company seeks to deliver means far more than suggestions of songs that you might like, as the document's description makes clear:

Targeted content delivery has long been an accepted means of conveying a desired message to an audience. Instead of creating a single message and delivering it to every member of the general public, content providers attempt to identify a particular segment of the population that is likely to have the greatest interest in their message. For example, a content provider might wish to convey a message regarding a service offered in a particular city. To convey this message, the content provider could send out a flyer to all residents of the city. However, if the service is only of interest to residents that own their own home, then targeting all residents of the city is suboptimal for the content provider. Instead, the content provider will attempt to segment the population of city residents into home owners and non-home owners and then only distribute their message to the segment of the population that are home owners. Population segmentation enables content providers to optimize their resources.

The development of digital content delivery has enabled new techniques of identifying population segments. For example, segments characterized by mobile device users or users that visit social networking sites. However, these segmentation techniques are often overly simplistic or too broad because they are based on a limited number of user characteristics.

Although the description could apply to entertainment or informative content, the phrasing -- targeting recipients for a flyer about a company's services -- paints a clear picture of advertising and marketing. In fact, Apple explicitly notes that the "assembled content package" directed at the user "can include one or more types of advertisements from one or more advertisers."

How it works

The first independent claim gives a technical explanation of the process that Apple uses:

A computer-implemented method, the method comprising: obtaining user characteristic data associated with an identified user requesting a content package; associating the identified user with one or more pre-defined behavioral segments based on behavior patterns in the user characteristic data associated with one or more content types; inferring a current mode of the identified user based on the temporal relationships of the behavior patterns; selecting one of the behavioral segments based on the current mode; assembling a content package for the identified user comprising invitational content related to the selected one of the behavioral segments; and ranking the behavioral segments based on their temporal relationship to the current mode.

As the description more plainly explains, Apple can compile user characteristics, including the following examples explicitly mentioned by the patent application:

-- A unique device identification number
-- Information about the software running on the device
-- The application requesting content, which could be an ad
-- The type of Internet connection the device is using
-- A unique account number with information stored as user characteristics
-- Interactions with content
-- Web searches performed by the user
-- Information from public databases, like census data of the predominant population in the area in which the user lives, associated with the user

There is far more that Apple could learn by inference. Purchase a product targeted at a particular age demographic like teenagers and the company can assume that the person is a teenager. As Apple explains: "In short, the user shares a good deal of information about him, even if the data isn't about personal attributes, when he makes a request for data. All of this data can be stored in association with the user as a user characteristic."

Another competitive patent weapon?

There are also the competitive implications of the patent application, if granted as drafted. That first independent claim, key to potential legal protection, is effectively a broadly written statement that discusses obtaining "user characteristic data" and matching it to "pre-defined behavioral segments based on behavior patterns" based on a "current mode" that can indicate such things as frequency or duration of interactions with types of content, and then sending out the appropriate "content package," which likely means advertising. That pretty well sums up the heart of segmented behavioral marketing.

This is only an application, not a granted patent. But it does give insight into how Apple approaches behavioral marketing, and perhaps why it wants to so carefully guard user data -- because it's a valuable commodity.

  • Erik Sherman On Twitter»

    Erik Sherman is a widely published writer and editor who also does select ghosting and corporate work. The views expressed in this column belong to Sherman and do not represent the views of CBS Interactive. Follow him on Twitter at @ErikSherman or on Facebook.

Comments