Second Thoughts on a GPhone: Privacy and Targeted Ads

A couple of months ago, the Wall Street Journal provided some speculation about a Google phone in an article titled Google Pushes Tailored Phones To Win Lucrative Ad Market.

It’s difficult to tell how much is speculation, but Google CEO Eric Schmidt has gone on record as saying that he believes that consumers will watch targeted ads in exchange for free cell phone service.

How exactly would targeted ads work on a cell phone? Would they check out whom you’re calling, and target ads based upon whether your calls go to the local cheap pizzeria for delivery instead of making reservations at very expensive eateries?

If you prefer to stay at 5 star hotels or at basic budget lodgings, how might the ads that you see on your phone differ? If most of the people you call have Italian last names, might pasta feature prominently in the targeted ads that you see?

A patent application from Google looks at user activity on a phone to determine the ads show. The actual content of calls doesn’t seem to play a role in deciding what advertising is shown (except maybe for interactions with menu systems), but called numbers of the user of the calling device could be.

The costs of the calls might be subsidized by ad revenues, subject to appropriate user permission/authorization.

Information derived from calling activity might be weighted to determine which ads to show by looking at:

  • Duration of calls
  • Costs of calls
  • The Frequency that a number was called
  • The time lapsed since a call was last made
  • Frequency of calls to the same number
  • Number of calls to the same number
  • Frequency of calls to a type of person
  • Frequency of calls to a type of organization
  • Frequency of calls to a type of business
  • Number of calls to a type of person
  • Number of calls to a type of organization
  • Number of calls to a type of business
  • Location based upon the area code of called number

The patent application:

Targeting and/or scoring advertisements using information derived from called telephone numbers or the called telephone numbers themselves
Invented by Shumeet Baluja and Michael Chu
US Patent Application 20070239529
Publisehd October 11, 2007
Filed: March 30, 2006

Abstract

Called telephone number(s) from a client device are stored. The called number(s) is used to lookup information related to the called number(s) (e.g., a name and address).

Additional information might be obtained about the called party including the type of business, classifications of the products of the business, etc.

This information may then used to select and/or store advertisements to serve to the client device (or to the user of the device, if known, when that user is on another device such as a Web device).

Advertisements can be scored according to how well they match the information derived from the called telephone number(s).

Choosing Keywords Related to Calls for Targeted Ads

You call a movie theatre, and the system stores the phone number, and uses a reverse-directory lookup on the number to see that a “movie theater” was called.

In the future, you may receive ads associated with movie theaters, movies, DVD’s, home theater electronic systems, etc., in the future.

Targeting and/or Scoring Ads Based Upon Information Derived From Calls

Numbers dialed might be used to look up related information, such as:

  • Geographic locations of the called numbers
  • Names of persons called
  • Names of businesses called
  • Names of organizations called
  • Types of business called
  • Types of organizations called

This kind of information might be used to build a profile for a specific client device (mobile phone, PDA, Desktop phone using VOIP, etc.) or for the user of the calling device.

A technology like that used for pages carrying Adsense might be used in combination with keywords created for that profile, to decide which ads to target, by building a page document that contains a string of the keywords in the profile, such as the store names and categories of the stores called by the user.

While that page might not make compelling reading, it could be used by ad targeting mechanisms, like Google’s AdSense technology) to decide upon potential ads to be served.

How Information from a Call Might be Expanded

The patent filing lists a few examples of how information from calls, and information looked up from those calls could be expanded to help find targeted ads:

1) Business name and Yellow Page Listing — can determine the type of business

  • Best Buy -> electronics retailer
  • Dick’s -> sporting goods
  • Lowes -> Home Improvement

2) Business name and databases — to determine services and/or products offered by the business

  • Midas->car service, mufflers, brakes, oil change
  • Eddie Bauer -> {inventory}

3) A person’s given name or surname — in conjunction with etymological knowledge to determine nationality and/or ethnicity information

  • Perez -> Spanish
  • Chow -> Chinese
  • Miller -> English
  • Xanthopoulos -> Greek
  • Petrov -> Russian
  • Hashimoto -> Japanese
  • Sudano -> Italian
  • Patel -> Indian
  • Beck -> German
  • O’Hara -> Irish

4) Place Names — together with available information

  • Orlando, Fla. -> Disney, family vacations
  • Bentonville, Ak. -> WalMart, retailing
  • Las Vegas, Nev. -> gambling, shows

For example, the term “Orlando, Fla.” might be entered as a search query and the top number of search results could be analyzed to determine related information.

Another example, the term “Orlando, Fla.” might be used as a search term in an online encyclopedia and an entry, if any, could be analyzed to obtain related information. That could also be done with other “seed” information such as a business name, an organization name, or a person’s name.

5) Economic Strata of Callers

For example, an economic stratum of the caller (a call to a Motel 6 or a Kia dealer versus a call to a Four Seasons Hotel or a Mercedes dealer) could be determined by further classification and/or analysis.

For example, the average price of the products sold by the merchants or the average selling price of the products produced by the manufacturers (perhaps normalized to the prices of similar products–$15,000 is inexpensive for a new car, but is expensive for a watch) can be analyzed.

As another example, the mean income (e.g., from the US census data bureau) of the locations of the places that are called can be analyzed.

6) Key presses in response to voice message system prompts

As an example of using such key presses, instead of simply noting that the user called a local theater, by analyzing the voice prompts (which might have been previously crawled (e.g., a list of numbers of voice message systems could be called and crawled by entering numbers) and analyzed (e.g., using speech recognition for example)) and responsive key press responses, it might be learned that the user was interested in a specific movie (“Finding Nemo” versus “The Matrix”), not just that they called a movie theater.

This could result in ads being served related to that specific movie, or based on the genre of the movie (e.g., children’s movies versus action and science fiction movies).

Such a system may require a device that would call dialed numbers (that were followed with further dialed digits), and utilizing speech recognition technology, learn which terms were associated with each of the possible choices.

Conclusion

The patent filing contains more details on ad targeting and how information can be derived from calls, and calling activity. It also provides a number of additional examples of the use of calling information and the creation of profiles with that information.

The collection of calling information in this manner does seem pretty invasive. The word “privacy” doesn’t appear once in the patent application, but it also doesn’t say anything about capturing information from calls themselves.

If a system like this was created, it could involve the collection of a lot of information that people likely wouldn’t want known by others. How long would that data be retained? How would it be protected?

Does it potentially violate US laws involving pen registers, which are devices or programs that “record all numbers dialed from a particular telephone line?”

If someone agrees to allow their calls to be tracked and monitored, and their call history recored, and another person who hasn’t agreed to such a condition borrows the phone to make calls, is their privacy being violated?

Share

2 thoughts on “Second Thoughts on a GPhone: Privacy and Targeted Ads”

Comments are closed.