Digital Media Buzz > Behavioral Targeting: I Always Feel Like Somebody’s Watching Me

Behavioral Targeting: I Always Feel Like Somebody’s Watching Me

Next Page « 1 2 »

By Jon Donley

In late 2007, Facebook launched a new program called Beacon, a collaboration between the growing social media site and more than 20 vendor sites, including Blockbuster, eBay and Travelocity. Using behavioral targeting technology, the service collected data on customers of the vendor sites, and then shared it with the users’ Facebook friends.

An uproar by Facebook users and privacy advocates resulted in changes to the program and an apology from CEO Mark Zuckerberg. Users now have to agree to participate in the program up front, rather than having to opt out.

It wasn’t the most auspicious time for Facebook’s gaffe. The Federal Trade Commission was in the process of launching a review of behavioral advertising — or behavioral targeting — for possible regulation. A staff report on that review was unveiled earlier this year, with recommendations that stopped short of mandating regulations, but put the industry on notice that only a proven record of self-regulation will stop such a move.

And Congress this week launched hearings to consider whether new federal laws are needed to protect consumer privacy in the context of behavioral advertising.

The 411 on Behavioral Advertising
The FTC defines behavioral advertising as “the practice of tracking an individual’s online activities in order to deliver advertising tailored to the individual’s interests.”

Through a mixture of web traffic analytics, traditional marketing practices updated for the web, and various tracking programs, advertisers follow individuals online from site to site, or within a specific site, learning habits and preferences, and noting transactions and other activities. Studies show that most web users don’t know the data collection is taking place; without some measure of technical savvy, the activity is invisible. Usually this growing profile of the individual is not personally identifiable.

The increasingly sophisticated models of behavioral tracking allow advertisers to more surgically deliver ads to individuals whose online activities indicate they might be interested in particular goods or services. This not only increases the efficiency of ad purchases, but allows more formalized ROI measurement.

Key tools and concepts of behavioral targeting include:

• Cookies — Tiny bits of text paced on the users’ computers, allowing data to be tracked by the site or service that installed the cookie. Cookies are ubiquitous on the web, and often are required for full functionality of the site. Web browsers are normally set, by default, to accept first-party cookies (from the site you are browsing), but can be set to reject cookies altogether.
• Web bugs or beacons — A one-pixel graphic designed to be invisible to the user, which is stored in the browser cache and carries tags and tracking information. Facebook uses a combination of cookies and beacons in its Beacon product.
• Flash cookies — Information stored within Flash files, similar in function to a traditional text cookie.
• First-party tracking — The use of cookie-powered web analytics tools to observe and learn from consumer behavior on a publisher’s website. Information can be used not only in ad delivery, but in observing user behavior to identify popular content and guide decisions on design and navigation. A prime example is Amazon.com, which gives suggestions and sends offers based on your own behavior and purchases on the service, as well those of other users who have bought or searched for similar items.
• Advertiser network tracking — Much internet advertising is delivered by ad networks. The network can use cookies to collect data about individual users as they move from one website to another within the advertising network. The network can sell and deliver ads based on this data.
• Third-party tracking — The collection of user information by one site, which then sells or shares consumer data to other networks or sites for targeted advertising.
• Deep-packet inspection — A controversial practice in which your Internet service provider tracks your “click stream,” or surfing habits, and sells targeted advertising based on your habits. While this method solves some of the problems with cookies, such as user deletion, it is seen as a greater threat to privacy, similar to the telephone company listening to your phone calls and selling the information to advertisers.
• Personally identifiable information (PII) — Data such as name, address, Social Security number and telephone number, which specifically identify someone.
• Non-PII — More general information, such as the IP address, purchases and browsing history, which cannot be used to directly identify someone.
• Contextual advertising — The delivery of advertising based on the content of the page. Ads delivered on a Google search results page, for example, are based on the context of the search. Google AdSense banners and text ads are based on keywords purchased by advertisers, and delivered to sites featuring related content.
• Retargeting — Delivering targeted ads to consumers who have visited a site in the past, with the idea of reminding a user of previous interest, and possibly spurring a reconsideration of the product or service.
• Interest-based targeting — Delivering ads based on the user’s history of interest in a topic. Google announced this type of behavioral targeting in March. The feature, which is in beta, and is to be opened to the public later this year, allows AdSense customers to choose from hundreds of topical target areas, and suggest more.
• Predictive targeting — Using demographic profiles and psychographics to deliver advertising based on statistical inferences of what a particular profile is likely to want to buy. This data-mining is more scientific than intuitive and estimates what is likely to interest a particular group.

Next Page « 1 2 »