Wells Fargo & Co. v. WhenU. Com, Inc.

MEMORANDUM OPINION AND ORDER DENYING PLAINTIFFS’ MOTION FOR PRELIMINARY INJUNCTION EDMUNDS, District Judge. INTRODUCTION Almost everyone who has surfed the Internet on his or her computer has encountered advertisements that pop-up from time to time. While the average Internet user may find the advertisements annoying, the question before the Court is whether they violate trademark or copyright law. Plaintiffs Wells Fargo & Co. and Quicken Loans, Inc. have asked the Court for a preliminary injunction against Defendant WhenU.com, Inc., whose business is Internet contextual advertising. For the reasons stated in this opinion, *737 plaintiffs’ motion for preliminary injunction is DENIED. FINDINGS OF FACT The Parties 1. Plaintiff WFC Holdings Corporation is a corporation incorporated in Delaware with its principal place of business in San Francisco, California. Complaint ¶ 6. 2. Plaintiff Wells Fargo & Company (collectively with WFC Holdings Corporation, “Wells Fargo”) is a corporation incorporated in Delaware with its principal place of business in San Francisco, California, and is the parent company of WFC Holdings Corporation. Complaint ¶ 7. 3. Plaintiff Quicken Loans Inc. (“Quicken Loans”) is a corporation incorporated in Michigan with its principal place of business in Livonia, Michigan. Complaint ¶ 9. 4. Plaintiffs Wells Fargo and Quicken Loans operate websites through which certain financial services are offered. Complaint ¶¶ 8,10. 5. Defendant WhenU.com, Inc. (“WhenU”) is a corporation incorporated in Delaware with its principal place of business in New York, New York. Answer ¶11. Plaintiffs’ Trademarks and Copyrights 6. The term “Wells Fargo” is a registered trademark of plaintiff Wells Fargo. PX 115. Wells Fargo has also received trademark registration for the logo it uses on its website. PX 117. 7. “QUICKEN LOANS” is a federally registered trademark, U.S. Reg. No. 2,528,282, owned by Intuit Inc. Tr. I (Ta-zelaar) 77. Intuit granted Quicken Loans an exclusive, nontransferable, non-assignable, perpetual license to the “QUICKEN LOANS” mark. Tr. I (Tazelaar) 77. See also PX 120. 1 8. Both Wells Fargo and Quicken Loans filed for copyright registration of their websites earlier this year. PX 114; PX 119. The copyright office granted registrations to the website “computer program.” PX 192; PX 193. Plaintiffs’ Businesses 9. The business of Quicken Loans consists of offering mortgages to customers. Tr. I (Stapp) 71:1-3. Quicken Loans services customer transactions through the Internet via the Quicken Loans website. Tr. I (Stapp) 68:10-15. 10. Quicken Loans customers appear to have some sophistication concerning mortgages. Currently, 85% to 90% of the Quicken Loans business consists of customers seeking to refinance prior mortgages, as opposed to first time home buyers. Tr. I (Stapp) 82:25-83:2, 140:17-141:3. Approximately 40% of these individuals are repeat customers who are already familiar with Quicken Loans, and the process of obtaining mortgages online. Tr. I (Stapp) 83:3-5. 11. Wells Fargo is a financial services company offering customers online access to various financial services and products. Tr. I (Tazelaar) 162:9-15. Many of these services and products are the sort that would likely be used by relatively sophisticated consumers, such as business services, brokerage services, wealth management services, and estate planning. PX 103, ¶ 5. Wells Fargo has offered financial products through the Internet since 1995. Tr. I (Tazelaar) 164:5-7. 12. Wells Fargo has made extensive use of its name and marks in interstate commerce in the United States and *738 throughout the world. See Pis.’ Proposed Findings of Fact ¶ 18. The Wells Fargo marks have been associated with the Wells Fargo business since its founding in 1852 and they are distinctive designations of the corporation, its services, and its website. See id. Quicken Loans also has made extensive use of its name and mark in interstate commerce in the United States; and its mark now is associated with the corporation and is a distinctive designation of the corporation, its services and its website. See id. ¶ 23. Wells Fargo and Quicken Loans have expended substantial resources to advertise their products and services, including advertisement on the World Wide Web. See id. ¶¶21 & 24. Both entities are widely recognized as industry leaders in their respective areas of service. See id. ¶¶ 19 & 22. The Business of WhenU 13. WhenU delivers online “contextual marketing” to computers via its proprietary software product, “SaveNow.” 2 Tr. VII (Naider) 19:25-20:5. 14. Contextual marketing technology endeavors to market products and services to consumers who have a demonstrable interest in those products and services. Tr. VII (Naider) 22:23-24:20. Traditionally, contextual marketing has been conducted by assembling large databases containing a wide variety of personal information about individual potential customers and their past purchasing behavior. Tr. VII (Naider) 23:24-24:20, 27:9-18; DX 501, ¶23. 3 15. WhenU’s proprietary software allows WhenU to deliver contextually relevant advertising at the moment the consumer demonstrates an interest in the product or service, without any knowledge of the consumer’s past history or personal characteristics. Tr. VII (Naider) 27:19-28:9. 16. WhenU’s participating consumers receive contextually relevant advertisements, delivered to them computer screens (also known as “desktops”). These advertisements are selected by SaveNow, based on a proprietary analysis of the consumer’s immediate interests, as reflected by the consumer’s Internet browsing activity. Tr. VII (Naider) 19:25-20:5; DX 501, ¶¶ 29-36. 4 17. Since launching its service in approximately early 2001, WhenU has delivered online marketing for more than four hundred advertisers, including such well-known companies as Bank of America, Citibank, Verizon, JPMorgan Chase, Panasonic, Cingular Wireless, Merck, and ING Bank. Tr. VII (Naider) 32:22-33:8; Tr. VIII (Naider) 47:16-18. How WhenU Distributes Its Software 18. WhenU offers its software under two brand names: “Save” and “SaveNow.” The two applications are identical in function; they differ only in their identifying *739 descriptions and method of distribution. Tr. VII (Naider) 33:18-35:13. 19. Consumers typically download the “Save” and “SaveNow” software in return for obtaining a free software application. Tr. VII (Naider) 54:7-14. In some cases, consumers are offered a choice between paying for a “premium” version of the desired application, or obtaining the desired application for free, but bundled with Save. Tr. VII (Naider) 55:11-20. For example, the Bearshare software application is marketed to consumers in two forms: a premium version that costs $19.95 to download, and a free version that comes bundled with Save. Tr. VII (Naider) 34:8-14. WhenU has also developed its own freeware applications, including an application called “Weathercast,” which are bundled with Save software. Tr. VII (Naider) 20:16-18, 30:6-31:4. 20. The SaveNow software is also typically obtained as part of a “bundle” with another software program, such as the popular Living Coral or Living Waterfall screen savers, but the user is not obligated to keep SaveNow in order to use the free software. Tr. VII (Naider) 34:15-22, 54:7-12. 21. WhenU shares the revenue generated from its bundled software with its bundling partners. Tr. VII (Naider) 54:18-24. The bundling of revenue-generating, advertising software (“adware”) with free software programs (“freeware”) is a common practice. Many software companies rely on the revenue generated by advertising software in order to offer freeware for free and to provide service and support for their freeware programs. DX 501, ¶ 37; DX 523, ¶¶ 74-75. 22. SaveNow is also available for download at WhenU’s website (DX 501, ¶ 40), at the websites of WhenU’s free applications such as mmu.getweathercast.com (DX 501, ¶ 39), and at certain third-party websites via a software download prompt screen that offers a user surfing the Internet the opportunity to download, for example, Weathercast and Save. Tr. VII (Naider) 53:7-16; DX 538, ¶ 13; DX 539. Although many users claim not to be aware that SaveNow has been loaded on to their computer, the Court finds that some user assent is required before SaveNow is downloaded. The fact that assent may be in the form of a reflexive agreement required for some other bundled program does not negate the fact that the computer user must affirmatively ask for or agree to the download. The Download Process and the SaveNow License Agreement 23. Although there are variations in the SaveNow download process, depending upon the other applications that the computer user is installing, certain key features of the download remain constant. For example, during the installation process, the consumer always receives a notice stating that SaveNow is part of the download, and explaining how SaveNow functions. Tr. VII (Naider) 57:18-58:22; DX 501, ¶ 38. 24. Regardless of the method of distribution, to proceed with the installation of SaveNow, the consumer must affirmatively accept a license agreement for SaveNow (the “License Agreement”). Tr. VII (Naider) 61:13-15; DX 501, ¶ 38; DX 510. The License Agreement is presented to the user in a text box with a scroll bar. DX 510. This is the standard way in which license agreements are incorporated into software installations. Tr. VII (Naider) 57:1-17. 25. The License Agreement explains that the software generates contextually relevant advertisements and coupons, utilizing “pop-up” ,arid various other formats. Tr. VII (Naider) 57:19-58:23; DX 510. It also explains that WhenU reserves the right to update or upgrade the software at *740 its discretion. 5 Tr. VII (Naider) 60:5-10; DX 510. The software cannot be installed unless the consumer affirmatively accepts the terms of the License Agreement. Tr. VII (Naider) 60:13-15. 26. Plaintiffs’ computer expert Benjamin Edelman testified about how he obtained WhenU’s software via a software download prompt screen at a website called Lyricsdownload.com, using the phrase “drive by download” to describe this method of distribution. On cross-examination, Mr. Edelman conceded that the software download prompt screen offers the user the opportunity to read the License Agreement and tells the user that accepting the software is deemed to be an assent to its terms. Tr. VI (Edelman) 100:4-24. See also DX 538, ¶ 13; DX 539. Mr. Edelman also conceded that the distribution of software via a download prompt screen is a common practice, and is used by a variety of distributors for a variety of different purposes. Tr. VI (Edelman) 98:25-99:9. 6 Uninstalling WhenU’s Software 27. Consumers can uninstall WhenU’s software from their computers if they no longer wish to have it. Tr. IX (Reinhold) 33:5-17; Tr. VII (Naider) 52:5-7. Once uninstalled, the software will cease to operate or show advertisements or coupons on the consumer’s computer. DX 501, ¶ 41. 28. When a user removes or “uninstalls” a program bundled with Save, the Save software is automatically uninstalled along with it. Tr. VII (Naider) 33:21-34:2. The Save software supports the associated program and cannot be uninstalled without also uninstalling that associated program. Tr. VII (Naider) 49:15-50:16. SaveNow can be uninstalled separately from any freeware program with which it was downloaded. The World Wide Web 29. Since 1996, millions of computer users have become regular users of the Internet and the World Wide Web. Tr. IV (Edelman) 111-12. The Internet is a network of millions of interconnected computers, through which text, images and sounds are transported and displayed by an application called the World Wide Web (the “Web”). Tr. IV (Edelman) 106-07. 30. Much of the information on the Web is stored in the form of “web pages,” which can be accessed through a computer connected to the Internet (available through commercial Internet service providers or “ISPs”), and viewed on a PC user’s computer screen using a computer program called a “browser,” such as Microsoft Internet Explorer or Netscape Navigator. Tr. IV (Edelman) 107-09. 31. Web pages are only perceivable by the user by viewing them on the computer screen through the use of a browser. Tr. IV (Edelman) 109; Tr. VI (Naider) 28. 32. A web page is identified by its own unique Uniform Resource Locator (“URL”) (e.g., http://urww.wellsfargo.com), which ordinarily incorporates its name or *741 the trademark of the website operator 0e.g., Wells Fargo). Tr. I (Tazelaar) 165. 33. “Websites” are locations on the World Wide Web containing a collection of web pages, much like pages of a book. Tr. IV (Edelman) 108. The 3-dimensional metaphor of the computer screen as a “desktop” is often used to discuss the display of text and images on a user’s computer screen. Tr. IV (Edelman) 116-17. 34. The computer screen is composed of a series of picture elements (called “pixels”). Tr. IV (Edelman) 117-19. 35. Pixels are arranged in a single layer of horizontal and vertical rows that form a grid on the computer screen. Tr. TV (Edelman) 117-18; Tr. IX (Reinhold) 44. The particular color of each individual pixel which, taken together, make up the image displayed on the 2-dimensional computer screen, is determined by instructions received from the underlying computer program. Tr. IV (Edelman) 117-19. 36. A series of events must transpire in order for a user to view a web page via an Internet browser on his or her computer screen. Tr. IV (Edelman) 119-20. 37. First, the remote server on which the computer code for a particular website is maintained sends the code to the user’s web browser. Tr. IV (Edelman) 120. 38. Second, the PC’s browser then reads the code to determine how each pixel that makes up the computer screen should illuminate in order to create the specific on-screen display for that particular website. Tr. IV (Edelman) 120. 39. Third, the PC’s browser then conveys specific instructions to the Windows operating system, which, in turn, will send these instructions to the PC’s video card. Tr. IV (Edelman) 120. These instructions are stored in the video memory frame buffer portion of the PC’s video card. Tr. IV (Edelman) 121-22. 40. Finally, the video card, thereupon, causes each pixel on the computer screen to illuminate so as to create the specific 2-dimensional on-screen display of the website. Tr. IV (Edelman) 120. The Windows Environment in Which WhenU Operates 41. WhenU’s software is designed to operate within the Windows computer operating system, popularized by Microsoft. DX 501, ¶ 11. Windows is the most widely distributed software application ever written, currently in use by roughly 95% of computer users. Tr. IX (Reinhold) 36:13— 14. 42. In the Windows operating system, the computer “desktop” functions as a multitasking environment in which numerous software “applications,” such as spreadsheets, word processing programs, Internet browsing software, e-mail software and instant messaging software, may all run simultaneously. DX 523, ¶¶ 28-36; DX 501, ¶ 12. 43. This graphical computer “desktop” was intentionally designed to represent what a user would experience when using an actual physical desktop, with virtual replicas of file folders, text, files, spreadsheets, calendars, rolodexes, and so on. Tr. VIII (Reinhold) 127:22-128:7. The computer desktop thus gives the user the impression of operating in a three dimensional space in which items can be moved on top of and underneath each other. Tr. VIII (Reinhold) 128:8-129:9. Accordingly, while a display on a computer screen is literally two dimensional, it can properly be viewed as a three dimensional presentation as a matter of user perception. Tr. VIII (Reinhold) 127:15-20. 44. When a user opens a software application, it is launched in what is known as a “window,” a box on the user’s desktop within which all of the functions of that application are displayed and operate. Tr. *742 IV (Edelman) 110:3-13; DX 501, ¶ 13. An application is simultaneously represented by a button on the “task bar,” the strip that typically runs along the bottom of the desktop. DX 501, ¶ 13. A window can be enlarged or “maximized” to fill the entire computer screen, or reduced to take up a smaller area. Tr. TV (Edelman) 112:9-19; DX 501, ¶ 13. 45. Over the past two decades, operating a computer by manipulating overlapping windows has become a familiar process to personal computer users everywhere. Tr. IX 16:1-11; DX 523, ¶35. Connecting to the Internet in the Windows Environment 46. To access the Internet in the Windows environment, a user must establish a connection to the Internet, either over a telephone modem or some other form of Internet connection. Once such a connection has been established, the user will typically launch a software application known as a “browser,” such as Internet Explorer or Netscape. Tr. IV (Edelman) 108:16-22; DX 501, ¶ 15. A user can have multiple browser windows open simultaneously. Tr. I (Stapp) 116:20-117:8. 47. When launched by the user, the browser, like any other Windows-based software application, opens in a new window. Tr. IV (Edelman) 112:11-12. Within this window, the user interacts with the browser to access various websites on the Internet. Tr. IV (Edelman) 107:18-108:22; DX 501, ¶ 15. 48. The user can directly access data contained in a particular website by entering the website’s address (or “URL”) into the address box in any such open browser window. Alternatively, the user may search for websites of interest by utilizing a search engine, such as Yahool or Google, and then access those websites by clicking on the resulting links displayed as listings. A user may also reach a particular website by clicking on a “hyperlink” embedded in the text or graphics of a webpage. Tr. VIII (Reinhold) 129:10-24; DX 523, ¶21. 49. When a user attempts to access a webpage, the server hosting the webpage sends information in the form of a file back to the user’s browser program. Tr. VIII (Reinhold) 130:8-15. That file consists simply of lines of text written in Hypertext Markup Language or “HTML.” Tr. VIII (Reinhold) 131:6-13, 132:5-25; DX 523, ¶¶16, 22; DX 570. 50. The browser then interprets the HTML code file, and taking that information in conjunction with the user’s own browser settings, requests that the Microsoft Windows operating system open a window to display the webpage. The operating system, in turn, takes all that information from the browser, combines it with information concerning the user’s hardware configuration and the competing claims of other software programs, and then displays content in a window. Tr. VIII 130:16-23; DX 523, ¶¶ 37-40. 51. The HTML code identifies various elements that help determine how the webpage will ultimately be rendered on a computer user’s screen, but it does not provide an actual pixel-by-pixel mapping for rendering the webpage. Tr. VIII (Reinhold) 131:25-132:4; Tr. IX (Reinhold) 4:10-14. A wide variety of other factors will have a significant effect on the ultimate appearance of the webpage on the user’s screen, including the user’s ability to customize the browser’s settings. Tr. IX (Reinhold) 7:13-8:14; Tr. VI (Edelman) 58:8-21. Windows Permits the Display of Multiple Websites in Multiple Ways 52. The Windows environment permits a user to have multiple browser windows open simultaneously, each displaying a different webpage. Tr. I (Stapp) 116:20-117:8. As with other applications open on a user’s desktop, each separate browser window can be opened or closed, mini *743 mized or maximized, and moved around the screen. Tr. VI (Edelman) 62:1-13. The user can select which window appears in front of which other windows at any given time, in much the same way as a person can re-order a stack of papers on his or her desk. DX 523, ¶ 36; Tr. VIII (Reinhold) 125:7-23,128:25-129:9; DX 501, ¶¶ 11-12. 53. There are many applications that a user can run while browsing the Internet that cause additional windows to appear automatically in front of an open browser window. For example, a user may be viewing a webpage (e.g., cnn.com), when the user’s electronic mail or instant-messaging software (e.g., Microsoft Outlook or AOL Instant Messenger) launches a message in front of the browser window that the user is viewing. Tr. VI (Edelman) 62:21-63:3; DX 523, ¶ 40; DX 501, ¶ 18; DX 505. 54. Users have ultimate control over what programs run on their computers, when they are run, and what they are commanded to do. The user owns and controls the computer and the computer display, including the pixels that generate that display. DX 523, ¶ 25; Tr. VI (Edel-man) 79:10-25. How WhenU Delivers Advertisements to Participating Consumers 55. Advertisements shown by WhenU software are set up by WhenU’s Advertising Operations team. Tr. VIII (Naider) 5:24-6:3. The Advertising Operations Team receives creative copy from an advertiser, places the ad on a WhenU server, then “maps” the advertisement using an ad set-up table. Tr. VIII (Naider) 6:4-15, 18:6-19. Each advertisement is assigned a name and a variety of parameters such as size, priority, and frequency. Tr. VIII (Naider) 6:4-15. 56. The Advertising Operations Team “maps” the ad by determining the various categories in the Directory (such as “Air Travel”) and keyword algorithms that will trigger the appearance of the advertisement, subject to priority and frequency limitations. Tr. VIII (Naider) 7:12-10:8. 57. When the Advertisement Operations Team is done, the data is automatically recorded into the proprietary WhenU Directory (the “Directory”). Tr. VIII (Naider) 17:21-18:14. The Directory is delivered to and saved on the consumer’s desktop when the consumer installs the software, and optimized and updated on a daily basis. DX 501, ¶ 28. 58. As of July 1, 2003, the Directory contained approximately 32,000 URLs and URL fragments, 29,000 search terms and 1,200 keyword algorithms. Tr. VII (Naider) 98:10-99:6. The Directory categorizes these elements into various categories in much the same way as a local Yellow Pages indexes businesses into categories. These categories are the “heart” of WhenU’s system for delivering advertisements. Tr. VIII (Naider) 8:19-9:4, 27:25-28:5; DX501, ¶28. 59. As a participating consumer browses the Internet, the SaveNow software studies the user’s browsing activity and compares it against the elements contained in the Directory. Simultaneously, the Sa-veNow software determines whether: (a) any of those elements are associated with a category in the Directory, and (b) whether those categories are associated with particular advertisements. If the software finds a match, it identifies the associated product or service category, determines whether appropriate ads are available to be displayed, and, if so, selects an ad based on the system’s priority rules, subject to internal frequency limitations. Tr. VIII (Naider) 13:21-14:25; DX 501, ¶¶ 30-31. 60. Web addresses and search terms are included in the WhenU Directory solely as an indicator of a consumer’s interest. DX 501, ¶ 32. For example, the www.wellsfargo.com web address is includ *744 ed in the “finanee.mortgage” category of the WhenU Directory in order to identify consumers who are potentially interested in mortgages. Thus, if a consumer were to enter into the address box in an open browser window or conduct a search using a search engine by typing in the words “Wells Fargo,” SaveNow would detect that activity and scan the proprietary directory for a match to a WhenU category such as “finanee.mortgage.” DX 501, ¶ 31. 61. The SaveNow software might also determine that the consumer is interested in a particular category of products or services if it found certain combinations of words (“keyword algorithms”) in the content of the webpage visited. For example, if a participating consumer accessed a webpage that contained two occurrences of the word “buying,” two occurrences of the word “home” and four occurrences of the word “mortgage,” the SaveNow software might determine that the consumer was interested in the “finanee.mortgage” category. Tr. VIII (Naider) 9:5-20; DX 501, ¶33. 7 62. Under WhenU’s category system, any given ad will ultimately be mapped to scores of discrete elements (ie., URLs, search terms, key word algorithms) that are related topically. Tr. VIII (Naider) 7:21-9:4. Thus, WhenU advertisements do not specifically target individual websites such as Wells Fargo and Quicken Loans. For example, Mr. Edelman implied in his original declaration that an advertisement for “GetSmart” was specifically targeted at the Quicken Loans homepage. PX 109, ¶ 26. In fact, the complete mapping of that ad in the Directory reveals that it is mapped to 13 separate categories, each of which represents many URLs and search terms. DX 569; Tr. VIII (Naider) 19:7-19. The URL for the Quicken Loans homepage is only one of over hundreds, if not thousands, of URLs that could trigger this ad via the category system. Tr. VIII (Naider) 21:2-11. 8 WhenU Sells Advertising On a Category Basis 63.WhenU sells advertising to advertisers on the basis of sales categories, which are grouped into certain product and service categories. Tr. VIII (Naider) 4:12-23, 63:21-64:5. These sales catego *745 ries are broader than the categories used for mapping advertisements. Tr. VIII (Naider) 7:21-8:6. Although the sales categories are made public for marketing purposes, the categories used to map ads are known only to the WhenU Advertising Operations Team, and are not disclosed or in any way promoted to WhenU’s advertisers, in-house sales team, or independent sales agents. Tr. (Naider) 34:13-18. 64. WhenU does not guarantee advertisers that their advertisements will appear when participating consumers access content from a particular website. WhenU guarantees only that advertisements will be shown to consumers who appear interested in a particular product or service sales category. Tr. VIII (Naider) 34:1-18; DX 501, ¶ 35. 65. WhenU does not allow its own advertisers to be excluded from any category as a condition of purchasing advertising from WhenU. Tr. VII (Naider) 45:25-46:11. Thus, WhenU can and does show ads for its advertisers’ competitors — ads that may well appear when a user has accessed the advertiser’s website. Id. 66. In sum, WhenU does not target specific websites either in its software or in selling its services to advertisers. Rather, WhenU’s advertisements are displayed according to the product category in which the consumer is interested and limited by factors such as the number of advertisements the consumer has already seen. Thus, it is the user’s actions on his or her desktop that ultimately determine whether that consumer will see a particular advertisement. DX 501, ¶¶ 31, 36, 46. WhenU Ads Are Displayed in Separate, Conspicuously Branded Windows, and Specifically Advise Participating Consumers That They Are From WhenU, and Are Not Sponsored by Any Website the User May Be Viewing 67. The advertisements and coupons that SaveNow delivers to a participating consumer’s desktop appear in a window (the “WhenU Window”) which is separate and distinct from any other window already open on the desktop. Tr, VI (Edel-man) 63:4-16; Tr. VII (Naider) 46:19-47:11; DX 523, ¶ 49; DX 501, ¶ 43. 68. SaveNow advertisements take various formats, such as: (1) a small format “pop-up” window that typically appears flush to the bottom right-hand corner of the consumer’s desktop; (2) a larger “pop-under” window that appears behind some or all of the browser windows that the consumer is viewing; (3) a horizontal “panoramic” window that runs along the bottom of the user’s computer screen. Regardless of the format used, the WhenU Window is a distinct, separate window unique to the SaveNow application and represented by its own button on the user’s task bar. Tr. VII (Naider) 43:16-45:20; DX 501, ¶ 44. 69. Many SaveNow advertisements— approximately 50% of the total — are pop-under ads. Tr. VII (Naider) 45:16-20. WhenU’s pop-under ads are designed not to be displayed to the user until after the user closes or minimizes the browser window containing the webpage that the user was viewing when the ad was triggered by the SaveNow software. Tr. VII (Naider) 44:15-23, 45:10-12. Thus, unless manipulated by the user, a SaveNow pop-under ad triggered by a Wells Fargo or Quicken Loans URL will not be displayed on the user’s screen at the same time as the webpage with that URL. DX 501, ¶ 49; Tr. VI (Edelman) 68:12-16; Tr. VII (Naider) 44:11-23. 70. SaveNow pop-up ads appear as a small box in the bottom right hand corner of the user’s computer screen. Tr. VII (Naider) 43:20-44:1; DX 501, ¶45. Depending on the site visited, the browser size, and the user’s screen resolution configuration, a SaveNow pop-up ad may or *746 may not appear in front of content in the underlying website. Tr. VI (Edelman) 65:15-24. 71. If the underlying webpage (or any other underlying windows, such as a Word document) is clicked on after the pop-up format advertisement is displayed, the pop-up will no longer appear at the front of the screen, although it will still be present in the user’s task bar at the bottom of the screen. DX 501, ¶ 54. 9 72. At any time a user may make any form of SaveNow advertisement disappear permanently by clicking on the “X” box at the top right hand corner of the ad. Tr. VII (Naider) 43:9-15. The “X” box is a standard feature of the Windows operating system and Internet users are generally familiar with its function. Tr. VII (Naider) 43:14-15; DX 501, ¶ 53; Tr. IX (Reinhold) 16:14-23. 73. The WhenU Window is labeled as such. SaveNow ads display a green symbol in the corner of the window and the SaveNow designation. Tr. VII (Naider) 36:17-19. Save ads contain a bull’s eye and the “SAVE I” designation. Tr. VII (Naider) 46:14-18. 10 74. All SaveNow advertisements contain a notice stating: “This is a WhenU offer and is not sponsored or displayed by the website you are visitiny. More.... ” Tr. VII (Naider)' 36:19-23. When the consumer clicks on the word “More..a dialog box opens that contains information about SaveNow and a direct link to the “Frequently Asked Questions” page of WhenU’s website. Tr. VII (Naider) 38:2-39:7; DX 501, ¶ 48. 75. SaveNow advertisements also contain a “?” symbol in the corner, which is also a typical feature of a window in the Microsoft Windows operating system. Clicking on the “?” symbol opens the same window as when consumers click on the “More...” link described above. Tr. VII (Naider) 37:22-38:10; DX 501, ¶ 48. 76. SaveNow does not automatically cause any advertiser’s webpage to be displayed on a user’s desktop. After the WhenU advertisement is displayed, the user can elect to access the advertiser’s website by clicking on the WhenU Window. Tr. VII (Naider) 39:10-20. The user also can elect not to access the advertiser’s webpage, and can easily close the WhenU window or minimize it for later viewing. Tr. VII (Naider) 42:24^43:15; DX 501, ¶¶ 52-54. 77. A user who clicks on a SaveNow advertisement is taken to the advertiser’s webpage. Tr. VII (Naider) 39:10-20; DX 501, ¶ 52. A consumer who has accessed the advertiser’s webpage can return to the webpage that was previously on the user’s screen by clicking on the Internet browser’s “Back” button. Tr. V (Edelman) 12:16-21; Tr. VII (Naider) 39:21-40:6; DX 501, ¶ 52. This function will work even if the user is accessing a secure webpage at the time the advertisement is clicked. Tr. IX (Reinhold) 27:10-28:16. WhenU Advertisements Do Not Use Plaintiffs’ Trademarks 78. WhenU’s advertisements do not use the words “WELLS FARGO,” “WELLS FARGO ONLINE,” “QUICKEN LOANS,” or any other trademark reg *747 istered to plaintiffs, in the advertisements themselves. DX 501, ¶ 32. 79. As discussed above, URLs are included in the Directory only to identify the website itself for the purpose of determining the interests of participating consumers. DX 501, ¶ 32. WhenU does not use any of plaintiffs’ trademarks to identify goods or services, to indicate any sponsorship or affiliation with the goods or services advertised by WhenU, or to identify the source or origin of any goods or services advertised by WhenU. 80. The use of keyword terms in connection with the delivery of advertisements is a common practice on the Internet, and is a source of revenue for search engines such as Google, and other Internet companies. DX 523, ¶¶ 54-56. WhenU Advertisements Do Not Appear “On” Plaintiffs’ Websites 81. Plaintiffs take the position that Sa-veNow advertisements appear “on” the plaintiffs’ websites. Such usage misconstrues the technical reality of the Internet. 82. A “website” is a series of related webpages, whose program code is located in separate, distinct servers controlled by the owners of the website. DX 523, ¶ 43; Tr. IV (Edelman) 108:8-11. When a user accesses a website, the website is not transferred to the user’s desktop. All that exists on the computer screen is an image generated by the user’s web browser based on instructions received from the text-based HTML code of the webpage. Tr. VIII (Reinhold) 130:8-23, 132:5-24; DX 523, ¶ 44. At that point, there is no connection between the individual user’s computer and the website. Tr. IX (Reinhold) 6:3-10. 83. Having caused the computer to display a page from one website, the user may request the computer to simultaneously display a page from a different website. The user will then have pages from two different websites on his desktop. If the user has called the second website page up in a new browser window, it will appear in front of the page from the first website and the only indication that the first website has been accessed will be the button on the user’s task bar. However, as discussed above, the user can easily size his browser windows to cause pages from both websites to appear on his screen, in windows that overlap or not, as the user chooses. Nothing that an individual does in the window displaying one website interferes with data transmitted to or from another website. Likewise, nothing that an individual does in a window displaying a different application (e.g., a word processing or instant messenger program) interferes with data transmitted to or from a website which the user has accessed. A website resides on its own servers where it is protected from tampering by “firewalls” and other Internet security procedures. DX 523, ¶¶ 42-43. 84. SaveNow interacts only with the web servers of WhenU or WhenU’s advertisers. Tr. VI (Edelman) 77:11-78:4; Tr. VII (Naider) 99:7-12. It has nothing to do with plaintiffs’ web servers. Tr. VIII (Naider) 13:21-14:25. As plaintiffs’ own expert acknowledged, SaveNow cannot and does not access the servers where plaintiffs’ websites are physically located. Tr. VI (Edelman) 77:11-78:4. 85. Defendant’s SaveNow software has no physical relationship to any other software application that may be open on a user’s desktop, including, but not limited to, windows in which images associated with plaintiffs’ websites are displayed. The fact that a window containing a Save-Now ad and the windows in which plaintiffs’ websites may be displayed may be visually stacked on top of each other on the user’s screen is purely a function of a computer’s graphical interface which is designed to make a computer “desktop” look *748 and act like a real desk. Tr. IX (Reinhold) 15:15-21,16:1-17:5; DX 523, ¶¶ 41-47. WhenU Advertisements Do Not “Modify” Plaintiffs’ Websites 86. SaveNow does not transmit, display or reproduce images of plaintiffs’ websites. Tr. VII (Naider) 82:4-5. All SaveNow does is display an image of an advertisement which contains a link to a site designated by the advertiser. The participating consumer may access that site, if the consumer chooses, by clicking on the link embedded in the advertisement. Tr. VII (Naider) 39:10-20. 87. Plaintiffs contend that because the appearance of a SaveNow advertisement alters the current content of video memory, plaintiffs’ webpages are “modified” whenever a SaveNow ad appears while a user is also displaying one of their webpag-es. The Court rejects this contention. 88. Once a computer browser renders a webpage on a window, a copy of the HTML code file associated with that web-page is saved into the computer’s general random access memory or “RAM.” Tr. IX (Reinhold) 11:17-22. The RAM copy of the HTML file is used to help the computer instantly redisplay the webpage image in the event that some other window has subsequently obscured all or part of the webpage image. Tr. IX (Reinhold) 12:7-13:5. The appearance of a SaveNow ad does not interfere with the storage of another webpage’s HTML code in RAM memory or erase another webpage’s HTML code. Tr. VI (Edelman) 76:15-77:8, 80:22-81:3; Tr. IX (Reinhold) 15:15-21, 27:3-9. 89. In addition to ordinary RAM memory, a computer maintains a temporary form of memory called video memory that forms part of the computer’s display system. Tr. IX (Reinhold) 13:6-12. The video memory simply contains a pixel-by-pixel “snapshot” of whatever happens to be displayed on a computer screen at any given instant. Tr. IX (Reinhold) 13:17-25. 90. Plaintiffs contend that because the appearance of a SaveNow advertisement alters the current content of video memory, plaintiffs’ webpages are “modified” every time a SaveNow ad appears while a user is also displaying one of their webpag-es. However, these pixels are part of the user’s physical computer, and are not part of any webpage that the user might happen to be viewing at the time. Tr. VI (Edelman) 79:10-25. 91. Further, because the pixel display is the means by which any image on a computer screen is generated, whenever the display changes, there is a corresponding change in the content of video memory. Tr. IX (Reinhold) 14:1-10. Video memory is modified when a user opens a new application, receives an instant message, or uses his mouse to move the cursor across the screen. Tr. VI (Edelman) 78:5-79:9; Tr. IX (Reinhold) 14:21-15:9. As long as a user is actively using the computer, the content of video memory is altered and updated every l/70th of a second. Tr. IX (Reinhold) 14:21-15:9. Accordingly, the alteration of video memory is an ephemeral occurrence, and does not constitute a modification of the plaintiffs’ webpages. WhenU Advertisements Do Not “Frame” Plaintiffs’ Websites 92. The Court rejects the contention that SaveNow “frames” plaintiffs’ websites. 93. Framing occurs when one webpage displays the content of another webpage within its own borders. If the outer window is moved, the framed page moves with it simultaneously; if the outer window is closed or minimized, the framed page closes or minimizes as well. Tr. IX (Reinhold) 29:1-17. The purpose of framing is to create a single seamless presentation that integrates the content of the two web- *749 pages into what appears to be single web-page. Tr. IX (Reinhold) 29:1-31:3. 94. SaveNow ads appear in entirely separate windows that can be moved independently without moving any other web-page, and can be closed without closing or in any way affecting any other webpage. Tr. IX (Reinhold) 29:18-30:21. Hence, Sa-veNow ads do not “frame” and are not “framed by” any other window, such as the window in which one of plaintiffs’ webpag-es might be displayed. WhenU Protects the Privacy and Security of Its Users 95. WhenU collects only the information necessary to run its system and to be able to compensate its partners and invoice its advertisers. Tr. VII (Naider) 76:4-11. This information consists of the URL, search term or keyword of the webpages that triggered the delivery of a WhenU advertisement, and whether the user clicked on the advertisement. Tr. VII (Naider) 75:19-76:3. This information is collected by WhenU on an aggregate basis and is not associated with an individual user or individual profile. Tr. VII (Naider) 76:17-77:20. WhenU does not collect a user’s “click stream data,” i.e., information concerning the history of webpages visited by the user. Tr. VII (Naider) 76:12-16, 83:18-84:23. Nor does "WhenU use cookies to track the activities of SaveNow users. DX 547; Tr. VII (Naider) 87:3-15; DX 547. 96. Like most entities that operate on the Internet, including Wells Fargo and Quicken Loans, WhenU uses the IP address of its users. The sole purpose of using the IP address is so that SaveNow can use the Internet to send ad display information to the WhenU servers. Tr. VII (Naider) 79:15-19. WhenU does not use IP addresses to identify individual users. Tr. VII (Naider) 80:21-28. 97. The Court additionally finds no support in the record for plaintiffs’ contention that WhenU somehow specifically targets “secure” webpages. Tr. VII (Naider) 102:1-18. Although it is possible to view a SaveNow advertisement while accessing a secure webpage, this does not disrupt the security of that webpage. A consumer who clicks on a SaveNow ad can easily return to the secure webpage that was previously on the user’s screen by clicking on the Internet browser’s “Back” button. Tr. IX (Reinhold) 27:10-28:16. It is also a standard feature of the Microsoft Windows operating system to provide a warning to users if they do something that would cause them to leave a secured webpage, giving the user the opportunity to cancel that decision. Tr. IX (Reinhold) 27:22-25; 28:10-16. Plaintiffs Have Failed To Submit Competent Evidence of Likelihood of Confusion from SaveNow Advertisements 98. Plaintiffs assert that a WhenU ad displayed on a computer screen at the same time as a consumer is viewing a page from one of their websites is inherently confusing to the consumer who, according to plaintiffs, believes it emanates from the plaintiffs’ websites. Memorandum in Support of Plaintiffs’ Motion for Preliminary Injunction (“Plaintiffs’ P.I. Mem.”), pp. 6-7. However, plaintiffs have not presented any evidence of actual consumer confusion. 99. By contrast, there is good reason to believe that the typical SaveNow user would not perceive a WhenU advertisement as sponsored by or affiliated with the plaintiffs’ websites. First, SaveNow users are accustomed to receiving offers from WhenU while surfing the Web. DX 501, ¶ 46. For example, a SaveNow user shopping for a financial services online would be exposed to SaveNow ads for obvious competitors such as Ameriquest Mortgage (PX 125), Brown & Company (PX 130), and LowerMyBills.com (PX 135). It is therefore unlikely that they would sudden *750 ly think that a SaveNow Ameriquest Mortgage ad comes from Wells Fargo or Quicken Loans. DX 523, ¶ 50. Second, SaveNow ads are identified by WhenU, and bear a prominent notice and disclaimer stating that they come from ‘WhenU” and are “not sponsored or displayed by the website you are visiting.” 11 Tr. IX (Reinhold) 25:5-20; 45:22-46:1. Third, SaveNow ads appear in a distinct window, bear all of the indicia of a distinct software application, and do not relate in any way to any other window on the user’s screen. Tr. IX (Reinhold) 16:12-17:5, 29:18-30:21. Internet users understand that different software applications run in different windows. Tr. IX (Reinhold) 16:1-11. 100. Plaintiffs argue that their customers are especially likely to be confused because they are “not particularly sophisticated.” Plaintiffs’ P.I. Mem., p. 14. However, plaintiffs have not supported that contention with evidence. To the contrary, Mr. Neal acknowledged that people who decide to obtain a mortgage online may be more knowledgeable about the Internet than ordinary users. Tr. Ill (Neal) 12:18-13:1. 101. In addition, consumers are likely to be attentive when attending to their financial affairs (Tr. II (Neal) 141:16— 142:4; Tr. TV (Jacoby) 21:7-19), and are especially attentive when obtaining a mortgage which is “for most people the largest financial decision they make.” Tr. I (Stapp) 141:16-19. See also Tr. I (Stapp) 141:23-142:14 (decision to obtain a mortgage involves study and research); Tr. Ill (Neal) 12:18-13:1 (the decision to obtain a mortgage is more significant in the life of a consumer than the decision to buy contact lenses). 102. The only evidence of potential consumer confusion presented by plaintiffs was the testimony of William Neal, whom plaintiffs offered as an expert in “consumer surveys [and] marketing research.” Tr. II (Neal) 86:24-87:1. Mr. Neal’s testimony was based on surveys he conducted for the plaintiff in: (a) Washingtonpost. Newsweek Interactive Co. (‘Washington Post”) v. The Gator Corp. (the “Gator Survey”), an action not involving WhenU, and (b) 1-800 Contacts, Inc. (“1-800”) v. WhenU, a lawsuit pending in the Southern District of New York (the “1-800 Survey”). See Tr. II (Neal) 88:3-8; PX158; PX 159. These plaintiffs were also represented by the Gibson, Dunn & Crutcher attorneys who represent plaintiffs in this case. Tr. II (Neal) 130:3-22. 103. Plaintiffs did not explain why Mr. Neal did not prepare a new survey for this matter. The Gator Survey was conducted in early June 2002. Tr. II (Neal) 123:4-21. The 1-800 Survey, which was modeled on the Gator Survey, took all of 14 days to complete. 12 Research involving the Internet may become obsolete in a matter of months. Tr. IV (Edelman) 100:1-12. See also Tr. VIII (Reinhold), 121:3-9 (Internet changes very rapidly). Mr. Neal’s testimony that the surveys’ results would not differ substantially if the data were collected today is based on speculation. Tr. Ill (Neal) 21:4-7. Indeed, Mr. Neal admitted that consumer perceptions of Internet advertising might have changed since he conducted the surveys, and that the only way to know for sure would have been by test *751 ing, which he did not do. Tr. Ill (Neal) 133:6-134:3. 13 104. Mr. Neal’s methodology was se-vérely criticized by defendant’s witness, Dr. Jacob Jacoby, an extremely well qualified expert in consumer behavior and research methodology. 14 Mr. Neal Did Not Show the Survey Respondents Any WhenU Pop-Up Ads or Other Stimuli 105. One flaw in Mr. Neal’s methodology was his failure to show the respondents any demonstrative stimulus. Tr. II (Neal) 134:15-19; PX 158; PX 159. Mr. Neal did not show the 1-800 respondents an exemplar’ of a WhenU ad, nor did he ensure that the respondents had WhenU ads in mind— as opposed to pop-up ads generated by Gator, or a search engine or a commercial website. Mr. Neal did not do anything to find out whether his respondents were even familiar with SaveNow ads. Tr. II (Neal) 138:20-141:2. Indeed, Mr. Neal conceded that he did not know with any degree of scientific certainty whether any of the respondents had ever seen a Save-Now ad. Tr. II (Neal) 139:7-10. 106. In lieu of showing any actual ads, Mr. Neal provided a generic description of a pop-up ad at the beginning of the surveys. PX 158, Tab C; Tr. II (Neal) 136:8-10, 138:9-19. This description explained that pop-up advertisements (a) usually appear in the middle of the user’s screen, (b) partially block out the content of the underlying web page, (c) may automatically take the user to another web site, and (d) may not close when the user clicks the “x” in the upper right-hand corner. PX 158, Tab C & PX 159. The survey instructed respondents to use this description in answering the questions in the survey (PX 158, Ex. C, p. 31), and Mr. Neal conceded he intended respondents to think of pop-up ads in the manner that he described them for purposes of responding to the survey. Tr. Vol. II (Neal) 138:9-19. 107. As Mr. Neal conceded (Tr. II (Neal) 156:4-9), his definition does not describe the defendant’s advertisements. 15 *752 WhenU pop-up, or small format, advertisements appear in the bottom right-hand eorner of the screen (Tr. VI (Naider) 35:18-36:12) and do not necessarily block any webpage content. See, e.g., PX 128. More than half of WhenU advertisements are pop-under ads which do not appear on the user’s screen until after the browser window has been closed. 16 Tr. VII (Naider) 44:11-45:17. WhenU ads do not take the user to another website without an affirmative click by the user (Tr. VIII (Naider) 108:3-5) and can always be closed by a click on the “x”. Tr. VI (Naider) 35:8-17. Mr. Neal conceded that a consumer might feel differently about a pop-up ad that automatically takes a consumer to another website (as he described) versus an ad that does not. Tr. II (Neal) 158:7-12. 108. The Court also rejects Mr. Neal’s effort to determine whether respondents were confused about the source of the ads based on their general recollection of pop-up ads they might have seen. Tr. II (Neal) 134:15-19, 135:24-136:10. As Dr. Jacoby testified, recall is used where the material issue is what the consumer remembers. The recall technique is not used to test confusion. Tr. IV (Jacoby) 31:19-34:4. 109. The Court also rejects Mr. Neal’s testimony that the sheer number of different SaveNow ads made it impossible to conduct a survey using actual SaveNow advertisements. Tr. III (Neal) 25:22-23:10. Mr. Neal could have used a sampling methodology to test a representative sample of WhenU advertisements, using any one of a variety of accepted scientific techniques. Tr. IV (Jacoby) 28:8-31:3. Mr. Neal’s explanation is further undercut by his concession that he would likely have used SaveNow ads in the 1-800 Survey had they been available to him. See PX 191 at 6. 17 110. Because Mr. Neal did not show the survey respondents WhenU ads, there is no way for anyone to know whether the respondents had WhenU ads in mind when they answered his questions. As Dr. Jaco-by testified: “There’s absolutely no scientific defensible foundation for concluding any WhenU advertising caused any of the data .... you can’t connect the dots.... ” Tr. IV (Jacoby) 34:5-11. The Survey’s Results Cannot Be Applied in This Case Because the Surveys Tested Universes of Respondents That Are Not Inclusive of the Relevant Universe in This Case 111. The surveys also do not provide reliable evidence of confusion because they did not sample the relevant universe of people who obtain mortgages or conduct banking services online (i.e., people who use or are likely to use plaintiffs’ websites). The results of a survey of contact lens users or a survey of readers of online periodicals cannot be extrapolated with accuracy to people who use or are likely to use plaintiffs’ online financial services. 18 *753 Tr. IV (Jacoby) 26:11-18. This is so even if all these people share similar demographic data. Tr. IV (Jacoby) 18:1-19:6. Of the approximately 156 million people who use the Internet, only 28 million shop for mortgages online. Tr. IV (Jacoby) 22:11-15. Furthermore, only 12-13% of consumers wear contact lenses. Tr. IV (Jacoby) 25:21-26:3. Given these statistics, there is no way to conclude that the contact lens wearing 1-800 respondents were also persons interested in obtaining mortgages online. Tr. IV (Jacoby) 19:7-12, 20:24 — 23:10, 25:19-17, 26:11-18. 112. Moreover, people conducting banking transactions or purchasing mortgages are more likely to pay careful attention than people buying contact lenses or perusing a periodical. Tr. II (Neal) 141:16-142:4; Tr. IV (Jacoby) 21:7-19. A meaningful survey must take into account not only the respondents’ demographic characteristics, but also what they are doing at the time' because consumers make decisions of different import with different mindsets. Tr. IV (Jacoby) 20:24-21:19. 113. Mr. Neal’s opinion that he could extrapolate the results of his previous surveys to users of online financial services does not appear to be well founded. Indeed, Mr. Neal acknowledged that there was no empirical evidence that supports his conclusion that the Wells Fargo and Quicken Loans website user populations are the same as the populations he developed for the 1-800 and Gator Surveys. Tr. Ill (Neal) 14:12-18. Mr. Neal’s Survey Questions Were Biased 114. Mr. Neal’s surveys also are unreliable because he used leading questions that may have skewed the survey results. Tr. IV (Jacoby) 34:12-24. 115. For example, Mr. Neal improperly told respondents that pop up ads appear “on” a website, thus suggesting the association he was trying to establish. Moreover, Mr. Neal admitted that some of his questions were biased. For example, Mr. Neal admitted that Question 10 (“Do you believe that WhenU.com was honest in informing you about what SaveNow software did?”) was a “loaded question” because of the suggestive nature of the word “honest.” PX 191 at p. 30; Tr. II (Neal) 166:8-167:7. Mr. Neal further admitted that, when he prepared the questionnaire, he was aware that his use of the word “honest” violated the standard rules of framing survey questionnaires. Tr. II (Neal) 167:3-12. See also Tr. II (Neal) 165:20-25. Mr. Neal Did Not Rule Out Obvious Alternative Explanations and Made Unsupported Analytical Leaps 116. Mr. Neal’s analysis of the survey data was flawed in other respects. For example, Mr. Neal concluded that 60% of respondents believed “pop-up ads are placed on the website on which they appear by the owners of that web site” (PX 110, ¶ 6(e)), relying on Question 4-1, which required respondents to agree, disagree or state no opinion as to the statement: “I believe that pop-up advertisements are placed on the website on which they appear by the owners of the website.” PX 158, Tab C, p. 32; Tr. II (Neal) 168:3-5. If the respondent indicated that he or she agreed, Mr. Neal interpreted that to mean the respondent believed that pop-up ads are always placed “on the website ... by the owners of that website”. Tr. II (Neal) 167:14-169:24. Mr. Neal testified that respondents who believed that pop-up ads *754 were sometimes displayed by website owners and sometimes not displayed by the website owners would not have answered “Agree,” but rather said they had “no opinion.” Tr. II (Neal) 168:6-169:19. He also failed to consider obvious alternative explanations for his results. 117. Mr. Neal’s testimony is counter-intuitive. A respondent who believed that some pop-up ads are displayed by website owners clearly had an opinion with respect to Q4.1. Moreover, the only evidence related to Mr. Neal’s assumption that “Agree” meant “always agree” demonstrates that the assumption is wrong. In response to Question 4-4 of the Gator Survey (“I believe that ‘Pop-Up’ advertisements are sponsored by the website on which they appear”) more than 66% “Agreed.” Under Mr. Neal’s interpretation, this meant that 66% of the Gator respondents thought that pop-up ads are always sponsored by “the website on which they appear.” However, in the next question of that survey, Mr. Neal posed the opposite question, “I believe that ‘Pop-up’ advertisements are sometimes not sponsored by or authorized by the website on which they appear.” If, as Mr. Neal believes, the 66% who agreed with Question 4-4 believed that pop-up ads are always “placed on the website ... by the owners of that website,” then a maximum of 34% of these respondents could possibly have agreed with Question 4-5, that pop-up ads are only sometimes not sponsored by the website owner. Tr. II (Neal) 172:9-173:14. However, 47% of respondents agreed with Q 4-5. This data indicates an error in Mr. Neal’s reasoning. 118. By failing to consider and account for these and other alternative explanations for his results, Mr. Neal violated basic standards of scientific practice and rendered his results unreliable. Tr. IV (Ja-coby) 42:1-43:4. The Survey Was Not Properly Administered, Did Not Contain Control Questions to Generate an Error Rate, and Employed a Design that Rendered the Results Uninterpretable 119. Mr. Neal failed to employ an experimental design that established causation. As Dr. Jacoby testified, unless a control group is used to account for the effects of “noise,” i.e., extrinsic factors such as pre-existing beliefs other than the stimulus at issue that could contribute to a survey’s results, the survey’s results are uninterpretable. Mr. Neal acknowledged that he did not use a true control group (Tr. II (Neal) 92:17-21) and that, as a result, his conclusions only had a 51% certainty level. Tr. Ill (Neal) 24:6-25:8. While Mr. Neal asserted that a certainty level of 51% is all that can be achieved in the social sciences (Tr. Ill (Neal) 23:13-24), his testimony is contradicted by Dr. Jacoby, who testified that the certainty level required in the social sciences is at least 95% (Tr. TV (Jacoby) 52:25-54:14), and who provided an example of the kind of survey design that could have been employed to establish causation. Tr. IV (Ja-coby) 52:25-54:14. 120. Other problems with Mr. Neal’s survey include: (a) the use of an Internet panel (Tr. IV (Jacoby) 45:13-47:7); (b) the failure to use standard procedures to avoid yea saying (Tr. IV (Jacoby) 36:24-39:20); (e) the use of compound questions (Tr. IV (Jacoby) 37:5-6); (d) the failure to correct for possible error through the use of control or filter questions (for example, by testing response rates to questions about fictional computer programs) (Tr. IV (Ja-coby) 43:8-44:21); and (e) the failure to provide an independent check on respondents’ understanding of the questions through the use of an in-person interviewer or administrator. Tr. IV (Jacoby) 43:25-44:3, 45:17-46:4. *755 Plaintiffs Have Not Shown Irreparable Harm 121. Plaintiffs assert that WhenU advertisements cause them irreparable harm by imposing non-compensable reputational injury on their marks. However, plaintiffs have failed to come forward with concrete evidence of even a single customer or potential customer who failed to purchase products or services from them because of WhenU. 122. Quicken Loans knew everything it needed to know to commence an action against WhenU by November 2002, and Wells Fargo knew everything it needed to know by December 2002. Indeed, Wells Fargo had prepared screen shots of Save-Now for use in litigation as early as November 2002 (Complaint, ¶ 2); Quicken Loans prepared similar screen shots as early as September 2002. PX 136. Nonetheless, this motion was not filed until May 20, 2003. The dilatory behavior of the plaintiffs in prosecuting their claims, and their strategic decision to defer a trademark case while they fulfilled the jurisdictional requirements for a copyright claim, are inconsistent with a finding that WhenU’s ads are causing the plaintiffs’ irreparable injury. WhenU Does Not Link to Plaintiffs’ Websites 123. Plaintiffs further contend that WhenU’s advertising injures them because it puts them at risk with their regulators. Specifically, plaintiffs have proffered certain federal banking regulations relating to weblinking, and suggested that these regulations apply to WhenU’s advertisements. However, plaintiffs have offered no competent testimony to that effect and a simple reading of the alleged regulations on which plaintiffs rely shows that they are concerned only with federal banks which employ “weblinking” with various websites and with which they have “joint marketing relationships.” See, e.g., Electronic Activities, 67 F. Reg. 34, 992 at 35,002 (May 17, 2002). 124. The SaveNow software does not link to any website other than the websites of WhenU and WhenU’s advertisers and WhenU has no co-branding relationships with any banks. Tr. VIII (Naider) 107:19-108:2. No regulatory agency has ever approached WhenU to express concerns about the effects of WhenU software on the regulation of banks or financial services companies. Tr. VIII (Naider) 108:6-13. Indeed, plaintiffs have failed to proffer any evidence that federal regulators have ever inquired about or expressed concerns about WhenU or WhenU advertising. Plaintiff Quicken Loans Has Actually Benefitted from SaveNow 125. Although plaintiffs both assert having been harmed by SaveNow, Quicken Loans has actually benefitted from WhenU advertising. 126. The Quicken Loans website is part of the overall Quicken.com website. The Quicken.com website is owned by Intuit, which owns TurboTax and other entities. DX 554; Tr. I (Stapp) 76:22-25, 109:5-24. Quicken Loans relies heavily on Intuit’s “Quicken” brand to attract its customers. Tr. I (Stapp) 89:3-6. 127. As part of the Quicken.com website, Quicken Loans benefits from any increase in the traffic to the Quicken.com home page or any other component of the Quicken.com website, including Turbo Tax. Tr. I (Stapp) 135:2-10. In recognition of this synergy, Quicken Loans has conducted joint marketing campaigns with Turbo Tax and other Intuit entities to drive traffic to the Quicken website. DX 557; Tr. I (Stapp) 134:5-13. 128. Turbo Tax uses WhenU to advertise its products. DX 558-560; Tr. I (Stapp) 137:2-5; Tr. VII (Naider) 92:2-93:12. WhenU’s Turbo Tax advertising *756 campaign has been very effective in bringing visitors to the joint Turbo Tax/Quicken website. DX 559. Accordingly, the Court finds that Quicken Loans has benefitted from the use of WhenU software by Turbo Tax to drive traffic to the Quicken family of entities. Tr. I (Stapp) 135:11-15. 129. In contrast, the only harm alleged by Quicken Loans stems from the possibility that a recipient of a SaveNow ad may conclude the SaveNow advertiser has a superior offer and thereby be “diverted” from the Quicken Loans webpage. Quicken Loans has not demonstrated that even a single customer, who would have otherwise purchased services from Quicken Loans, did not do so because of “diversion” by a WhenU advertisement. Indeed, Mr. Stapp admitted that only 3-5 percent of people who access the Quicken Loans site even bother to fill out an application form, much less procure services from Quicken Loans. Tr. I (Stapp) 84:12-16. An Injunction Will Harm WhenU and the Public 130. Entry of a preliminary injunction would seriously harm WhenU. A court opinion casting doubt on the legality of WhenU’s core business model would result in the loss of many of WhenU’s largest advertisers, costing WhenU millions of dollars in lost revenue. Tr. VIII (Naider) 36:23-37:9. Because of the long planning cycles involved in the advertising business, this damage could not be recouped even if WhenU eventually prevailed on the merits. Tr. VIII (Naider) 37:10-19. 131. A number of advertisers, including American Express, Bank of America and General Motors, have already discontinued their campaigns with WhenU out of concern that further use of WhenU’s advertising services will embroil them in litigation. In the opinion of WhenU’s CEO, WhenU would lose key advertisers should this Court issue a preliminary injunction. DX 501, ¶ 58; Tr. VIII (Naider) 36:25-37:9. 132. WhenU employs the services of 50 individuals to maintain its operations, and relies on the efforts of some 60 or 70 independent sales representatives, many of whom derive most of their revenue from sales generated on behalf of WhenU. Tr. VII (Naider) 32:2-16. WhenU’s business success depends heavily on attracting and recruiting talented personnel. The effect of an injunction and the associated financial losses would be to prevent WhenU from recruiting talented people and to increase the likelihood that present employees would leave the company. Tr. VIII (Naider) 37:20-38:5. 133. The issuance of a preliminary injunction would have an adverse effect on WhenU’s ability and incentive to improve its contextual advertising technology to deliver more specific real-time advertising. WhenU is a start-up company, and its technology is constantly evolving. Tr. VII (Naider) 28:18-19. 134. Harm to WhenU would harm the public as well. WhenU benefits participating consumers by improving access to relevant, useful and money-saving information about products and services that interest them. WhenU’s advertisements increase the choices available to consumers and thereby promote competition. DX 501, ¶¶ 35, 55. 135. A preliminary injunction could also chill the efforts of other companies seeking to develop forms of “push technology”—technology that delivers information to the desktop without need for consumers to make an active request each time they see the information. Tr. VII (Naider) 29:22-31:20. CONCLUSIONS OF LAW I. Preliminary Injunction Standard In considering whether to issue a preliminary injunction, the court must *757 consider four factors: “(1) whether the movant has shown a strong likelihood of success on the merits; (2) whether the movant will suffer irreparable harm if the injunction is not issued; (3) whether the issuance of the injunction would cause substantial harm to others; and (4) whether the public interest would be served by issuing the injunction.” Overstreet v. Lexington-Fayette Urban County Gov’t, 305 F.3d 566, 573 (6th Cir.2002). See also Rock and Roll Hall of Fame Museum, Inc. v. Gentile Prods., 134 F.3d 749, 753 (6th Cir.1998) (noting that first factor is whether movants have shown a “strong likelihood” of success on the merits). A movant must also demonstrate that “failure to issue the injunction is likely to result in irreparable harm” to him. United States v. Miami Univ., 294 F.3d 797, 816 (6th Cir.2002) (quoting Kallstrom v. City of Columbus, 136 F.3d 1055, 1068 (6th Cir.1998)). A preliminary injunction is “an extraordinary remedy involving the exercise of a very far-reaching power.” Leary v. Daeschner, 228 F.3d 729, 739 (6th Cir.2000) (quoting Direx Israel, Ltd. v. Breakthrough Med. Corp., 952 F.2d 802, 811 (4th Cir.1991)). It is because preliminary injunctive relief is such a “drastic” remedy that plaintiffs must show circumstances clearly demand its entry. Mazurek v. Armstrong, 520 U.S. 968, 972, 117 S.Ct. 1865, 138 L.Ed.2d 162 (1997) (quoting 11A Charles A. Wright, Arthur R. Miller & Mary Kay Kane, Federal Practice and Procedure § 2948 (2d ed.1995)). II. Plaintiffs Have Not Demonstrated A Strong Likelihood of Success on the Merits of Their Trademark Claims Plaintiffs contend that WhenU infringes on their trademarks in violation of Section 32(1) of the Lanham Act. That section provides in relevant part: Any person who shall, without the consent of the registrant- (a) use in commerce any reproduction, counterfeit, copy, or colorable imitation of a registered mark in connection with the sale, offering for sale, distribution, or advertising of any goods or services on or in connection with which such use is likely to cause confusion, or to cause mistake, or to deceive; ... 15 U.S.C. § 1114(l)(a). To establish a claim for trademark infringement, plaintiffs must prove: (1) ownership of a valid mark that is entitled to protection under the Lanham Act, and (2) that WhenU’s use of the mark is likely to cause confusion within the consuming public. See Champions Golf Club, Inc. v. The Champions Golf Club, Inc., 78 F.3d 1111, 1114 (6th Cir.1996). It is established that plaintiffs own valid marks entitled to protection. The only issue is whether plaintiffs can establish the second element of their trademark infringement claim. A. Defendant Does Not “Use” the Plaintiffs’ Marks “In Commerce” The Lanham Act only forbids the “use in commerce... of a registered mark in connection with the sale, offering for sale, distribution or advertising” of goods or services. 15 U.S.C. § 1114(l)(a). See also 15 U.S.C. § 1127 (a trademark is deemed to be used in commerce only “when it is used or displayed in the sale or advertising of services”). There can be no liability under the Lanham Act absent the use of a trademark in a way that identifies the products and services being advertised by the defendant. See, e.g., DaimlerChrysler AG v. Bloom, 315 F.3d 932, 936 (8th Cir.2003); Holiday Inns, Inc. v. 800 Reservation, Inc., 86 F.3d 619, 623-25 (6th Cir.1996). Plaintiffs’ trademarks do not appear in WhenU ads or coupons. The only trademarks that appear in a WhenU ad are WhenU’s own marks and the marks of its advertisers. FF ¶¶ 78-91. Thus, this is not the “usual trademark case” where “the defendant is using a mark to identify *758 its goods that is similar to the plaintiffs trademark.” Interactive Prods. Corp. v. a2z Mobile Office Solutions, Inc., 326 F.3d 687, 695 (6th Cir.2003). The Lanham Act defines “use in commerce” as “the bona fide use of a mark in the ordinary course of trade.” 15 U.S.C. § 1127. With respect to services, a mark is used in commerce “when it is used or displayed in the sale or advertising of services and the services are rendered in commerce ...” Id. “If [a defendant is] using [a plaintiffs] trademark in a ‘non-trademark’ way that is, in a way that does not identify the source of a product then trademark infringement and false designation of origin laws do not apply.” Interactive Prods. Corp., 326 F.3d at 695 (finding that defendant’s use of plaintiffs mark in post-domain path name of defendant’s website did not constitute “use in commerce” as it did not identify source of goods and was not used with intent to confuse or mislead consumers). Plaintiffs argue that WhenU uses their marks in at least three ways. First, WhenU hinders Internet users from accessing plaintiffs websites. Second, WhenU deliberately positions its “pop-up” advertisements in close proximity to plaintiffs’ trademarks. And third, WhenU uses plaintiffs’ marks to trigger delivery of advertisements. 1. Defendant Does Not Hinder Access to Plaintiffs’ Sites Plaintiffs contend that when consumers attempt to access plaintiffs’ websites and WhenU advertisements “pop-up,” users may mistakenly be diverted to the advertisers’ websites or may become so frustrated, angry, or confused by the advertisements that they stop visiting plaintiffs’ sites or fail to continue searching for the types of services that seem to trigger such advertisements. Plaintiffs cite two cases to demonstrate that this is “use” under the Act: People For Ethical Treatment Of Animals [PETA] v. Doughney, 263 F.3d 359, 365 (4th Cir.2001), and Planned Parenthood Fed’n of America, Inc. [Planned Parenthood] v. Bucci, No. 97 Civ. 0629CKMW), 1997 WL 133313 at *4 (S.D.N.Y.1997). The conduct in neither case, however, is analogous to WhenU’s conduct. In PETA the Fourth Circuit Court of Appeals held that the defendant used the plaintiffs mark “in connection with” goods or services when he registered the domain name “peta.org,” used the name to establish a website for the purported organization “People Eating Tasty Animals,” and provided links on the website to various sources for meat, fur, leather, hunting, animal research, and other organizations with views generally antithetical to PETA’s views. PETA, 263 F.3d at 365. The court concluded that the defendant did not actually have to sell or advertise goods or services on his www.peta.org website to use PETA’s mark in commerce. Id. Rather he only had to prevent users from obtaining or using PETA’s goods or services or had to link his website to other’s goods or services. Id. In reaching this conclusion, the PETA court primarily relied on the reasoning of the District Court for the Southern District of New York in Planned Parenthood v. Bucci, a factually similar case. The defendant in Planned Parenthood registered the domain name “plannedparent-hood.com” and created a website using that name which contained information antithetical to Planned Parenthood’s views. Id. at 365 (summarizing Planned Parenthood). The district court reasoned that the defendant used the plaintiffs mark “in connection with” the distribution of services for the following reason: ... it is likely to prevent some Internet users from reaching plaintiffs own Internet web site. Prospective users of *759 plaintiffs services who mistakenly access defendant’s web site may fail to continue to search for plaintiffs own home page, due to anger, frustration, or the belief that plaintiffs home page does not exist. Id. The PETA court relied on the Planned Parenthood court’s reasoning, even though the defendant’s website, unlike the deceptive Planned Parenthood site, contained a hyperlink to PETA’s official site. Id. at 363. WhenU’s “use” of plaintiffs’ marks is not analogous to the defendants’ use in PETA and Planned Parenthood. WhenU only uses plaintiffs’ marks in its directory, to which the typical consumer does not have access, in order to determine what advertisements to direct to consumers. Unlike PETA and Planned Parenthood, a consumer entering the domain name or URL address for either Wells Fargo or Quicken Loans in fact accesses the Wells Fargo or Quicken Loans websites. Plaintiffs present no evidence to suggest that consumers are unable to reach their sites as a result of the simultaneous appearance of WhenU’s advertisements on their computer screens. To view plaintiffs’ websites in full, consumers only need to move, minimize, or close the advertisement windows. Thus, this Court finds PETA and Planned Parenthood unpersuasive in that defendant does not hinder access to plaintiffs’ sites. 2. The Fact that Some WhenU Advertisements Appear on a Computer Screen at the Same Time Plaintiffs’ Webpages are Visible in a Separate Window Does Not Constitute a Use in Commerce of the Plaintiffs’ Marks a. The Positioning of Defendant’s Advertisements Does Not Constitute “Framing” Plaintiffs argue that WhenU positions its “pop-up” advertisements in such a way that consumers see one display containing WhenU’s advertisements and plaintiffs’ websites and trademarks. This positioning, plaintiffs contend, gives consumers the “impression that the pop-up is affiliated with or approved by [plaintiffs.” Plaintiffs complain that WhenU therefore is relying on, and thus using, plaintiffs’ marks and their associated reputation. To support their claim that this constitutes “use in commerce,” plaintiffs rely on Hard Rock Cafe International (USA), Inc. v. Morton, No. 97 Civ. 9483(RPP), 1999 WL 717995 (S.D.N.Y.1999). Peter Morton (“Morton”), one of the defendants in Hard Rock Cafe, co-founded the first Hard Rock Café in London, England in 1971 with Isaac Tigrett (“Tigrett”). Id. at *2. Morton and Tigrett owned and operated a number of Hard Rock Cafés until a dispute between them in 1985 resulted in the division of their properties. Id. Tigrett subsequently sold his Hard Rock Café interests. Id. By 1990, the Rank Group PLC (“Rank”) had acquired those interests. Id. The plaintiff in Hard Rock Cafe, Hard Rock Café International (USA) Inc. (HRCI), is a subsidiary of Rank. Id. at *1. In 1995, Morton opened The Hard Rock Hotel and Casino in Las Vegas, Nevada. Id. at *2. The following year, Morton sold his interests in the Hard Rock Café business to Rank, although he retained ownership of the Hard Rock Hotel and Casino in Las Vegas and the rights to develop Hard Rock Hotels and Casinos in certain defined territories. Id. at *1. Morton also sold Rank “all of the assets of and utilized in the conduct of the Business,” including certain intellectual property rights; although by a separate agreement entered into on the same date (the “License Agreement”), he obtained a license to use certain service marks and trademarks. Id. More specifically, the License Agreement grant *760 ed Morton permission to operate a website and to make certain uses of the Hard Rock Hotel mark. Morton was restricted, however, from using or exploiting the mark outside the Morton Territories in connection with the sale of merchandise. Id. at *24. In 1997, HRCI commenced a six count lawsuit against Morton claiming that he breached the License Agreement through his operation of an Internet website for the Las Vegas Hard Rock Hotel and Casino (the “Hard Rock Hotel site”). Id. at *1. Relevant to the pending matter is HRCI’s claim that Morton’s site contained a link that opened the website of a third party, Tunes Network, Inc. (“Tunes”), which sold compact disk recordings of music (“CDS”). Id. at *2. A visitor to Morton’s site accessed the Tunes’ site in the following way: When the computer user clicks on the “Shop” icon on the first page of the Hard Rock Hotel web site, a page appears that includes an icon reading “record store.” As of December 1997, clicking on this icon took the user to a Tunes page “framed” by a border on the left and a border on the top, each of which contained a Hard Rock Hotel logo. In the Tunes portion of the page, the computer user can listen to portions of various music CDS and/or purchase music CDS. The CDS are sold by Tunes, not by Hard Rock Hotel. No Hard Rock Hotel logo is shown on the CD packaging but Hard Rock Hotel receives a 5% commission for each of these sales. Id. at *15. Morton argued that the CDS were sold by Tunes, not Hard Rock Hotel, and that the hyperlinks on his website did not involve further “use” of HRCI’s marks because hyperlinks “are merely technical connections between two independent sources of content.” Id. at *25. The District Court for the Southern District of New York concluded otherwise. The court held that Morton used HRCI’s marks in violation of the Licensing Agreement because through the framing mechanism described above, “the Hard Rock Hotel logo appears around the border of a computer screen otherwise filled by the Tunes web page.” Id. at *24-25. The court explained: Framing is far more than a “technical connection between two independent sources of material.” Through framing, the Hard Rock Hotel Mark and the Tunes site are combined together into a single visual presentation and the Hard Rock Hotel Mark is used to promote the sale of CDS by Tunes. Because the Tunes material appears as a window within the original linking page, it is not clear to the computer user that she or he has left the Hard Rock Hotel web site. The domain name appearing at the top of the computer screen ... continues to indicate the domain name of Hard Rock Hotel, not that of Tunes. The Tunes web page is reached in the same fashion as any other section of the Hard Rock Hotel web site, by clicking on a button labeled “record store” which resembles the other buttons leading to web pages maintained by Hard Rock Hotel. The spinning globe, Hard Rock Hotel’s logo, appears not only to the side of the framed Tunes web page, but also within the Tunes menu bar, on the Tunes page itself. The Hard Rock Hotel web site and the Tunes web page are thus smoothly integrated. In light of this seamless presentation of the Tunes web page within the Hard Rock Hotel web site, the only possible conclusion is that the Hard Rock Hotel Mark is used or exploited to advertise and sell CDS. Id. (internal citations to record omitted). The court further noted that “[f]raming is a flexible device, and in other cases the distinction between the two sources of ma *761 terial appearing on the screen might be clear to the computer user. Such was not the case here.” Id. at *25 n. 16. In contrast, in the instant case, when WhenU’s advertisements pop-up and partially overlap plaintiffs’ sites on the computer screen, it seems apparent to the user that what is appearing on his or her screen are two distinct sources of material. Unlike the framing mechanism in Hard Rock Cafe, the windows containing WhenU’s advertisements are not situated over plaintiffs’ websites in such a way that plaintiffs’ marks appear to be part of the WhenU window. In other words, the presentation of the two windows is not “seamless.” Hard Rock Hotel’s logo, the spinning globe, actually appeared “to the side of the framed Tunes web page” and “within the Tunes menu bar, on the Tune’s page itself.” Id. at *24-25. Plaintiffs’ marks are neither displayed or appear to be displayed on WhenU’s windows, and the fact that WhenU advertisements appear on a computer screen at the same time plaintiffs’ webpages are visible in a separate window does not constitute a use in commerce of plaintiffs’ mark. b. WhenU is Engaged In Legitimate Comparative Advertising The juxtaposition of WhenU’s advertisements with plaintiffs’ websites in separate windows on a participating consumer’s computer screen is a form of comparative advertising. U-Haul Int’l, Inc. v. WhenU. com, Inc., 279 F.Supp.2d 723, 728 (E.D.Va., 2003) (even if mere simultaneous appearance of defendant’s ad and plaintiffs’ marks constituted a “use” within the meaning of the Lanham Act, it is still immune from liability as a form of legitimate comparative advertising). Comparative advertising rests on the premise that a competitor’s trademark may appear at the same time as the trademark owner’s. 3 J. Thomas McCarthy, McCarthy on Trademarks & Unfair Competition § 25:52 (4th ed.2003) (hereinafter “_ McCarthy”). See, e.g., G.D. Searle & Co. v. Hudson Pharm. Corp., 715 F.2d 837, 841 (3rd Cir.1983) (inclusion of Metamucil mark on competitor’s product not a prohibited use under Lanham Act); Diversified Mktg., Inc. v. Estee Lauder, Inc., 705 F.Supp. 128, 132 (S.D.N.Y.1988) (finding phrase “If You Like ESTEE LAUDER.. .You’ll Love BEAUTY USA” lawful comparative advertising). In accusing WhenU of “free riding” on their trademarks, plaintiffs ignore the fact that trademark laws are concerned with source identification. They are not meant to protect “consumer good will (sic) created through extensive, skillful, and costly advertising.” Smith v. Chanel, Inc., 402 F.2d 562, 566 (9th Cir.1968). The rule favoring comparative advertising “rests upon the traditionally accepted premise that the only legally relevant function of a trademark is to impart information as to the source or sponsorship of the product.” Id. Comparative advertisements may therefore make use of competitors’ trademarks even if the advertiser reaps the benefit of “the product recognition engendered by the owner’s popularization, through expensive advertising, of the mark.” Anti-Monopoly, Inc. v. Gen. Mills Fun Group, 611 F.2d 296, 301 n. 2 (9th Cir.1979). The so-called comparative advertising cases cited by plaintiffs do not hold otherwise. (Reply Memorandum in Support of Plaintiffs’ Motion for Preliminary Injunction (“Plaintiffs’ Reply Mem.”) at p. 2.) Oral-B Labs., Inc. v. Mi-Lor Corp., 810 F.2d 20 (2d Cir.1987), did not involve comparative advertising. August Storck K.G. v. Nabisco, Inc., 59 F.3d 616 (7th Cir.1995), ruled that the simultaneous presentation of the plaintiffs trademarks on the defendant’s packaging was lawful competitive advertising, despite the finding of the *762 district court that the packaging created a possibility of confusion. Id. at 618. Noting that “some [consumers] are bound to misunderstand no matter how careful a producer is,” the Storck court criticized the district court’s overstatement of the plaintiffs “private injury” and its relative disregard of the higher “public interest in competition.” Id. at 618-619. c. The Inclusion of URLs that Include Portions of Plaintiffs’ Trademarked Names in the Scrambled WhenU Directory is Not a Use in Commerce Plaintiffs also argue that WhenU’s inclusion of their marks in the WhenU Directory constitutes “use in commerce.” The inclusion of web addresses in WhenU’s proprietary Directory is done to identify the category the participating consumer is interested in, such as mortgages, and to dispatch a contextually relevant advertisement to that consumer. The advertisement that is displayed does not bear the plaintiffs’ trademarks. To the contrary, it bears WhenU’s marks and branding as well as those of the advertiser. Thus, the SaveNow software only uses URLs to identify the website itself, just like one would have to use the word “Macy’s” to describe the Macy’s department store. This does not constitute the “use” of any trademark belonging to plaintiffs, as that term is used in the Lanham Act, because WhenU does not use any of the plaintiffs’ trademarks to indicate anything about the source of the products and services it advertises. Bird v. Parsons, 289 F.3d 865, 877-78 (6th Cir.2002) (when a domain name is used to indicate an address on the Internet, and not to identify the source of goods and services, it is not functioning as a trademark); U-Haul, 279 F.Supp.2d at 727-28. The cases cited by plaintiffs (Plaintiffs’ Reply Mem. pp. 1-2) do not support their argument that the inclusion of URLs in the WhenU directory constitutes the use of a trademark under the Lanham Act. In Eli Lilly & Co. v. Natural Answers, Inc., 233 F.3d 456, 465 (7th Cir.2000), and New York State Society of Certified Public Accountants v. Eric Louis Assoc., Inc., 79 F.Supp.2d. 331, 342 (S.D.N.Y.1999), the courts considered whether a defendant’s use of a plaintiffs trademark as metatags in its website constitutes trademark infringement. Only a handful of courts have addressed this issue. Most of those courts hold that such use is “use in commerce” and violates the Lanham Act if such use also causes customer confusion or initial interest confusion and is not a “fair use” of the plaintiffs mark. 19 See, e.g., id.; Brookfield, 174 F.3d at 1064-65 (concluding that use of metatags constitutes “use in commerce” and violates trademark laws because it created “initial interest confusion”); Niton Corp. v. Radiation Monitoring Devices, Inc., 27 F.Supp.2d 102 (D.Mass.1998)(holding that direct copying *763 of plaintiffs metatags and HTML code constituted trademark infringement); Trans Union L.L.C. v. Credit Research, Inc., 142 F.Supp.2d 1029 (N.D.Ill.2001)(holding that use of plaintiffs trade name as metatag in defendant’s website was a permissible fair use as metatag simply described defendants and the content of their website); Bihari v. Gross, 119 F.Supp.2d 309, 322-23 (S.D.N.Y.2000)(holding that use of meta-tags constitutes “use in commerce” but finding no trademark violation because defendants only used mark in metatag to fairly identify the content of his websites and did so in good faith that is, without the intention of capitalizing on the plaintiffs reputation and goodwill). The Brookfield court offered the following analogy to explain its holding: Using another’s trademark in one’s me-tatags is much like posting a sign with another’s trademark in front of one’s store. Suppose West Coast’s competitor (let’s call it “Blockbuster”) puts up a billboard on a highway reading “West Coast Video: 2 miles ahead at Exit 7”where West Coast is really located at Exit 8 but Blockbuster is located at Exit 7. Customer’s looking for West Coast’s store will pull off at Exit 7 and drive around looking for it. Unable to locate West Coast, but seeing the Blockbuster store right by the highway entrance, they may simply rent there. Brookfield, 174 F.3d at 1063. But, as the only court to address whether “keying” a plaintiff’s trademark is “use in commerce” stated, “keying” (a process which is more analogous to WhenU’s method of triggering ads than using metatags), is not analogous to the “devious placement of a road sign bearing false information.” See Playboy Enters., Inc. v. Netscape Communications Corp., 55 F.Supp.2d. 1070, 1076 (C.D.Cal.), aff'd, 202 F.3d 278 (9th Cir.1999). In Playboy, Playboy Enterprises filed a trademark infringement action against two Internet search engine operators, Netscape Communications Corp. (“Netscape”) and Excite, Inc. (“Excite”). Id. at 1071. Playboy challenged Netscape’s and Excite’s sale of its trademarks “playboy” and “playmate” as keyword search terms to prompt banner advertisements for competitors’ adult entertainment sites. Id. at 1072. Netscape and Excite keyed hundreds of additional terms to trigger these particular advertisements. Id. In holding that defendants were not using Playboy’s trademarks in commerce, the district court primarily focused on the fact that Internet users only could enter the generic word “playboy” or “playmate” as search terms, which the court concluded were English words in their own right. Id. at 1073. But the court also concluded that Netscape’s and Excite’s use of Playboy’s trademarks was distinguishable from the defendant’s use of the plaintiffs mark in Brookfield. Id. The Playboy court therefore offered this analogy as a more appropriate comparison to the defendants’ conduct: This case presents a scenario more akin to a driver pulling off the freeway in response to a sign that reads “Fast Food Burgers” to find a well-known fast food burger restaurant, next to which stands a billboard that reads: “Better Burgers: 1 Block Further.” The driver, previously enticed by the prospect of a burger from the well-known restaurant, now decides she wants to explore other options. Assuming that the same entity owns the land on which both the burger restaurant and the competitor’s billboard stand, should that entity be liable to the burger restaurant for diverting the driver? Id. at 1075. While the court did not answer this question explicitly, the analogy *764 itself and the court’s ultimate holding indicate that the answer would be no. The Playboy court’s analogy presents a closer scenario to the present case than that used in Brookfield, and it supports the conclusion that the inclusion of plaintiffs’ marks in defendant’s Directory is not a use in commerce. B. Plaintiffs Have Not Demonstrated Likelihood of Confusion 1. The Legal Standard Although the Court’s holding that defendant has not impermissibly used plaintiffs’ marks makes it unnecessary to reach the issue of likelihood of confusion, plaintiffs’ failure to establish this element of their claim further weakens their request for injunctive relief. In typical trademark cases, courts determine whether a likelihood of confusion exists by examining and weighing the following factors: (1) the strength of the senior mark; (2) relatedness of the goods and services; (3) the similarity of the marks; (4) evidence of actual confusion; (5) the marketing channels used; (6) likely degree of purchaser care; (7) the intent of the defendant in selecting the mark; and (8) the likelihood of expansion of the product lines. Holiday Inns, 86 F.3d at 623 (citing Frisch’s Rest., Inc. v. Elby’s Big Boy, Inc., 670 F.2d 642, 648 (6th Cir.1982)). These factors are not exhaustive and courts may consider some or none of them, or expand upon them. As the Sixth Circuit cautioned, “ ‘[e]ach case presents its own complex set of circumstances and not all of these factors may be particularly helpful in any given case.’ ” Interactive Prods., 326 F.3d at 694 (quoting Homeowners Group, Inc. v. Home Mktg. Specialists, Inc., 931 F.2d 1100, 1107 (6th Cir.1991)). In the Internet setting in particular, courts have begun to realize that consumer confusion can occur even though the consumer is not actually confused as to the source of goods or services at the point of sale or upon reaching the website to which he or she was “hijacked.” See, e.g., Brookfield, 174 F.3d at 1062; Eli Lilly, 233 F.3d at 465; N.Y. State Society of Certified Public Accts., 79 F.Supp.2d. at 342. This doctrine, referred to as “initial interest confusion,” has been recognized by a handful of courts as actionable under the Lanham Act. See id. In Playboy, the district court explained this doctrine as follows: Initial interest confusion, as coined by the Ninth Circuit, is a brand of confusion particularly applicable to the Internet. Generally speaking, initial interest confusion may result when a user conducts a search using a trademark term and the results of the search include web sites not sponsored by the holder of the trademark search term, but rather of competitors. The Ninth Circuit reasoned that the user may be diverted to an unsponsored site, and only realize that she has been diverted upon arriving at the competitor’s site. Once there, however, even though the user knows she is not in the site initially sought, she may stay. In that way, the competitor has captured the trademark holder’s potential visitors or customers. Playboy, 55 F.Supp.2d. at 1074 (citing Brookfield, 174 F.3d at 1062-64). The Sixth Circuit Court of Appeals, however, has not adopted the initial interest confusion doctrine and has not even acknowledged the doctrine in recent Internet trademark cases. See, e.g., Taubman Co. v. Webfeats, 319 F.3d 770 (6th Cir.2003); Interactive Prods., 326 F.3d at 694. The Sixth Circuit has stated on several occasions that “the only important question [in a trademark infringement action] is whether there is a likelihood of confusion between the paHies’ goods or services,” that is regarding the origin of the goods offered by the parties. Taubman Co., 319 F.3d at 776 (citing Bird, 289 F.3d at 877) (emphasis *765 in original); Interactive Prods., 326 F.3d at 694. The Taubman court wrote, “[u]n-der Lanham Act jurisprudence, it is irrelevant whether customers would be confused as to the origin of the web sites, unless there is confusion as to the origin of the respective products.” Taubman Co., 319 F.3d at 776. Thus, this Court must apply the Frisch factors to determine whether a likelihood of confusion exists in the present case. The factors in dispute are evidence of actual confusion, marketing channels used, and likely degree of purchaser care. For each of those factors, plaintiffs rely on the survey evidence presented by their expert William Neal. For the reasons set forth above and below, the Court finds Mr. Neal’s testimony is not persuasive on this critical issue. 2. Plaintiffs’ Survey Evidence is Unpersuasive a. The Legal Standard Under Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), the Court must ensure that expert testimony “is not only relevant, but reliable.” See also Kumho Tire Co. v. Carmichael, 526 U.S. 137, 147, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999). Daubert and Kumho are codified in Federal Rule of Evidence 702. The Daubert standard applies to survey evidence. See, e.g., The Sports Authority, Inc. v. Abercrombie & Fitch, Inc., 965 F.Supp. 925, 933 (E.D.Mich.1997) (holding that “[tjhe proponent of a consumer survey has the burden of establishing that it was conducted in accordance with accepted principles of survey research”) (citation omitted). Defendant’s expert, Dr. Jacoby, testified that Mr. Neal failed to satisfy fundamental principles of survey research in seven major ways: (a)The surveys did not sample the appropriate universe of respondents; (b) The surveys did not use any demonstrative stimuli or otherwise replicate actual market conditions; (c) The survey questionnaires were biased and leading; (d) Mr. Neal drew unwarranted inferences and failed to take into account obvious alternative explanations; (e) The surveys did not ask the kind of control questions needed to generate an error rate; (f) The surveys were not administered properly, including the use of a panel of regular survey respondents; (g) The surveys did not employ a design that established causation, rendering the surveys results uninterpretable. See FF ¶¶ 102-120; DX 566. These seven items roughly correspond to the factors recommended by the Federal Judicial Center’s Manual for Complex Litigation, (3rd Ed.1995) (hereafter “MCL”) for assessing the validity of a survey. See MCL § 21498. See also Sports Authority, 965 F.Supp. at 933 (applying the MCL factors). These factors also reflect the considerations specifically mentioned in Daubert and in the advisory committee’s notes to Rule 702 of the Federal Rules of Evidence. See Fed.R.Evid. 702, advisory committee’s notes, 2000 Amendments (hereafter “Rule 702 Advisory Committee Notes”). It is not necessary for the Court to review the parties’ arguments as to all of these factors. It is clear from a review of several of the most critical factors that Mr. Neal’s surveys do not provide reliable evidence of likelihood of confusion. See, e.g., Simon Property Group, L.P. v. mySimon, Inc., 104 F.Supp.2d 1033 (S.D.Ind.2000); Winning Ways v. Holloway Sportswear, 913 F.Supp. 1454 (D.Kan.1996). *766 b. The Neal Surveys Lack Probative Value Because They Do Not Remotely Approximate Actual Market Conditions “To have substantial probative value, a survey ... must ... be designed to examine the impression presented to the consumer by the accused product. Therefore, a survey must use the proper stimulus, one that tests for confusion by replicating marketplace conditions.” Conopco, Inc. v. Cosmair, Inc., 49 F.Supp.2d 242, 253 (S.D.N.Y.1999) (citation omitted). A survey that fails to adequately replicate market conditions is entitled to little weight, if any. See, e.g., We Media, Inc. v. Gen. Elec. Co., 218 F.Supp.2d 463, 474 (S.D.N.Y.2002); Nat’l Distillers Prod. Co. v. Refreshment Brands, Inc., 198 F.Supp.2d 474, 484 (S.D.N.Y.2002); Cumberland Packing v. Monsanto Co., 32 F.Supp.2d 561, 575 (E.D.N.Y.1999); Coherent, Inc. v. Coherent Tech., Inc., 935 F.2d 1122, 1126 (10th Cir.1991). Mr. Neal’s survey did not replicate market conditions, and therefore has little probative value in establishing likelihood of confusion. See, e.g., Tr. III (Neal) 25:22-24. The Court cannot conclude that consumers are confused about the source of WhenU ads absent testing of WhenU ads. There are many kinds of ads on the Internet (e.g., pop-up ads, pop-under ads, panoramic ads, animated ads) and they come from many different sources. For example, there are pop-up ads generated by software applications that a user downloads on his computer, such as SaveNow; there are pop-up ads generated by internet portals like America Online (Tr. VIII (Naider) 59:16-21); and there are pop-up ads generated by commercial websites. Tr. IV (Edelman) 110:22-111:2. Given the many kinds of Internet ads, and the many different entities who generate them, the Court cannot conclude that the kind of people who use plaintiffs’ websites are confused about the origin of WhenU’s ads without evidence of how those individuals perceive WhenU ads. The Court finds that, at a minimum, survey respondents should have been shown the item that is said to be infringing or confusing. 20 Mr. Neal did not make any attempt to replicate the respondents’ experiences in encountering a WhenU ad. Mr. Neal did not show respondents WhenU pop-up ads, screen shots or other demonstrative stimuli. FF ¶¶ 105-110, and did not take any measures to ensure that the respondents had WhenU ads in mind when responding to the survey. Id. To the contrary, Mr. Neal virtually ensured that his respondents did not have WhenU ads in mind in prefacing his questions with a definition that describes WhenU pop-up (or small format) ads inaccurately, and completely excludes WhenU pop-under ads, panoramic ads and coupons. FF ¶¶ 106-107. 21 *767 c. Plaintiffs’ Surveys Lack Probative Value Because They Do Not Survey the Appropriate Population Identification of the proper universe is recognized as a critical element in the development of a survey. See, e.g., Federal Judicial Center, Reference Manual on Scientific Evidence (2d Ed.2000) (hereafter “MSE”), p. 239 n. 41 (citations omitted); MCL § 21.493; Amstar Corp. v. Domino’s Pizza, Inc., 615 F.2d 252, 264 (5th Cir.1980); Conopco, 49 F.Supp.2d at 253. Selection of a proper universe is so critical that “even if the proper questions are asked in a proper manner, if the wrong persons are asked, the results are likely to be irrelevant.” 5 McCarthy, § 32:159 at 32-250.3. See also 5 McCarthy, §§ 32.160-61 at 32-251-57; Jordache Enterp., Inc. v. Levi Strauss & Co., 841 F.Supp. 506, 518 (S.D.N.Y.1993); Am. Footwear Corp. v. Gen. Footwear Co., 609 F.2d 655, 660-61, n. 4 (2d Cir.1979); Richard J. Leighton, Using Dauberb-Kumho Gatekeeping to Admit and Exclude Surveys in Lanham Act Advertising and Trademark Cases, 92 TMR 743, 763 and n. 98 (July/August 2002) (hereafter “Leighton”). A survey must use respondents from the appropriate universe because “there may be systemic differences in the responses given... by persons [with a particular] characteristic or preference and the responses given to those same questions.. .by persons who do not have that ... characteristic or preference.” Federal Evidence Practice Guide (Matthew Bender 2003) § [4][6][i]. In this case, the relevant universe consists of people who are likely to bank or look for a mortgage online. Because of the rapidity with which the Internet changes, the relevant universe is also current Internet users, i.e., persons familiar with Internet advertising as it exists today and with today’s comfort level in using the Internet. As plaintiffs’ own witnesses acknowledged, the Internet changes rapidly (Tr. II (Neal) 132:22-133:4; Tr. IV (Edel-man) 99:25 (“the internet changes awfully quickly”)), and users develop greater understanding of Internet content with exposure to it. Tr. I (Stapp) 128:8-129:15 (discussing user understanding of sponsored search results). Because Mr. Neal failed to survey the appropriate universe of people, the results of his 1-800 and Gator Surveys cannot be extrapolated to this case, regardless of whether those respondents were demographically representative of the general population of Internet users, as Mr. Neal contends. Tr. II (Neal) 96:7-15. Only a fraction of Internet users bank and obtain mortgages online; therefore, there is no way to know whether any of the Gator and 1-800 respondents were users or potential users of plaintiffs’ websites. FF ¶ 111— 113. Furthermore, there are obvious differences between conducting financial transactions online and buying contact lenses online or reading an online periodical. FF ¶ 111-113. Indeed, the record suggests a relatively high degree of sophistication on the part of plaintiffs’ customers, both with respect to the Internet and financial services. FF ¶¶ 112. d. The Survey Questionnaires Were Unclear and Leading A survey is not reliable if it suggests to the respondents an answer that would not *768 otherwise have occurred to them. More specifically, “[i]t is improper to suggest a business relationship where the respondent may previously have had no thought of any such connection.” 5 McCarthy 32:172. 22 The 1-800 Survey was ostensibly designed to determine whether respondents believed — incorrectly—that there was a relationship between WhenU ads and plaintiffs’ websites. FF ¶ 114-118. The Court cannot find Mr. Neal’s conclusions on this subject reliable, because the survey questionnaire repeatedly suggested to respondents a link between pop-up ads and websites. FF ¶ 115. As set forth in detail in the Court’s Findings of Fact, the questions in Mr. Neal’s survey were flawed in numerous other respects. See FF ¶¶ 114-120. Accordingly, the Court cannot place any weight on the data they produced. It was also incumbent on Mr. Neal to “adequately account for obvious alternative explanations” for his survey data. Fed. R.Evid. 702, advisory committee’s notes (citing, Claar v. Burlington Northern R.R., 29 F.3d 499, 502-504 (9th Cir.1994)). As set forth above, Mr. Neal failed to rule out alternative explanations for his surveys’ results. e. The Survey Was Not Properly Administered, Contained No Control Questions to Generate an Error Rate, and Employed a Design that Rendered the Results Uninterpretable Mr. Neal’s surveys suffered from numerous other flaws, including the use of an Internet Panel (see FF ¶ 120) and the failure to use control questions to generate an error rate. (see FF ¶ 120). Of even greater significance, Mr. Neal failed to employ an experimental design that established causation. FF ¶ 119. See, e.g., Cumberland Packing, 32 F.Supp.2d at 574 (“In a test of a causal proposition the appropriate use of controls is crucial.”). Courts have widely recognized the need for consumer surveys to adjust for so-called “background noise,” i.e., extrinsic factors, preexisting beliefs, general confusion or other factors, other than the stimulus at issue, that contribute to a survey’s results. See, e.g., Greenpoint Fin. Corp. v. The Sperry & Hutchinson Co., Inc., 116 F.Supp.2d 405, 409 (S.D.N.Y.2000); Winning Ways, 913 F.Supp. at 1475-76. See also Smith Kline Beecham Consumer Healthcare L.P. v. Johnson & Johnson-Merck Consumer Pharm. Co., 2001 U.S. Dist. LEXIS 7061 at *38-39 (S.D.N.Y. June 1, 2001). As Dr. Jacoby testified, a survey design must include a control group in order to account for the effects of “noise.” FF ¶ 119. The control group “functions as a baseline and provides a measure of the degree to which respondents are likely to *769 give an answer ... not as a result of the [thing at issue], but because of other factors, such as the survey’s questions, the survey’s procedures ... or some other potential influence on a respondent’s answer such as pre-existing beliefs.” Novartis Consumer Health, Inc. v. Johnson & Johnson-Merck Consumer Pharms. Co., 129 F.Supp.2d 351, 365 n. 10 (D.N.J.2000). “By adding an appropriate control group, the survey expert can test exactly the influence of the stimulus.” 5 McCarthy § 32:187 (quoting MSE, p. 250). Had Mr. Neal used a control group, he might have been able to make a “causal inference” that was “clear and unambiguous.” Id. Thus, the Court finds that plaintiffs have failed to show that they are likely to prevail on the merits of their trademark infringement claim. Plaintiffs have not established that defendant used their work in commerce within the meaning of federal trademark law, and have not established any likelihood of confusion in defendant’s “use” of their marks. 23 III. Plaintiffs Have Not Shown Strong Likelihood of Success on the Merits of Their Copyright Claims A. Plaintiffs Have Not Shown Any Infringement of Their Right to Prepare Derivative Works Plaintiffs claim that WhenU violates their exclusive right to prepare “derivative works.” Complaint ¶ 128. To prevail on this claim, plaintiffs would have to show that WhenU has incorporated the plaintiffs’ websites into a new work. 17 U.S.C. § 101 (defining derivative work as “[a] work based on one or more preexisting works” that is “recast, transformed or adopted”). Plaintiffs have not made the necessary showing as to WhenU because WhenU merely provides a software product to computer users. The SaveNow software does not access plaintiffs’ websites; therefore, it does not incorporate them into a new work. FF ¶¶ 84-94. Accordingly, plaintiffs’ claim that the defendant violates their right to create derivative works can only be understood as a contributory copyright theory. Moreover, SaveNow users do not infringe plaintiffs’ right to prepare derivative works because consumers who cause the display of WhenU advertisements or coupons on their screens do not alter plaintiffs’ websites. Plaintiffs’ websites reside on separate servers. The WhenU Window has no physical relationship to plaintiffs’ websites, and does not modify the content displayed in any other open window. FF ¶¶ 84-94. Even if the presence of an overlapping window could be said to change the appearance of the underlying window on a computer screen, the mere alteration of the manner in which an individual consumer’s computer displays the content sent by plaintiffs’ websites does not create a “derivative work.” Lewis Galoob Toys v. Nintendo of Am., 780 F.Supp. 1283, 1291 (N.D.Cal.l991)(indicating that “the consumer may experiment with the product and create new variations of play, for personal enjoyment, without creating a derivative work”), aff'd, 964 F.2d 965 (9th Cir.1992). *770 New York Times Co., Inc. v. Tasini, 533 U.S. 483, 121 S.Ct. 2381, 150 L.Ed.2d 500 (2001), does not hold to the contrary. Unlike the publisher in Tasini, WhenU is not copying or making additions to or deletions from plaintiffs’ actual copyrighted works. Contrary to plaintiffs’ claim, WhenU has not added anything to plaintiffs’ web pages. If one were able to look at the HTML code of plaintiffs’ sites, one would not see any changes as a result of WhenU’s advertisements. In this respect, the effect of WhenU’s advertisements on plaintiffs’ sites is more akin to the affect of a video game accessory in Lewis Galoob Toys. See also U-Haul, 279 F.Supp.2d at 730 (Tasini not applicable where no reproduction of plaintiffs work occurs). In Lewis Galoob Toys, the seller of a video game accessory known as the “Game Genie Video Game Enhancer” (“Game Genie”) filed an action seeking a declaratory judgment that it was not violating or contributing to the violation of the defendant’s video game copyrights. The defendant, Nintendo of America (“Nintendo”), markets and sells a home video games hardware system known as the Nintendo Entertainment System (“NES”) and compatible video game cartridges. Id. at 1285. The plaintiff markets and sells the Game Genie, which fits between the NES control desk and compatible video game cartridges and allows players to temporarily alter certain attributes of video games. Id. at 1286, 1288. Nintendo argued that the attachment of the Game Genie to its copyrighted works created a derivative work. The district court rejected Nintendo’s argument, finding that a consumer utilizing the Game Genie for noncommercial, private enjoyment “neither generates a fixed transferable copy of the work, nor exhibits or performs the work for commercial gain.” Id. at 1291. The court explained, [Ijinherent in the concept of a “derivative work” is the ability for that work to exist on its own, fixed and transferable from the original work ... [o]nce the Game Genie and its attached game cartridge are disconnected from the NES, or the power is turned off, those changes disappear and the video game reverts to its original form. Id. WhenU’s conduct affects plaintiffs’ sites in a comparable manner. It only temporarily changes the way the sites are viewed by consumers. As soon as the advertisements are “disconnected” — that is closed or minimized — plaintiffs’ sites revert to their original form. If anything, WhenU’s advertisements modify their sites far less that the Game Genie altered users’ NES video game experience. The Court also finds irrelevant a series of cases cited by plaintiffs in which copyrighted material was not merely altered, but also publicly re-transmitted in the altered form. See, e.g., WGN Cont'l Broad. Co. v. United Video, Inc., 693 F.2d 622 (7th Cir.1982) (re-broadcast of television program with integrated text deleted); Gilliam v. Am. Broadcasting Cos., Inc., 538 F.2d 14 (2d Cir.1976) (public television broadcast of edited work); Nat’l Bank of Commerce v. Shaklee Corp., 503 F.Supp. 533 (W.D.Tex.1980) (distribution of books with addition of unauthorized advertising materials). In marked contrast, plaintiffs here do not allege any general or public re-transmission of the alleged derivative work by computer users. Plaintiffs base their allegations of copyright violation on the assertion that, because WhenU ads modify the pixels on a computer user’s on-screen display, this modification creates a “derivative work.” The Court finds this argument unpersuasive in light of plaintiffs’ expert’s admission that pixels form part of the hardware of a computer and are owned and con *771 trolled by the computer user who chooses what to display on the screen. Plaintiffs do not have any property interest in the content of a user’s pixels, much less a copyright interest. FF ¶¶ 90-91. Further, in order for a work to qualify as a derivative work, it must be independently copyrightable. Woods v. Bourne Co., 60 F.3d 978, 990 (2d Cir.1995). To be independently copyrightable, it must be “fixed”—that is, it must be “sufficiently permanent or stable to permit it to be ... reproduced.” See 17 U.S.C. §§ 101, 102; Medforms, Inc. v. Healthcare Mgmt. Solutions, Inc., 290 F.3d 98, 107 (2d Cir.2002). See also Lewis Galoob Toys, 964 F.2d at 967 (noting that “[a] derivative work must incorporate a protected work in some concrete or permanent ‘form’ ”); Micro Star v. Formgen, Inc., 154 F.3d 1107, 1111 n. 4 (9th Cir.1998) (noting, by way of example, that covering a television screen with pink cellophane, while modifying the appearance of the copyrighted programs, would not create a derivative work “because it does not incorporate the modified image in any permanent or concrete form”). The pixels on a computer screen are updated every l/70th of a second. FF ¶ 91. The alteration of pixels is therefore far too transitory an occurrence to form a basis for a copyright violation. The appearance of a WhenU advertisement on a consumer’s computer screen at the same time as one of the plaintiffs’ web pages is also a transitory occurrence that might never be duplicated exactly on that or another person’s computer screen. U-Haul, 279 F.Supp.2d at 731. Accordingly, the WhenU advertisement does not create a work that is sufficiently permanent to be independently copyrightable, and hence does not create a derivative work. Since SaveNow users do not infringe the plaintiffs’ right to prepare derivative works, WhenU is not liable for contributory infringement. See Matthew Bender & Co. v. West Pub., 158 F.3d 693, 706-07 (2d Cir.1998). IV. Plaintiffs Have Not Suffered Irreparable Harm A finding of irreparable harm is “the single most important prerequisite that the Court must examine when ruling upon a motion for a preliminary injunction.” MetroBanc v. Fed. Home Loan Bank Bd., 666 F.Supp. 981, 984 (E.D.Mich.1987). Irreparable harm means more than merely “substantial” harm. Ramik v. Darling Int’l, Inc., 161 F.Supp.2d 772, 778 (E.D.Mich.2001). In this matter, plaintiffs have not demonstrated irreparable or even substantial injury. Although they argue that WhenU is “free-riding” on their reputation and goodwill, to the extent plaintiffs could demonstrate that WhenU took advantage of their marks to benefit itself without compensation, plaintiffs’ injury would be purely monetary in nature, and compensa-ble by paying damages for the alleged unjust enrichment. Because the Court finds that plaintiffs’ copyright claims are without merit, plaintiffs are not entitled to any presumption of irreparable harm. Ronald Mayotte & Assocs. v. MGC Bldg. Co., 885 F.Supp. 148, 153 (E.D.Mich.1994). However, even if plaintiffs had a valid copyright claim, the presumption is rebutted in this case by plaintiffs’ own conduct, which is inconsistent with their assertion of irreparable injury. A. Plaintiffs’ Delay in Asserting Their Rights Rebuts Any Claim of Irreparable Injury Although plaintiffs were aware of the alleged effects of SaveNow on their businesses as early as August, 2002, they delayed nine months before bringing a motion for injunctive relief. Plaintiffs’ delay *772 in seeking a preliminary injunction undermines their allegation of irreparable harm. Citibank, N.A. v. Citytrust, 756 F.2d 273, 276 (2d Cir.1985) (“Significant delay in applying for injunctive relief in a trademark case tends to neutralize any presumption that infringement alone will cause irreparable harm pending trial, and such delay alone may justify denial of a preliminary injunction for trademark infringement.”). See also Ramik, 161 F.Supp.2d at 778. B. WhenU Poses No Threat to Plaintiffs’ Relationship with Their Banking Regulators At the preliminary injunction hearing, plaintiffs raised allegations concerning alleged violations of federal banking law in an effort to bolster them showing of injury. These allegations, which plaintiffs failed to assert in their complaint or motion for preliminary injunction, do not appear to have merit. The apparent legal basis for plaintiffs’ concern is a regulation issued by the Office of the Comptroller Currency (“OCC”). 24 Electronic Activities, 67 Fed.Reg. 34,992 (May 17, 2002) (codified at 12 C.F.R. pt. 7). That regulation prescribes rules concerning the practice of national banks who “enter into joint marketing relationships with third-parties through the Internet.” Electronic Activities, 67 Fed.Reg. at 35,002. These rules place certain regulatory obligations on banks that “share electronic space, including a co-branded web site” with others. Shared Electronic Space, 12 C.F.R. § 7.5010 (2003). WhenU is not involved in a joint marketing relationship with any bank and does not share a co-branded website with any bank. Accordingly, this regulation has no bearing on this case. The Court notes that despite spending many days entering evidence into the record, plaintiffs did not introduce any evidence suggesting that bank regulators are concerned about the effects of WhenU software. The Court finds plaintiffs’ suggestions concerning the regulatory impact of WhenU to be unsubstantiated. Y. Issuance Of A Preliminary Injunction Would Harm WhenU and Others Issuance of a preliminary injunction would significantly harm WhenU’s business. Injunctive relief would disrupt WhenU’s established relationships with advertisers and harm WhenU’s ability to obtain new advertisers. Tr. (Naider) VIII, 36:23-37:19. Moreover, the harm to WhenU would not only include the loss of client business that would take years to regain, but also the potential, permanent loss of talented and specially trained staff. Tr. (Naider) Vol. VIII, 37:23-38:5. Issuance of a preliminary injunction would also cause harm to WhenU advertisers, who would lose the ability to have their competitive offers delivered to potential customers by SaveNow simply because those customers view content from the plaintiffs’ websites. DX 501, ¶¶ 55-59. Granting an injunction to protect plaintiffs from the rigors of competition also threatens the integrity of the competitive process. Plaintiffs’ objection to WhenU’s advertising is that it presents customers with alternative choices for procuring the services offered by plaintiffs, increasing the chance that prospective customers will entertain more attractive offers. Federal policy has long favored such comparative advertising and disfavored restrictions on such advertising. See, e.g., In Regard to *773 Comparative Advertising, 16 C.F.R. § 14.15(c) (2003). 25 The Court concludes that, based on the evidence presented, a preliminary injunction will cause significant harm to defendant, defendant’s clients and users, and the general public. Denying the motion will not damage plaintiffs other than in a manner compensable by an award of monetary damages in the event plaintiffs eventually prevail on the merits. Accordingly, the Court finds that the balance of the equities weighs against granting plaintiffs’ motion. CONCLUSION Based on all of the foregoing, Plaintiffs’ motion for preliminary injunction is DENIED. Footnotes [1] Quicken Loans is authorized to bring this lawsuit by virtue of agreement with Intuit Inc., which gave Quicken Loans the right to bring this suit against unauthorized uses of the licensed Quicken Loans Marks. PX 104, ¶ 11. [2] Except as specifically indicated, references to SaveNow refer to both WhenU's proprietary Save and SaveNow software programs. [3] DX 501 is Mr. Naider's Affidavit in Opposition to Plaintiffs’ Motion for a Preliminary Injunction, and DX 502-DX 518 are the exhibits to that Affidavit. Mr. Naider’s Affidavit together with the exhibits thereto were also received in evidence as PX 506. [4] SaveNow also makes available to participating consumers "dollars off," "percentage off” and other savings coupons for products and services at hundreds of online retail merchants. Such a coupon might remind a consumer, for example, of a free shipping offer that is available by using a particular product code when purchasing the product. A team of WhenU content researchers tests and updates these offers on a daily basis. The coupons offered by SaveNow afford participating consumers the opportunity to make significant savings while shopping online. Tr. VII (Naider) 21:4-22:22; DX 501, ¶27; DX 507. [5] Mr. Edelman was apparently not thinking of the License Agreement when he testified that WhenU does not obtain user consent to update the Directory. Tr. V (Edelman) 81:14— 82:6. [6] For a brief period of time, consumers whose Internet browser security settings had been set below the levels recommended by Microsoft could receive Save through software download prompts without being required to agree to the License Agreement. Tr. VII (Naider) 61:19-62:14; DX 538, ¶¶ 14-15. This problem affected only a tiny percentage of users. Tr. VII (Naider) 65:23-66:24. Currently, even users with browsers set to the lowest security settings instead of the "medium” security setting recommended by Microsoft cannot download Save without indicating their assent to the License Agreement. Tr. VII (Naider) 61:15-24. [7] Most of the terms in the directory are related to the delivery of SaveNow advertisements; however, some terms do not trigger ads but are in the database for administrative purposes. Thus, Mr. Edelman was mistaken in his belief that WhenU targets secured websites (Tr. V (Edelman) 61:23-65:2; Tr. VI (Edelman) 21:11-22:8) or uses its keyword system to target specific websites. Tr. V (Edelman) 61:23-69:7; PX 109, ¶ 31. The elements of the WhenU Directory Mr. Edel-man cited in his testimony are not used to trigger advertisements. Tr. VI (Edelman) 116:6-117:25; Tr. VII (Naider) 102:1-12:8. [8] Indeed, initially Mr. Edelman contended that SaveNow lacked any functionality for mapping advertisements to categories. PX 109, ¶ 30 ("Avi Naider's declaration alleges that ... websites are grouped into substantive categories with which WhenU staff then associate multiple advertisements .... If WhenU's software ... has such a functionality, I could not locate it ...”). See generally PX 109, ¶¶ 26-30; Tr. IV (Edelman) 134:15-136:16. Ultimately, Mr. Edelman conceded he had been mistaken. Tr. VI (Edelman) 140:13-141:24; Tr. IX (Edelman) 84:3-6, 87:15-88:1. Because the Court finds that Mr. Edelman's understanding of the complex Sa-veNow software is incomplete and imperfect, and because Mr. Naider has a much better understanding of the software, the Court has credited Mr. Naider's testimony on the subject of the operation of SaveNow to the extent it conflicts with Mr. Edelman’s testimony. Tr. IX (Edelman) 87:15-88:1; Tr. VIII (Naider) 44:17-23. On cross-examination of Mr. Naider, plaintiffs' counsel elicited that Mr. Naider does not have a degree in computer science. Tr. VIII (Naider) 43:13-17. Interestingly, neither of the parties' computer experts has a degree in computer science. PX 109; DX 523. [9] WhenU’s pop-up advertisements have a "refresh” function that may cause them to come back to the front of the screen one time after retreating behind the window the user has activated. This feature operates only once. Tr. VII (Naider) 41:3-12. [10] Although all WhenU ads are branded by WhenU, WhenU has distributed at least some ads without the advertiser’s name. However, in September 2003, WhenU implemented a policy requiring that all graphic advertisements display the name of the advertiser. Tr. VII (Naider) 52:17-19. [11] Plaintiffs attempted to show through Dr. Jacoby that disclaimers do not work. However, Dr. Jacoby testified that disclaimers can be effective. Tr. IV (Jacoby) 84:22-85:24. The fact that Dr. Jacoby authored articles about empirical research he did on two disclaimers which had not been prominently displayed and which he determined to be ineffective is therefore irrelevant. Tr. IV (Jacoby) 83:7-87:4. [12] It took ten days to administer and collect the data. Tr. II (Neal) 123:25-124:15, 132:7-11. Mr. Neal took another four days to analyze the data and write his report. Tr. Ill (Neal) 46:14-47; 11; PX 159. [13] Mr. Neal acknowledged that consumers who see WhenU ads over and over again with their specific format might come to recognize those ads as coming from a particular source Tr. II (Neal) 146:25-147-11. On re-direct, counsel elicited from Mr. Neal that some WhenU ads are branded “Save,” and some are branded "SaveNow.” Tr. Ill (Neal) 23:2-6. However, a user who downloads Save-Now sees ads with the SaveNow brand; a user who downloads Save sees ads with the Save brand. Tr. VII (Naider) 46:14-18. [14] During the hearing, plaintiffs’ counsel suggested that unlike Mr. Neal, Dr. Jacoby is a "professional witness.” Tr. IV (Jacoby) 75:12-15. However, Mr. Neal is being compensated by the plaintiffs, and has been retained by plaintiffs' counsel on at least four occasions. Tr. II (Neal) 130:3-22. Counsel also cross-examined Dr. Jacoby on the handful of cases in which his findings had been criticized (Tr. IV (Jacoby) 70:16-75:4), suggesting that Dr. Jacoby had intentionally "rigged a survey to get a particular result.” Tr. IV (Jacoby) 74:21-75:4. I have read those cases, and I have also read cases that laud Dr. Jacoby's credentials and research. See, e.g., Indianapolis Colts v. Metropolitan Baltimore Football Club Ltd. Partnership, 34 F.3d 410, 415 (7th Cir.1994) (Judge Posner characterizing Dr. Jacoby’s survey evidence and related testimony as having "all the trappings of social scientific rigor"); Hill's Pet Nutrition v. Nutro Products, 258 F.Supp.2d 1197, 1210 (D.Kan.2003) (finding “credentials and testimony of Dr. Jacoby to be impeccable, and his rationale thoroughly persuasive”); Beacon Mutual Ins. Co., v. Onebeacon Ins. Group, 253 F.Supp.2d 221, 225-26 (D.R.I.2003). I do not find it surprising that a witness who has testified in over 100 cases (Tr. IV (Jacoby) 8:24-9:2) has been criticized from time to time, and find that Dr. Jacoby is a highly qualified and well-recognized expert in consumer confusion and survey evidence. [15] These inaccuracies were likely due to the fact that Mr. Neal knew almost nothing about SaveNow ads when he designed the 1-800 Survey, never having seen SaveNow in operation. Tr. II (Neal) 45:10-46:12. In formulat *752 ing the definition, Mr. Neal relied on information he received from 1-800's attorneys and a perusal of WhenU's website. Tr. II (Neal) 128:6-129:7. While Mr. Neal testified on his re-direct examination that he had seen Save-Now ads in operation prior to conducting the 1-800 survey (Tr. III (Neal) 22:10-18), this testimony is inconsistent with Mr. Neal's pri- or testimony in the 1-800 case and on cross-examination. Tr. II (Neal) 128:16-19; Tr. III (Neal) 43:4-45:19. [16] Mr. Neal testified that he had limited his survey to the pop-up format because he believed (erroneously) that it was the predominant format. Tr. II (Neal) 156:16-21. [17] This testimony is perplexing as Mr. Neal testified that plaintiffs' counsel provided him with some screen shots of SaveNow ads. Tr. II (Neal) 128:20-25. [18] The Gator Survey used a sample of people who reported that they had accessed an online newspaper or magazine within the last two months. PX 159, Ex. 4, p. 19. The 1-800 Survey relied on a sample of people who reported that they were contact lens users or *753 anticipated getting contact lenses within the next year and would "consider using the Internet to shop for or purchase contact lenses.” PX 158, Ex. B (S7, S11, S12). There were no questions in either survey concerning the use or likely use of online financial services. PX 158; PX 159. [19] “Metatags are HTML code intended to describe the contents of the web site." Brookfield, 174 F.3d at 1045. It is interesting to note that in many of the cases involving meta-tags, the defendant's use of the plaintiff's mark was not limited to use as a metatag on the defendant's website. For example, in New York State Society of Certified Public Accountants, the defendant also used plaintiff's Internet domain name NYSSCPA to establish his Website at NYSSCPA.com (“.org” was the top-level domain of plaintiff's site). New York State Society of Certified Public Accts., 79 F.Supp.2d. at 338-39. In Brookfield, the defendant used plaintiff's mark as the domain name for its Website. Brookfield, 174 F.3d at 1043-44. And in Eli Lilly, the court focused on the defendant's use of the plaintiff's mark “Prozac” in the name of its herbal alternative "Herbrozac,” in deciding that defendant infringed upon plaintiffs mark, rather than defendant's inclusion of the mark as metatags on its website. Eli Lilly, 233 F.3d at 462-64. [20] Plaintiffs' expert William Neal erroneously alleged that an article by defendant's expert (PX 163) discussed testing for confusion without showing a stimulus. PX 190 at 8. In fact, Dr. Jacoby was testing materiality, not confusion. PX 163 at 2. ("Reflecting different theories of the case, [the FTC's and Jacoby's surveys] focused on differing components ... The FTC’s survey (Stewart 1987) addressed the question: Were the representations likely to mislead? In contrast, Kraft's survey (Jaco-by 1988) addressed the question: Were the representations material? ”). [21] In Pharmacia & Upjohn Co. v. Generation Health, 1997 WL 750605, 44 U.S.P.Q.2d (BNA) 1091 (W.D.Mich. July 15, 1997), the court encountered a survey with a similar defect, in a case involving two anti-cholesterol products with similar names. Plaintiff, the maker of Colestid, obtained a survey to measure the confusion caused by the similarity of the Colestid name to the name of defendant's product Cholestin. Consumers were played a radio advertisement for one of the brands, and then asked to identify the product which they had just heard advertised. Since plaintiff did not advertise Colestid on the radio, the respondents who heard the Colestid "adver *767 tisement” actually heard Cholestin's radio advertisement, with references to Cholestin replaced with Colestid. As a consequence, the Colestid “advertisement" did not accurately describe Colestid (e.g., it stated incorrectly that Colestid was available "over-the-counter” at “local food, drug, and discount stores," that Colestid "works naturally" to reduce cholesterol and that Colestid helped raise "good" cholesterol). The court found that "by exposing some participants to a highly unrealistic situation, the survey says little about the likelihood of confusion of the marks ..."Id. at *15 n. 6. [22] For example in Simon Property Group v. mySimon, the survey’s questions “implicitly suggested] to the respondent the possibility of a business connection between the SPG and mySimon home pages that the respondent may not have made on his or her own.’’ 104 F.Supp.2d at 1048. See also Universal City Studios, Inc. v. Nintendo Co., 746 F.2d 112, 118 (2d Cir.1984) ("To the best of your knowledge, was the Donkey Kong game made with the approval or under the authority of the people who produce the King Kong movies?” improperly suggested a connection to respondents); Beneficial Corp. v. Beneficial Capital Corp., 529 F.Supp. 445 (S.D.N.Y.1982) (rejecting the question, "Do you think there may or may not be a business connection between Beneficial Capital Corp. and the Beneficial Finance System Companies?” on the grounds that it was a leading question "not well suited to eliciting an uninfluenced reaction”). [23] The fact that WhenU advertisements are conspicuously branded, and state on their face, "This is a WhenU offer and is not sponsored or displayed by the website you are visiting. More ...” further dispels any likelihood of confusion. See FF ¶ 74. In this Circuit, courts have found disclaimers such as the ones employed by WhenU to be “very informative” and have relied on them in declining to find confusion. Taubman Co. v. Webfeats, 319 F.3d 770, 776 (6th Cir.2003) (citing Holiday Inns, Inc. v. 800 Reservation, Inc., 86 F.3d 619 (6th Cir.1996)). [24] Plaintiff Quicken Loans is not a national bank and hence not subject to regulation by the OCC. Although Quicken Loans has applied for a banking license, its application is for a license as a federal savings bank, which is subject to regulation by Office of Thrift Supervision, not the OCC. [25] Finally, as a technical matter, the Court does not agree with plaintiffs’ contention that the issuance of injunctive relief can be narrowly tailored by simply ordering WhenU to place all of plaintiffs' URLs on a "suppress list.” Tr. VI (Edelman) 34:4-10. Putting plaintiffs’ URLs on a suppress list might not satisfy the host of concerns raised by plaintiffs in this lawsuit. Tr. VIII (Naider) 106:8-107:18; Tr. VI (Edelman) 161:14-162:13. More importantly, the Court finds this argument rather disingenuous, because were the Court to create a precedent because this partial remedy might be relatively easy to implement (by addition to the suppress list), thousands of new plaintiffs could simply follow the lead of Wells Fargo and Quicken Loans, leaving WhenU without a business.