Having a presence on the World Wide Web used to be a simple matter. Companies would promote themselves and their products on webpages and hope that enough consumers would show enough interest at least to justify the cost of the website. As the Internet evolved into a vast commercial enterprise, businesses began to apply the same marketing analysis as they did offline, and technology developed to assist them. Several years ago, businesses like WebSideStory began offering dedicated web-tracking services. These services can capture and analyze many aspects of web traffic and create a multitude of customized reports. Such digital market research has become indispensable to many online businesses. Its use has also raised many concerns about privacy, generating numerous articles. We instead here address and summarize some of the significant yet largely unrecognized ways that trademark attorneys can use this data to assist their clients.
Most trademark lawyers are familiar with "cookies," which are strings of text that a web server places on a user's hard drive when delivering a requested webpage. As a result, the web server that delivered the cookie, and sometimes other servers as well, can then read the cookie when delivering another webpage to the same computer. This allows the web server to associate certain activity on a website with that computer, and, presumably, its user. Among other things, cookies can reveal how users move through a website once they arrive. For a variety of reasons, often having to do with privacy, cookies have been the focus of much attention and are presumed by many to be the primary way that websites gather data about their visitors. In fact, websites can and do collect significant amounts of information, including how a visitor arrives at a particular site, through other means. Many websites use a variety of technologies to track the online movements of their visitors.
As an example, websites can engage in "referrer logging," the technical term for collecting the URLs that have sent visitors to a particular webpage. Businesses use this data to evaluate the effectiveness of their affiliate marketing or other e-commerce strategies. But this information can be equally valuable to the trademark attorney. These logs may reveal whether unscrupulous entities have registered domain names confusingly similar to a client's trademark, such as common misspellings or domain names that are composites of a trademark. For instance, The Hertz Corporation's discovery that many visitors arrived from www.herts.com, a website offering discount car rentals, could be probative of actual confusion in an action against the registrant of that domain name. (The examples in this article are hypothetical and are for illustrative purposes only. In fact, the website www.herts.com is inactive, and the authors of this article do not intend to imply that its owner has engaged in any misconduct.)
Web-tracking statistics also include the identity of referring search engines and the number of visitors from each. Businesses find this information helpful in determining which search engines are generating traffic to their websites. These statistics also tally the actual keywords searched that eventually led visitors to the site. Thus this data identifies which search terms the general public associates with a client's product or service. As a result, it can help evaluate the conduct of a potential defendant in a trademark infringement action. For instance, a discovery by Calvin Klein that "colored jeans" was a search term used by ten percent of the visitors to its site could be very useful in a case against a company selling jeans designated "Cal Kline: the colored jeans."
The keywords can also be searched in a search engine such as Google or a web-tracking site such as Overture to determine whether competitors are diverting traffic from a client's website by purchasing sponsored links based on those keywords. Overture also shows the number of searches made of a word or phrase, and even suggests keywords for particular types of business. (See .) This data can be evidence of secondary meaning or of the fame of a mark.
Web-tracking data also reveals the amount of traffic from country code domains (ccTLDs), which can be of great value to a client trying to establish international renown or merely to demonstrate that its mark is being used in a foreign jurisdiction in order to establish foreign rights. In the litigation context, this data may also provide a factual basis to assert jurisdiction in the United States over a foreign defendant. In a licensing arrangement, it can help audit compliance with territorial restrictions.
Bear in mind that web-tracking data is not 100% accurate or complete. Data on referring URLs normally cannot distinguish among those users accessing a website from a "bookmark" in their browsers, those clicking on a link in an e-mail message, and those typing the URL directly into the browser. Domain names do not necessarily reveal a visitor's physical location (for example, if the visitor is logged into a corporate network while browsing). On the other hand, technology has been developed that can track even more aspects of users' online behavior. Examples are pixel tags, keystroke-loggers, and various kinds of "spyware." Although these methods have benign as well as nefarious uses, concerns about breaching privacy have caused many companies to avoid them.
Web-tracking data can help gauge damages, which can be increased for willfulness under 15 U.S.C. § 1117. The data can show how a defendant modified its webpages to exploit consumers' recognition of a plaintiff's trademark. An example from traditional marketing will help illustrate this technique. Direct-mail solicitations for a particular product often vary in one aspect, such as the tagline on an envelope meant to get the recipients' attention. This seemingly minor difference is then tracked to determine which promotional slogan was most successful. Because response rates in direct marketing are so low, even a small increase for one slogan over another is economically significant.
The same can be done on the web. Consider the example of a travel aggregator that uses the trademarks of well-known hotels in its banner advertisements. The aggregator can track which mark most often led consumers to the order-placing webpage, generating sales. The aggregator's resulting change of its banner ads could be evidence of willful infringement. Calculating damages can further benefit from some highly focused statistics showing the revenue per referring URL, and the browse-to-buy ratio - the ratio of the number of orders to the number of visits per visitor.
As in print marketing, the layout of webpages is important. The location of a hyperlink can affect its click-through rate. Many companies track this statistic. Discovery in litigation could disclose that the defendant researched which area of its webpage generated the highest click-through rate, and then used the plaintiff's mark in that spot. This could be strong evidence of willful infringement.
Similarly, web-tracking data can reveal entry pages, the webpages at which visitors began navigating a website. The data can even show the revenue per entry page (the ratio of total orders to a particular entry page visited before the confirmation-of-purchase page). A defendant's modification of its homepage or other entry pages can be further evidence of willfulness.
The Internet offers many other interesting tools not yet fully exploited by the trademark bar. For instance, the Overture search engine has a tool showing the number of searches made of a word or phrase in the last month. See . This data can serve as evidence of secondary meaning, or of the fame of a particular mark in a case of dilution.
For instance, in Playboy Enters., Inc. v. Netscape Communs. Corp., 55 F. Supp. 2d 1070 (C.D. Cal. 1999), rev'd, 354 F.3d 1020 (9th Cir. 2004), Playboy alleged that Netscape's sale of the keywords "playboy" and "playmate" (to purchasers whose banner advertisements would then be displayed on webpages generated by searches for those terms) infringed its well-known PLAYBOY and PLAYMATE marks. Noting that Playboy had neither produced a survey nor demonstrated that users typing "playboy" or "playmate" into a search engine necessarily sought the services provided under the plaintiff's marks, the district court ruled that Playboy had failed to establish likelihood of confusion. Playboy, 55 F. Supp. 2d at 1084.
A review of statistics from Overture, however, reveals that in July 2004, most of the more than 109,000 searches for "playmate" were for "playboy playmate," "playmate gallery," "playmate of the year," and other phrases clearly associated with Playboy magazine. Had it been available to Playboy in 1999, this data would certainly have provided compelling evidence of secondary meaning and likelihood of confusion. (Even without such evidence, the Ninth Circuit subsequently rejected as absurd defendant's contention that it had used those words in their primary or generic sense. Playboy, 354 F.3d at 1028.)
Trademark law has changed drastically as a result of the Internet. Although much has been written on using the Internet to search and clear trademarks and to investigate potential infringement, web-tracking data remains an untapped resource. While this article provides a basic introduction to these tools, creative attorneys will no doubt find many other ways to benefit from them.
Stephen W. Feingold is a partner in the New York office of Pitney Hardin llp, where he leads the Trademark and Copyright Group. He is also Adjunct Professor of Trademark Law at Fordham University School of Law. Gerry A. Fifer and David H. McDonald are attorneys in the Group also based in the firm's New York office.