14.7 Customer Profiling and Behavioral Targeting

Advertisers are willing to pay more for ads that have a greater chance of reaching their target audience, and online firms have a number of targeting tools at their disposal. Much of this targeting occurs whenever you visit a Web site, where a behind-the-scenes software dialogue takes place between Web browser and Web server that can reveal a number of pieces of information, including IP address, the type of browser used, the computer type, its operating system, and unique identifiers, called cookies.

And remember, any server that serves you content can leverage these profiling technologies. You might be profiled not just by the Web site that you’re visiting (e.g., nytimes.com), but also by any ad networks that serve ads on that site (e.g., Platform-A, DoubleClick, Google AdSense, Microsoft adCenter).

IP addresses are leveraged extensively in customer profiling. An IP address not only helps with geolocation, it can also indicate a browser’s employer or university, which can be further matched with information such as firm size or industry. IBM has used IP targeting to tailor its college recruiting banner ads to specific schools, for example, “There Is Life After Boston College, Click Here to See Why.” That campaign garnered click-through rates ranging from 5.0 to 30 percent (Moss, 1999) compared to average rates that are currently well below 1 percent for untargeted banner ads. DoubleClick once even served a banner that included a personal message for an executive at then-client Modem Media. The ad, reading “Congratulations on the twins, John Nardone,” was served across hundreds of sites, but was only visible from computers on the Modem Media corporate network (Moss, 1999).

The ability to identify a surfer’s computer, browser, or operating system can also be used to target tech ads. For example, Google might pitch its Chrome browser to users detected running Internet Explorer, Firefox, or Safari; while Apple could target those “I’m a Mac” ads just to Windows users.

But perhaps the greatest degree of personalization and targeting comes from cookies. Visit a Web site for the first time, and in most cases, a behind-the-scenes dialogue takes place that goes something like this:

Server: Have I seen you before?

Browser: No.

Server: Then take this unique string of numbers and letters (called a cookie). I’ll use it to recognize you from now on.

The cookie is just a line of identifying text assigned and retrieved by a given Web server and stored on your computer by your browser. Upon accepting this cookie your browser has been tagged, like an animal. As you surf around the firm’s Web site, that cookie can be used to build a profile associated with your activities. If you’re on a portal like Yahoo! you might type in your zip code, enter stocks that you’d like to track, and identify the sports teams you’d like to see scores for. The next time you return to the Web site, your browser responds to the server’s “Have I see you before?” question with the equivalent of “Yes, you know me;,” and it presents the cookie that the site gave you earlier. The site can then match this cookie against your browsing profile, showing you the weather, stock quotes, sports scores, and other info that it thinks you’re interested in.

Cookies are used for lots of purposes. Retail Web sites like Amazon use cookies to pay attention to what you’ve shopped for and bought, tailoring Web sites to display products that the firm suspects you’ll be most interested in. Sites also use cookies to keep track of what you put in an online “shopping cart,” so if you quit browsing before making a purchase, these items will reappear the next time you visit. And many Web sites also use cookies as part of a “remember me” feature, storing user IDs and passwords. Beware this last one! If you check the “remember me” box on a public Web browser, the next person who uses that browser is potentially using your cookie, and can log in as you!

An organization can’t read cookies that it did not give you. So businessweek.com can’t tell if you’ve also got cookies from forbes.com. But you can see all of the cookies in your browser. Take a look and you’ll almost certainly see cookies from dozens of Web sites that you’ve never visited before. These are third-party cookies (sometimes called tracking cookies), and they are usually served by ad networks or other customer profiling firms.

By serving and tracking cookies in ads shown across partner sites, ad networks can build detailed browsing profiles that include sites visited, specific pages viewed, duration of visit, and the types of ads you’ve seen and responded to. And that surfing might give an advertising network a better guess at demographics like gender, age, marital status, and more. Visit a new parent site and expect to see diaper ads in the future, even when you’re surfing for news or sports scores!

While the Internet offers targeting technologies that go way beyond traditional television, print, and radio offerings, none of these techniques is perfect. Since users are regularly assigned different IP addresses as they connect and disconnect from various physical and Wi-Fi networks, IP targeting can’t reliably identify individual users. Cookies also have their weaknesses. They’re assigned by browsers and associated with a log-in account profile on that computer. That means that if several people use the same browser on the same computer without logging on to that machine as separate users, then all their Web surfing activity may be mixed into the same cookie profile. (One solution is to create different log-in accounts on that computer. Your PC will then keep separate cookies for each account.) Some users might also use different browsers on the same machine, or use different computers. Unless a firm has a way to match up these different cookies with a single user account or other user-identifying information, a site may be working with multiple, incomplete profiles.

References

Moss, M., “These Web Sites Know Who You Are,” ZDNet UK, October 13, 1999.