*265In September 2017, the Court held a bench trial between two Alzheimer's charities-Counterclaim Plaintiff Alzheimer's Disease and Related Disorders Association (hereafter, "Alzheimer's Association" or the "Association" or "Counterclaim Plaintiff") and Counterclaim Defendant Alzheimer's Foundation of America (hereafter, "Alzheimer's Foundation" or "AFA" or "Counterclaim Defendant")-on certain Lanham Act claims. Specifically, the Association contends that AFA's purchase of Association trademarks as search engine keywords and use of the two-word name "Alzheimer's Foundation" constitute trademark infringement and false designation of origin under the Lanham Act.
For the following reasons, the Court finds that the Association failed to prove that AFA's actions were likely to cause consumer confusion and enters judgment for the Counterclaim Defendant.
I. Procedural History
The history of the facts and prior proceedings in this matter is too voluminous to warrant full description here. Prior decisions of Judge Sweet, from whom this case was reassigned to the undersigned on November 21, 2016, provide additional background. See Alzheimer's Found. of Am., Inc. v. Alzheimer's Disease & Related Disorders Ass'n, Inc., 796 F.Supp.2d 458 (S.D.N.Y. 2011) ; No. 10-CV-3314, Dkt. No. 1701 ; Alzheimer's Found. of Am., Inc. v. Alzheimer's Disease & Related Disorders Ass'n, Inc., No. 10-CV-3314 (RWS), 2014 WL 4290455 (S.D.N.Y. Aug. 29, 2014) ; See Alzheimer's Found. of Am., Inc. v. Alzheimer's Disease & Related Disorders Ass'n, Inc., No. 10-CV-3314 (RWS), 2015 WL 4033019 (S.D.N.Y. June 29, 2015). For the purposes of this bench opinion, the Court offers the following summary.
In 2007, the two Alzheimer's charities began suing each other. The first suit, in Virginia state court, concerned which entity was entitled to a trust bequest when the trustee made the first check out to "Alzheimer's Foundation," but sent it to the Association. The state judge dismissed the suit in 2009.
In April 2010, AFA sued the Association in this Court related to other checks cashed by the Association but made payable to AFA. In June 2010, the Association countersued, alleging trademark infringement and related state and common-law claims. See No. 10-CV-5013, Dkt. No. 1. On May 25, 2011, Judge Sweet dismissed all of the claims brought by AFA and by the Association except for those under and related to the Lanham Act. Alzheimer's Found. of Am., Inc., 796 F.Supp.2d at 468. In 2014, Judge Sweet dismissed AFA's remaining claims.2 That left only the Lanham-related claims in the Association's countersuit for resolution.3 The Association amended their remaining counterclaims on September 13, 2013, and the Amended Counterclaims are operative here. See Amended Counterclaims ("Am. Count."), Dkt. No. 125.
*266In July 2014, the Association brought a motion for a preliminary injunction, see Dkt. Nos. 183, 185, seeking to enjoin AFA "from purchasing Association-owned trademark terms from any search engine provider and from utilizing the phrase 'Alzheimer's Foundation' ...in its advertising." See Alzheimer's Found., 2015 WL 4033019, at *1. Following a two-day hearing, on June 29, 2015, Judge Sweet denied the Association's motion for a preliminary injunction. See generally id. After performing much of the same analysis the Court undertakes in this post-trial opinion, Judge Sweet concluded that "[t]he Association has failed to clearly establish the three most important factors in the preliminary injunction analysis: a likelihood of success on the merits, the prospect of irreparable harm, and the balance of the hardships tipping in its favor." Id. at *14. While Judge Sweet acknowledged that "the Association makes strong points on several of the Polaroid factors," he found that "its case for confusion falls short of the 'clear or substantial likelihood of success' required for a preliminary injunction that would alter the status quo." Id. (quoting Sunward Elecs., Inc. v. McDonald, 362 F.3d 17, 24-25 (2d Cir. 2004) ).
Following reassignment to the undersigned, the Court set a motions and trial schedule. AFA submitted motions to exclude the report and testimony of both of the Association's expert witnesses, see Dkt. Nos. 269-71, 287-89, as well as a motion in limine to exclude one of the Association's theories of damages. See Dkt. Nos. 293-95. The Court reserved judgment on the motions. See Dkt. No. 286; Transcript of July 26, 2017 Final Pretrial Conference.
The Court held a six-day bench trial beginning on September 6, 2017 in accordance with its Individual Practices in Civil Cases for non-jury proceedings. Prior to trial, the parties submitted declarations of direct and rebuttal testimony as well as copies of anticipated exhibits and deposition designations that they intended to use at trial. The parties also submitted proposed findings of fact and conclusions of law. At trial, the parties called only those witnesses whom they intended to cross-examine and read into the record deposition designations and counter-designations for all other witnesses. In all, the Court received direct testimony from 25 witnesses, 13 of whom also provided live testimony, and admitted roughly 280 exhibits from both parties. Following trial, the parties submitted post-trial memoranda of law and updated proposed findings of fact and conclusions of law that were fully submitted on October 25, 2017.4 The Court then heard closing arguments on December 5, 2017.
II. Findings of Fact5
A. Counterclaim Plaintiff Alzheimer's Association and Its Marks
The Alzheimer's Association, formed in 1980, is the world's largest private non-profit *267funder of Alzheimer's research. FOF ¶¶ 1-2. Since 2005, Harry Johns has served as the Association's CEO. Ex. 256 ¶ 2.6 Based in Chicago, the Association is comprised of more than 80 local chapters across the nation providing services within each community. FOF ¶ 3. Until 2016, most of these chapters were independent, operating under contract with the Association; in late 2015 and early 2016, a number of the chapters merged into the Association, while a few others disaffiliated. See Tr. (Johns) 252:10-253:12; AFA FOF ¶ 6(viii)-(xi). In fiscal year 2016, the Association raised more than $160 million in contributions, an increase from the $87.9 million the Association reported in "total revenues, gains and other support" in fiscal year 2010. FOF ¶ 13, 25. In fiscal year 2016, the Association spent $133.6 million on program activities, including more than $44 million on public awareness and education. FOF ¶ 27. The Association had nearly 9 billion media impressions and more than 41 million website visits in fiscal year 2016. Ex. 256 ¶ 30.
Surveys show that the Association is somewhat well-known. The Association commissioned a firm named Copernicus to conduct consumer awareness surveys (the "Copernicus Surveys"). AFA FOF ¶ 65. Copernicus measured both unaided-that is, "top of mind"-and aided awareness of the Association among different demographic groups. AFA FOF ¶¶ 66, 69, 70. In 2015, when respondents in the general population were prompted for the first two "health charity organizations that support research, help provide and enhance care and support, build disease awareness and provide tips to reduce risks of different diseases" that come to mind, respondents were vastly more likely to name the American Cancer Society (46%), St. Jude Children's Research Hospital (28%), American Heart Association (19%), and Susan G. Komen for the Cure (10%), than they were to name the Alzheimer's Association (3%). See Ex. AF at ALZ0058612. When respondents were asked which organizations "involved in the fight against Alzheimer's disease" they have heard of-the Association registered 8-12% awareness throughout 2011-2015. AFA FOF ¶ 71. Amongst "Champions," a subpopulation comprised of the demographic groups that the Association targets with its advertising messages, unaided awareness of the Association among organizations "involved in the fight against Alzheimer's disease" has hovered between 10% and 20% between 2009 and 2015. See Ex. AF at ALZ0058614. The Copernicus Surveys show aided awareness of "Alzheimer's Association"-that is, where respondents are prompted with the specific name-at roughly 25-32% among the general population, and at 35-47% among "Champions" between 2011 and 2015. Id. at ALZ0058618. For AFA, those numbers were 15-19% and 25-32% respectively. Id. at ALZ0058619.
The Association owns a number of marks containing the term "Alzheimer's Association," including the standard character mark ALZFIEIMER'S ASSOCIATION®, U.S. Reg. No. 2,850,223 (the "'223 Mark"), which has been registered since June 8, 2004, but which has been in use since 1988. FOF ¶¶ 5-7. The Association also owns U.S. registrations for marks using the two-word name "Alzheimer's Association" along with other words or with graphical elements, as well as standard character marks for WALK TO END ALZHEIMER'S® and MEMORY WALK®. See FOF ¶¶ 8-9. Together, these marks are hereafter referred to as the "Association *268Marks," although the '223 Mark is most central to the present litigation. The walk-related marks refer to an event organized by the Association that raises awareness and funds for Alzheimer's care, support, and research; in 2016, nearly 500,000 participants took part in walks in 630 communities, raising more than $78.6 million. FOF ¶¶ 14-15. The Association maintains a website accessible at www.alz.org, on which it displays "alz.org" and "Alzheimer's Association" at the top of its landing page. AFA FOF ¶ 15. The Association's principal color is purple. AFA FOF ¶ 16.
The Association did not register, and has no rights to the terms "Alzheimer's" or "Association" standing alone. AFA FOF ¶¶ 82, 126.
B. Counterclaim Defendant Alzheimer's Foundation of America and its Marks
The Alzheimer's Foundation of America was founded in 2002 with the primary purpose of providing optimal care and services to individuals living with Alzheimer's disease and to their families and caregivers. AFA FOF ¶¶ 17-19. Since January 1, 2014, Charles J. Fuschillo, Jr. has served as AFA's President and CEO. Ex. DS ¶ 3. AFA has more than 2,600 member organizations throughout the country that collaborate on education, resources, best practices and advocacy to meet the needs of those with Alzheimer's and related illnesses and their families and caregivers. AFA FOF ¶ 20. AFA has awarded millions of dollars in grant funding to its member organizations for services such as respite care. AFA FOF ¶ 21. In 2010, AFA's "revenues, gains and other support" from "contributions and special events including telethons" was approximately $6.6 million. Ex. 118 at 59. AFA maintains a website at www.alzfdn.org, and principally uses the colors teal and white. AFA FOF ¶ 26.
AFA owns various composite and standard character trademarks. The organization's first registered mark-registered in 2006-is for "AFA Alzheimer's Foundation of America" with the organization's "heart in hands" logo, U.S. Reg. No. 3,179,067 (the '067 Mark), as shown below. Ex. AK.
In December 2009, AFA filed for the assumed or "doing business as" two-word name "Alzheimer's Foundation" in four jurisdictions. Exs. 29-31, 33. Before December 2009, AFA had not applied for any trademarks using the two-word name without "of America." FOF ¶¶ 41, 155. At that time, AFA also filed for a composite mark using the two-word moniker "Alzheimer's Foundation" and displaying its heart in hands logo. AFA now owns this composite mark, U.S. Reg. No. 3,948,014 (the '014 Mark), as shown below. Ex. AL.
AFA did not use the '014 Mark in commerce until October 2010. FOF ¶ 39.1. AFA disclaimed any exclusive right to the literal element of the '014 mark as well as all literal elements of the '067 composite mark. FOF ¶¶ 39, 154; Ex. AK.
Finally, and most crucially for the issues at bar, on July 3, 2014, AFA filed for a standard character mark for "Alzheimer's Foundation." FOF ¶ 208. On December 23, 2014, this mark was registered on the supplemental USPTO register as U.S. Reg. No. 4,661,324 (the '324 Mark). FOF ¶ 209;
*269Ex. AM. According to the registration, the first use in commerce of the two-word name dates to March 31, 2004. Ex. AM.
AFA claims to have identified itself by the two-word moniker "Alzheimer's Foundation" since its inception. See Tr. (Fuschillo) 47:9-48:1; Tr. (Hall) 555:9-14; Tr. (Steinberg) 564:9-12. Prior to 2009, AFA's uses of the two-word name were largely limited to the headlines of press releases, and to sponsored advertising online, although AFA did identify singular instances of use on a billboard in Times Square, in its Care Advantage magazine, and in a radio ad. Exs. 19, 97, BR-1-BR-4, CO, CP; Tr. (Fuschillo) 30:8-25, 34:16-20. AFA has also used the two-word name on its Twitter account, which is space limited. FOF ¶ 232. Throughout AFA's history, the most pervasive use of "Alzheimer's Foundation" without "of America" has come in its online advertising. AFA FOF ¶ 73.
The Association does not claim any rights to the term "Alzheimer's Foundation," and does not use the phrase to refer to itself in any of its ads nor on its website. AFA FOF ¶¶ 82-83.
C. Keyword Advertising and Competitor Campaigns
Search engines, such as Google and Bing, use algorithms to identify various web pages relevant to the search terms input by users. FOF ¶¶ 46, 48. Search engines list both "organic," or unpaid, and "sponsored," or paid, search results. FOF ¶ 47. The search engine results page displays sponsored ads first, followed by organic listings. AFA FOF ¶ 38. The term SEO, or "search engine optimization," describes techniques used to enhance a website's appearance in organic search results. FOF ¶ 49. Metatags, including keyword metatags, are tools used in SEO, although evidence suggests that Google abandoned using metatags in their search algorithm by 2009. Ex. DV ¶¶ 102; Tr. (Leone) 789:1-791:2. Metatags are contained in a website's coding and are generally not visible to users unless they use a function to view the web page code. Ex. 258 ¶ 8.
In marketing, the term "keywords" generally refers to search keywords, which are terms purchased by advertisers so that their ads will appear in sponsored results in response to certain searches. FOF ¶ 52. Google's advertising program is called "Google AdWords." Each time a sponsored ad appears on a results page, this counts as one impression. AFA FOF ¶ 37. Advertisers pay for their paid ads based on the number of times a user "clicks" on the ad and goes to the advertiser's website. AFA FOF ¶ 39. Google and Bing require that the advertiser's URL appears in the ad. AFA FOF ¶ 54. Paid ads are labeled as "ads" or "sponsored" in some manner on the search results page. AFA FOF ¶ 40.
Google and Bing conduct real-time auctions 24/7 in which advertisers or their agents bid on keywords that will result in having their online ads displayed on a search engine results page. AFA FOF ¶¶ 41-42. The bidding is not visible to the public, users do not know which keywords organizations have bid on, and the bidding is largely non-exclusive. AFA FOF ¶¶ 43-44. Google and Bing do not limit the number of advertisers that can bid on a keyword or the number of advertisers who may have their ads displayed in response to a search query. AFA FOF ¶ 51.
While Google and Bing each uses a proprietary algorithm to determine the order in which paid ads appear, generally, the higher the bid the more likely the bidder's online ad will appear at or near the top of the sponsored search engine results. AFA FOF ¶ 45; Ex. 258 ¶ 11. The proprietary algorithms are also reported to use relevancy in its calculations, the score for which may be improved by using the keywords entered by the user in the organization's *270paid ad or on its website. AFA FOF ¶¶ 46-47. SEM, or search engine marketing, refers to the process of buying keywords so that an advertiser's sponsored link appears prominently in response to particular searches. FOF ¶ 55.
Since prior to 2007, Google has not restricted the use of trademarks as keywords, even in cases in which it receives a complaint from the trademark holder. AFA FOF ¶ 48. Bing has a similar policy regarding trademarked keywords. See Ex. DV ¶ 43; Ex. DU ¶ 37. Google and Bing have never prevented AFA from purchasing Association Marks as keywords. AFA FOF ¶ 52. Moreover, advertisers often bid on and purchase competitors' trademarks as keywords. AFA FOF ¶ 55.
D. Parties' History of Online Advertising
Both AFA and the Association have used online marketing to attract a broader audience to their websites. AFA FOF ¶ 74. Part of AFA's objective in driving traffic to its website was to increase "conversions"-or online donations. See, e.g., Ex. 60. To that end, AFA retained consultants, including Correlative, to advise them on what keywords to purchase and to create the content of online ads. AFA FOF ¶ 78.
Since June 7, 2004, AFA has described itself in the header of online ads using the two-word name "Alzheimer's Foundation." AFA FOF ¶ 79. Ads using this two-word name resulted in over 30 million impressions between June 7, 2004 and the end of 2009. AFA FOF ¶ 80. AFA understood that Google limited the number of characters an advertiser could use in the title of its paid ad to 25 or 30. FOF ¶ 59. Evidence at trial did not conclusively establish Google's current policy regarding character limitations. See Tr. (Marquardt) 905:11-15. From 2010 to date, AFA has spent $300,000 to $500,000 annually to purchase keywords. AFA FOF ¶ 86.
From June 2004 to the present, AFA has also purchased Association Marks as keywords. FOF ¶¶ 68, 146. By late 2007, AFA was frequently purchasing "alzheimers association" and variants as search keywords. FOF ¶ 149. For part of this time-April 2012 to June 2014-AFA ran sponsored ads that used the word "Association" in the text. FOF ¶ 71. After correspondence from the Association's counsel, AFA discontinued the use of the word "Association" in the content of its ads, FOF ¶ 72, however AFA continues to purchase Association Marks as keywords and to advertise itself as "Alzheimer's Foundation" in responsive ads. FOF ¶¶ 73, 204.
The Association began its own keyword marketing program in April 2007. FOF ¶ 65. The Association spent approximately $150,000 annually on keyword advertising in 2015-16, but also utilized Google Grants for an additional $480,000 of keyword advertising. Tr. (Carson) 292:11-293:8. Initially, the Association purchased "Alzheimer's Foundation" as a keyword, but all foundation-related keywords were removed from the Association's SEM in August 2010. Exs. BU, CB. The Association did not believe it was acting improperly in purchasing foundation-related keywords. Tr. (Geiger) 874:25-875:18. The Association has never purchased as an SEM keyword the search term "Alzheimer's Foundation of America." FOF ¶¶ 245, 247. In August 2016, AFA's trademark counsel wrote to the Association to complain that Alzheimer's Association's Northern California and Northern Nevada chapter had purchased "Alzheimer's Foundation" as a keyword and displayed resulting advertising text with "Alzheimer's Foundation" in the title. Ex. 15. The chapter had recently merged with the Alzheimer's Association, FOF ¶ 237, and upon receipt of the letter, the chapter discontinued purchase of the keyword. FOF ¶ 240.
*271Presently, one of the at-issue ads run by AFA might show "Alzheimer's Foundation" in the ad title, in standard typeface without any graphic element, followed by the associated website URL and a brief description or tagline. For example, as can be seen below, a Google search of "alzheimer's association" from June 2014 shows AFA's ad as the top result, with the main header as "Alzheimer's Foundation-alzfdn.org" followed by a yellow marker reading "ad," the URL "www.alzfdn.org" and a tagline of "An Association of Care and Support. Reach Out to Us for Help...." See Ex. 71. The advertisement also contains links to pages within AFA's website for those looking to "Make a Donation," to learn "About AFA," or to "Reach Out to Us for Help." Id. The second and only other ad result is from the Association. Their header reads "alz.org-Alzheimer's Association," is also followed by the yellow "ad" marker and the URL for their donation page, "www.alz.org/donate." Id. The Association's tagline was "Honor a Loved One with a Tribute Donation-Support Research & Care" and included links to subtopics on their website. Exhibit 71 is reproduced here:
Other trial exhibits showing Google search results for "alzheimer's association" might have the ads in a different order, or might show different text used by each *272organization, but consumers are presented with the same basic context shown above. See, e.g., Exs. 90, 161, 212; see also Ex. 97 (listing the various headers and textual elements of AFA's advertisements). The parties introduced other examples in which AFA used "Alzheimer's Foundation" in its ad text. See, e.g., Exs. 148-58, 245.
Purchasing keywords such as "Alzheimer's Association" was part of the AFA's so-called competitors campaign. Ex. 49. In October 2011, Chris Leone, AFA's SEO consultant, recommended increasing the daily budget for the competitor's campaign, and AFA agreed. Exs. 50, 51. At times, the keyword "Alzheimer's Association" received more clicks on AFA's sponsored ad than those received from its own brand. FOF ¶ 174. In early 2012, Mr. Leone reported that campaigns targeting AFA's competitors performed the best. Exs. 54, 55. The competitors campaign comprised roughly 40% of AFA's keyword marketing budget. Ex. 60 at AFA0021419. AFA also worked with a vendor named Vizions on an SEO plan involving keywords such as "alzheimers association," Exs. 42, 43, but there is no specific evidence in the record that AFA adopted Vizions' recommendations.
AFA rejected recommendations from Mr. Leone to include Association Marks in the content of AFA's online ads or website. Ex. DU ¶¶ 68-74; Tr. (Leone) 775:14-778:17; Tr. (Steinberg) 568:24-570:5; Exs. 63, 96-98. However, as described above, AFA did run an ad with the text "An Association of Care and Support," Ex. 66, which AFA removed after receiving a complaint from the Association. Ex. H; Tr. (Foh) 345:20-346:16; Tr. (Johns) 147:17-150:3; Ex. DS ¶ 61. The ad used the words "Alzheimer's" and "Association" but they were not next to each other. Ex. DS ¶ 60; Tr. (Foh) 345:11-18; Tr. (Leone) 778:18-779:9.
At trial, Mr. Leone opined that the "common strategy" of competitor-targeted ads was effective here because "Alzheimer's Association is the larger organization" and "if someone is searching for Alzheimer's Association, which they are frequently, then we're providing a[n] ad as another option for them. It is the same thing that most competitors do." Tr. (Leone) 797:2-13. Mr. Leone did not recall AFA ever using the word "compare" in the content of its ads resulting from Association-related keywords. Tr. 791:18-21.
The Association also incorporates AFA's metatag coding into its arguments in this case. The Association claims that from early 2005 through spring 2008 and then again starting in mid-2011, AFA was coding its website with metatag keywords including Alzheimer's Association, Alzheimer's and Related Disorders Association, memory walk, and Maintain Your Brain-all terms used and trademarked by the Association. See Ex. 240B. Exhibit 240B, on which the Association relies in making this argument, was admitted at trial solely "for what it may permissibly be considered for." Tr. 988:3-5. AFA fairly notes that 240B does not show how or when AFA's website was coded and no witness was called to explain the meaning of 240B. Tr. 986:22-988:2. Regardless, the Court finds that AFA's metatag practices are largely immaterial to the present issue, as they have likely not been used by search engines since before 2009, and so have little effect on the ordinary prudent consumer. See Exs. CW; DV ¶¶ 99-102; Tr. (Leone) 789:1-791:2; see Network Automation, Inc. v. Advanced Sys. Concepts, Inc., 638 F.3d 1137, 1146 n.3 (9th Cir. 2011) ("Modern search engines such as Google no longer use metatags. Instead they rely on their own algorithms to find websites."); Ascentive, LLC v. Opinion Corp., 842 F.Supp.2d 450, 467 (E.D.N.Y. 2011) (citing *2734 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 25:69 (4th ed. 2013) ).
In order to make an online donation to AFA, a consumer could not simply enter a keyword or click on an online ad, but would have to visit AFA's website. AFA FOF ¶ 109. If after looking at AFA's website, a computer user determines that AFA is not the entity to which she wants to donate, the user can click back to return to her original screen page. Ex. DV ¶ 79; DU ¶ 93. To continue on to the donation page, a user would have to click on "Donate" or "Support Us" within the AFA website. AFA FOF ¶ 112. On the donation page, there are many references to "AFA" or "Alzheimer's Foundation of America," and no references to the Association or use of any the Association Marks. AFA FOF ¶ 113. At no point while on the AFA website during the donation process would a consumer see any of the Association Marks. AFA FOF ¶ 115. A similar multi-step process would be required to make an online donation to the Association. AFA FOF ¶ 116.
E. Non-Study Based Evidence of Confusion
The Association presents various types of evidence of purported confusion. First, the Association cites evidence that AFA has previously acknowledged confusion and, in fact, was the first to complain of confusion. In 2004, AFA's CEO at the time drafted a letter to the Association's then-CEO stating "I was therefore quite surprised when a routine web search under 'alzheimers foundation' led me to [the Association's] site. 'Alzheimer's Foundation of America' is a registered service mark of our organization. It distresses me that supporters of our respective organizations may be confused when searching the internet." Ex. 20. No witness was called to explain the letter, which may not relate to sponsored ads, and thus may be irrelevant to the purported confusion at issue here. Additionally, in AFA's suit against the Association, AFA alleged that the Association's "unauthorized conversion of checks made payable to Alzheimer's Foundation, constitutes an unauthorized use of Alzheimer's Foundation's Marks and is likely to continue to cause Plaintiff damage by confusing the general public...." FOF ¶ 106.4 (quoting Ex. 36 ¶ 61). The confusion caused by check-related practices is also not at issue here. Finally, in 2014, AFA's Kristen Leesment wrote to CEO Charles Fuschillo that she was "surprised" by the results of a survey AFA was taking of online donors that showed many people saying they donated before or were "introduced to AFA through a fund raising event," and speculated that "several respondents may have us confused with the Association or with 'the cause.' " Ex. 70 at AFA0021358.
Second, the Association points to check-related evidence. In 2009, in connection with litigation in Virginia, AFA learned that the Association receives checks made out to "Alzheimer's Foundation" and that it cashes those checks. FOF ¶ 151. Between 2007 and 2012, Alzheimer's Association received more than 5,700 checks made payable to "Alzheimer's Foundation" or a variant totaling over $1.5 million. Ex. CI. The Association notes that it receives many more checks made out to "Alzheimer's Foundation," than to "AFA" or "Alzheimer's Foundation of America." FOF ¶¶ 111-12.2. AFA, in turn, received more than 5,000 checks between 2006 and June 2016 made payable to "Alzheimer's Association" or near variants. FOF ¶ 107. While AFA receives checks made payable to a number of erroneous names, the most common error is checks payable to Alzheimer's Association. FOF ¶ 108. AFA's documents suggest that a large percentage of AFA's online donors are first-time donors, and that the average online donation, as well as check donation, is under $100.
*274See Ex. 60 at AFA0021415; FOF ¶¶ 262-63.
The parties have substantial disagreements about the value of this evidence with respect to the quantity of confusion and with respect to any causal links between the mislabeled checks and AFA's actions.
AFA claims that the number of checks received by the Association from January 2009 to January 2012 that were payable to AFA was only 0.1% of the total number of checks received, and 0.252% of the total value of checks received, but the Association disputes these figures as misleading. AFA FOF ¶ 154. The Association concedes that "the checks so received were a small percentage of the total number and value of checks received during the periods identified." Id. Similarly, Mr. Fuschillo opined that the number of checks AFA received annually that were made out to Alzheimer's Association was "de minimis," but could not identify how many total checks AFA received annually or the number made payable to the Association, and had done no personal review of these records. Tr. 63:9-74:15. Consequently, his opinion carries little weight.
In one instance, on July 30, 2014, the Association's cashiering department brought Christine Foh, the Association's general counsel, checks made payable to the Association and an obituary indicating donations were to be sent to the Association that were received alongside a printout of AFA's internet donation form. FOF ¶ 115. The Association followed upon the issue, with Ms. Foh calling a donor listed on the AFA form. FOF ¶ 116-25. The donor confirmed that the donation was intended for the Association. FOF ¶ 126. When asked why she had submitted the AFA form, the donor claimed that she had typed the Alzheimer's Association name into her web browser and clicked on the first entry that came up to download the form for donations and that she was unaware that the form she downloaded was for AFA or that she was on the AFA website. FOF ¶¶ 127-28. Aside from this one instance, the Association did not provide further evidence-anecdotal or otherwise-that directly confirmed that AFA's practices in purchasing Association Marks as keywords or in using the two-word name "Alzheimer's Foundation" caused the check-related confusion nor did the Association provide evidence that confirmed donors' intent one way or another. Moreover, no witness testified about other donors or potential donors contacting one organization looking to donate to the other, apart from people asking about the difference between the Association and AFA. See Tr. (McCullough) 585:2-6; Tr. (Geiger) 857:12-20; Tr. (Steinberg) 573:1-574:20. Third parties have also testified that they were not aware of or did not remember instances of confusion between the Association and AFA. AFA FOF ¶¶ 140-41.
Third, the Association points to an assortment of purported incidents of actual confusion, none of which have a clear link to any of AFA's allegedly infringing actions. For example, the parties engaged in a legal dispute concerning the estate of Sandra Lynch in which Ms. Lynch's will left a bequest to "Alzheimer's Foundation, Pennsylvania Chapter" even though the Association had a Pennsylvania chapter and AFA did not. FOF ¶¶ 129-32. Nonetheless, no evidence was adduced that Ms. Lynch's confusion was born from any of AFA's behaviors at issue here. See Ex. 11. Additionally, the Association presented evidence of NBC's Today show running a quotation the Association has provided alongside AFA's logo. FOF ¶ 134; Ex. 13. However, the logo mistakenly used by NBC was the '067 Mark with the words "of America," and is not at issue in this litigation. The Association also submitted a video clip from Celebrity Family Feud in which the host suggested the show was *275raising money for "Alzheimer's Foundation," when it was actually raising money for the Association. FOF ¶ 133; Ex. 12. Again, no evidence was presented that explained the reason for the mistake, let alone evidence connecting it to AFA's actions.
F. Dr. Van Liere's Studies
In addition to the evidence presented above, the Association also presented the survey evidence and opinion testimony of Dr. Kent Van Liere, Ph.D., Vice President at NERA Economic Consulting. FOF ¶ 135; Ex. 262. Over the course of 30 years, Dr. Van Liere has designed and analyzed hundreds of surveys focused on marketing related issues, including likelihood of confusion surveys, and his testimony has been admitted by a variety of courts. FOF ¶ 136.
Van Liere and his firm conducted two studies, both of which were the subject of a Daubert motion brought by AFA to exclude Van Liere's testimony. AFA did not challenge Dr. Van Liere's credentials as an expert. Instead, AFA hired its own expert, Hal Poret, to review and provide opinions regarding Van Liere's. surveys and conclusions. AFA FOF ¶ 170; Ex. DX; Ex. DY. Mr. Poret did not conduct his own survey. The Court denied AFA's Daubert motion at the end of Dr. Van Liere's examination, Tr. 740:6-10, and again at the close of trial, deciding instead to "[take] into account all of the record evidence." See Tr. 996:24-997:2.
1) Study 1
In his first study, Dr. Van Liere conducted a lineup, or "Squirt" style study meant to gauge the likelihood of confusion between the standard character marks "Alzheimer's Association" and "Alzheimer's Foundation." FOF ¶ 139. Study 1 operated as follows: Survey respondents were shown the Alzheimer's Association standard character mark. Then, after a "distractor" question, they were shown either a list of names of charitable organizations that included "Alzheimer's Foundation" (test condition), or a list of names of charitable organizations that included "Alzheimer's Trust" (control condition). Ex. 262 ¶ 15. Respondents were then asked questions meant to measure whether they believed the Alzheimer's Foundation or Alzheimer's Trust was the same as, or affiliated with, the first organization they were shown-namely, "Alzheimer's Association." Id. ¶ 16; see Ex. 191 (screen prints of Study 1). In the test condition, 50% of respondents believed that Alzheimer's Foundation was the same as, or affiliated with, Alzheimer's Association. Ex. 262 ¶ 23. In the control condition, only 16% believed so about Alzheimer's Trust. Id. As a result, Study 1 was said to demonstrate 34% net confusion. Id. Based on this result, Dr. Van Liere opined that "AFA's use of the name 'Alzheimer's Foundation' is likely to confuse a substantial proportion of consumers in the relevant population into believing that 'Alzheimer's Foundation' is the same as, or is affiliated with Alzheimer's Association." Ex. 262 ¶ 33; FOF ¶ 138.
Through Mr. Poret's testimony, AFA makes three criticisms of Study 1. First, it alleges that Van Liere failed to use the most relevant universe of consumers. Ex. DX ¶¶ 77-104. Van Liere determined that the relevant universe for testing should be people who either donated in the recent past or were likely to donate in the near future to "charitable organizations focused on Alzheimer's disease." Ex. 262 ¶ 10; Tr. (Van Liere) 641:7-12. To qualify for the study, prospective participants could have either answered that they had donated to an Alzheimer's organization within the past 12 months or that they would consider donating to an Alzheimer's organization within the next 12 months. AFA FOF ¶ 214. The screening data for Study 1 showed that at least 85% of the respondents *276for the Study 1 qualified on the basis of a potential future donation, and not on the basis of any past donation. AFA FOF ¶ 215. Moreover, roughly 65% of respondents in Study 1 said they would consider donating to at least three of the four types of charitable organizations about which they were questioned (the others being breast cancer, animal rescue, and homelessness). AFA FOF ¶ 217; Ex. DX ¶¶ 80, 91. Mr. Poret concluded that "[t]here is simply nothing in the data that suggests that most respondents are anything more than members of the general public who might theoretically consider donating to any of a number of charities." Ex. DX ¶ 102.
While Poret's criticism is well-taken, the Court sees no substantial error in the definition of the survey universe. Relevant consumers who would search for Alzheimer's Association on the internet would not only include those who have made donations in the past year. And while it is likely that respondents were inflating their willingness to give in the near future, it would be manifest error to exclude those who happened to also be considering a donation to a homelessness charity, for example. Finally, a stricter screen of prospective participants may not have substantially altered the results given the relatively low awareness of either charity.
Second, AFA argues that Study 1 fails to replicate marketplace conditions. Study 1 showed respondents the bare words "Alzheimer's Association" and "Alzheimer's Foundation," with both phrases in plain, black and white, and bolded text. AFA FOF ¶ 173. AFA argues that in the real world, consumers would see color, stylization, and additional content that surrounds and defines those terms, and so Van Liere's word association bears no relationship to marketplace conditions. See Ex. DX ¶¶ 110-126; Dkt. No. 287 at 18. The Association responds that since the two marks would appear in close proximity, in large, plain text, and without any accompanying logos or colors in the online search context most at issue, a study of this nature is appropriate to test for confusion between the marks. Ex. 262 ¶¶ 15, 18-19, Tr. (Van Liere) 664:24-665:7. Still, AFA notes that the marks are still seen with accompanying content or context, which is different for each mark, including different URLs and descriptive words. See Exs. AX, AY, AZ, BA, DL, DQ; Ex. DX ¶¶ 111-16.
While it is true that in instances with competing ad results, consumers would see both marks simultaneously, and surrounded by distinguishing content, the Court cannot conclude that Van Liere's word association test, however artificial, is unreliable or lacks probative value as a result. That the comparison is between the two marks outside the context relevant to this litigation does lessen its probative value in determining actual confusion, but it does not negate its value to the general analysis of whether AFA's actions are likely to confuse consumers.
AFA's third major criticism of Study 1 more clearly militates against placing great weight on the survey results. AFA questions Van Liere's choice of and process in choosing "Alzheimer's Trust" as a control, arguing that the choice artificially inflated the net confusion numbers. The Court agrees.
A control term should "share[ ] as many characteristics of the test stimulus as possible, with the key exception being the characteristics whose influence is being assessed." AFA FOF ¶ 192. A control term that generates higher levels of confusion would yield lower net confusion when compared to the test stimulus-"Alzheimer's Foundation."
Dr. Van Liere spoke with his staff about what a proper control would be. AFA FOF ¶ 191. Ultimately, Dr. Van Liere selected *277"Alzheimer's Trust" as the control for the lineup survey. AFA FOF ¶ 193. NERA staff had "pre-tested" two controls, "Alzheimer's Charity" and "National Alzheimer's Foundation," which, by generating more confusion, would have yielded net confusion rates for Alzheimer's Foundation of 12% and 11% respectively, far lower than the 34% Van Liere computed using "Alzheimer's Trust." Ex. DX ¶¶ 44-57, 145, 148-49; Exs. DG, DI. Dr. Van Liere considered using "Alzheimer's Charity" as a control, but switched to using "Alzheimer's Trust." AFA FOF ¶ 185. Dr. Van Liere testified that he rejected "Alzheimer's Charity" because it "sounded like a product category rather than a single real entity." Ex. 262 ¶ 22.
Dr. Van Liere testified that his staff never disclosed the results of these pre-test surveys to him, even though the results were disclosed to the Association's attorneys. AFA FOF ¶ 184. He also claimed to be entirely ignorant about the methodology employed by NERA staff in conducting the pre-test studies. AFA FOF ¶ 183. At trial, the Court asked Dr. Van Liere: "Do you understand your ultimate product to ever have been influenced in any way by what's done in the pretest studies?" Dr. Van Liere answered: "That's a hard question to answer, the way you've asked it. Certainly I design the work. I do the work. I oversee the work. I say yes or no to how we're going to ultimately do my study. That's how I do it. I can't really kind of evaluate influences, either from other studies we've done in other matters or other studies that staff have worked on for other people. I wouldn't want to make a general statement about that, how that influence passes." Tr. 698:9-19. Based on the substance of his answers, and his demeanor on the witness stand, the Court did not find Dr. Van Liere candid on the subject of the influence of pretest surveys. While it is legitimate to run a pre-test or pilot study for the purposes of improving a study, no credible explanation was offered for the changes made between the pre-test survey and the reported survey, and this suggests a potentially improper purpose. See Louis Vuitton Malletier v. Dooney & Bourke, Inc., 340 F.Supp.2d 415, 450 n.196 (S.D.N.Y. 2004), vacated in part on other grounds, 454 F.3d 108 (2d Cir. 2006).
But even putting this issue aside, "Alzheimer's Trust" is a weak control term. Dr. Van Liere testified that he and his staff chose "Alzheimer's Trust" because they wanted a two-word combination that sounded like a real organization but that was not actually a real brand in the marketplace, and not for any reason related to the pre-test surveys. Tr. (Van Liere) 720:24-721:11. Having a two-word control term was important to filter out any inflation of confusion that might result if "Alzheimer's Foundation" were the only two-word name on a list being compared with the two-word mark "Alzheimer's Association." Tr. (Van Liere) 727:10-728:6. Nonetheless, the presence of the word "Trust" as a descriptor for a charity, is both more unique than "Foundation," and more easily distinguished from "Association." As AFA notes, "trust" is an ambiguous word that also denotes other meanings, such as "to have confidence in" or "estate." Tr. (Van Liere) 687:15-688:5; Ex. DX ¶¶ 129, 133-35. For example, "Alzheimer's Federation" is also a two-word name that sounds like a real organization but is not, and that could more clearly control for the confusion created by reasons other than Foundation-specific reasons. Dr. Van Liere could have also used "Alzheimer's Foundation of America" as a control. The Association's contention is that AFA's use of the two-word mark in particular causes consumer confusion, so what better way to test that proposition than to compare its use with that of AFA's full, and undisputedly non-infringing name?
*278While the Court recognizes that the perfect control may not exist, as a measure of consumer confusion between two marks, Study 1-a test of mere words, shown sequentially, and with the unique word "trust" in the control term-is not particularly persuasive, especially in light of evidence that other potential controls, including those comprised of three or five words ("National Foundation for Alzheimer's Research" and "Alzheimer's Awareness Foundation") and one that sounds artificially generic ("Alzheimer's Charity"), registered substantially higher levels of confusion. See Ex. DX ¶ 21; Ex. DG; Tr. (Van Liere) 688:6-689:2, 693:24-696:5.
2) Study 2
The second study Dr. Van Liere and his staff performed was designed to replicate the process consumers go through in searching "Alzheimer's Association" and viewing the results. The Association argues that its results show that consumers are both initially confused by the results, and that this initial interest confusion is not dispelled when they are taken to AFA's home page. Defendant and Counterclaim Plaintiff Alzheimer's Disease and Related Disorders Association, Inc.'s Post-Trial Memorandum of Law ("Post-Trial Br."), Dkt. No. 345, at 30.
In Study 2, respondents were asked to type "Alzheimer's Association" into a search box in either Bing or Google and then were either shown a page of results including AFA's disputed ads and other keyword ads (test condition), or a page of organic search results without ads (control condition). FOF ¶ 249. Respondents were asked to click on the link or links they thought would take them to the website of the organization for which they searched. Id. Those who clicked on AFA's ad were then shown AFA's web page and asked if they thought it was the web page of the organization for which they had searched. Id. If respondents did not select the "Alzheimer's Foundation" or control links, the process repeated, this time asking the respondents to click on the link or links, if any, that they believe would take them to the website of an organization affiliated with the organization they searched for. FOF ¶ 250.
In the test condition, 42% believed a sponsored ad for "Alzheimer's Foundation" was the same as or affiliated with the "Alzheimer's Association." Ex. 262 ¶¶ 3, 31, 34. In the control condition, only 22% expressed confusion, yielding a net rate of 20% about the sponsorship of the links. Id. Moreover, of the 42% who were confused in the test condition, 74% remained confused as to source or affiliation after viewing a screenshot of AFA's web page (or 31% of the total respondents in the test condition). Id. As a result, Dr. Van Liere expressed the opinion that AFA's keyword advertising campaign using Association Marks led to a likelihood of confusion at both the initial interest page and at the landing page. Id.
AFA lodges four main criticisms of Study 2. The first regards how Van Liere defined the survey population. As the Court noted above, although the population may not be a perfect reflection of those consumers who are likely to search for or donate to the Alzheimer's Association, neither is it substantially erroneous.
Second, AFA criticizes the internet search survey for failing to replicate market conditions. AFA argues that the survey did not ensure that respondents were aware of the Association when taking the survey. Rather, the study design assumes that participants asked to type in "Alzheimer's Association" know that it is a particular organization. Consumers cannot mistakenly associate the AFA ad or website with the Association if they were not aware of the Association's existence to begin with. Alzheimer's Foundation of America, *279Inc.'s Post-Trial Memorandum of Law ("AFA Post-Trial Br."), Dkt. No. 348, at 31-32; Ex. DX ¶¶ 171-76. To the extent that consumers in the real world are searching for "Alzheimer's Association" as a generic category, and do not have the Association in mind when searching, they cannot be confused by AFA's ad or webpage.
Upon the Court's evaluation of the language used in the study, Dr. Van Liere's asserted belief that "people understood that they were searching for a specific organization," is fair and accurate. Tr. (Van Liere) 705:7-706:2. Nonetheless, the fact remains that consumers entering those search words in the real world may not have that particular understanding. Thus, while this criticism does not undermine Study 2, it does lessen the results' applicability to the task at hand.
Third, AFA criticizes the test stimulus as biased. AFA notes that all respondents presented with the test stimulus were shown a search results page in which the first sponsored ad was that of AFA, and there was no sponsored ad for the Association. AFA FOF ¶ 204. In real life, the Association also bid on its own Marks as keywords, and often would have appeared in the sponsored ads section as well. See Exs. 71, 161. Additionally, the order of the ads would change based on the organizations' respective bids. AFA charges that choosing a test condition without the Association's own ad, and failing to rotate the test condition so that sometimes another organization's ad appeared first, created a bias in favor of choosing the AFA ad-the first one listed-as the organization they had searched for. Tr. (Van Liere) 710:9-711:22; Ex. DX ¶¶ 199-207, 224-225; Ex. 188 at 11; Ex. 90. The bias created by the order of options is well-established. See, e.g., Jon A. Krosnick & Duane F. Alwin, An Evaluation of a Cognitive Theory of Response-Order Effects in Survey Measurement, 51 Pub. Opinion Q. 2, 201-219 (Summer 1987); Dana R. Carney & Mahzarin R. Banaji, First Is Best, PLOS One (June 27, 2012). Further, the study's control, as discussed below, does not account for this bias.
However, while failure to rotate the ad results did create a bias that suggests the actual confusion percentage would be lower, Study 2 did use a real web page with real ads that did actually result from typing in "Alzheimer's Association" at one point in time. See Ex. 188 at 11; Ex. 90. Consequently, while the net confusion rate might be inflated based on the failure to rotate the order of the ad results,7 if the Court were to find a likelihood of confusion based on the one test stimulus used, it would constitute a Lanham Act violation just the same.
Fourth, and finally, AFA criticizes Study 2's control stimulus. Instead of using a control with sponsored ads altered to remove the alleged infringing elements-namely, by removing AFA's ad altogether or its use of the two-word "Alzheimer's Foundation"-Study 2 used the same Google or Bing search results with the sponsored ads entirely removed. AFA FOF ¶ 207; see Ex. 188 at 12. A control respondent was deemed "confused" if she clicked on a Wikipedia link or news link instead of on the link for the Association's webpage itself.
The Association points to Study 2 for support for its contention that the initial interest confusion Van Liere claims to have found was not dispelled when confused consumers were taken to AFA's home page. Ex. 262 ¶ 31. AFA argues that *280using a Wikipedia page or a news article as a control to measure noise is flawed because it would not control for the tendency of respondents "to answer that a webpage for an Alzheimer's organization is a web page for the term they searched for." Ex. DX ¶ 28; AFA Post-Trial Br. at 33. This criticism, while valid, fails to undermine the central finding of Van Liere's study; after clicking on the wrong link, the AFA landing page (with its logos, text, and unique color) does little to correct any preexisting confusion.
However, Study 2's control stimulus was flawed for AFA's second proffered reason. By failing to control for the presence of sponsored ads, Study 2 ends up inflating any net confusion because the first search result in the control is the Association's website, that is, the correct answer, while the first search result in the test condition is the allegedly infringing AFA advertisement. Ex. DX ¶¶ 224-36. A better control stimulus would have contained non-infringing sponsored ads at the top that, if clicked, counted towards the confusion rate for the control. Dr. Van Liere testified that a control should "share[ ] as many characteristics of the test stimulus as possible, with the key exception being the characteristics whose influence is being assessed." Tr. (Van Liere) 654:18-23. He defends his choice by broadly defining the "key elements at issue" as "the sponsored links," and stating that a proper control is identical to the test condition but for that issue. Ex. 262 ¶ 30; Tr. (Van Liere) 723:22-724:15. A more accurate description of what is "at issue" is the allegedly infringing AFA ad, and not sponsored ads in general. Accordingly, only the offending item should have been removed or replaced.
* * *
In sum, Study 1 was designed to measure whether use of the two-word mark "Alzheimer's Foundation" was likely to cause confusion and Study 2 was designed to test whether AFA's sponsored link advertisements on search engines when consumers search for the Association was likely to cause confusion. Ex. 262 ¶ 9. In light of the deficiencies discussed above, the Court cannot conclude that the results from the studies constitute credible evidence of consumer confusion. In particular, Van Liere's decision to use the controls of "Alzheimer's Trust" and a results page without sponsored ads in Studies 1 and 2 respectively undermines the persuasiveness of his results. Ultimately, while the Court denied Counterclaim Defendant's Daubert motion, the evidence in the record and revealed at trial strongly militates against placing much weight on the study evidence.
G. Evidence of Willful Action
The Association marshals many pieces of circumstantial evidence in an attempt to tell a story in which AFA acted willfully, or in bad faith, when adopting the two-word name "Alzheimer's Foundation" and when purchasing Association Marks as keywords.
First, the Association presents a timeline of events. On April 13, 2004, Eric Hall, the then-CEO of AFA, wrote Sheldon Goldberg, then-CEO of the Association, to complain about confusion that may result when a search for "alzheimers foundation" leads consumers to the Association's website. Ex. 20. Nonetheless, by June 2004, AFA was purchasing "Alzheimer's Association" as a keyword. Ex. 97. An email from an outside SEO consultant to Hall in December 2007 suggests that AFA was aware this was a trademarked phrase. Ex. 24.
In 2007, the parties engaged in litigation in Virginia state court concerning which organization was the intended beneficiary of a gift from a trust. FOF ¶ 33. In the litigation, AFA learned that the Association *281had been depositing checks it received made payable to "Alzheimer's Foundation." FOF ¶ 151; Ex. 176. Shortly after the end of the Virginia litigation in December 2009, AFA registered the two-word name "Alzheimer's Foundation" as an assumed name in New York, Texas, Washington, DC, and California. Exs. 29-31, 33. As outlined above, in April 2010 AFA sued the Association for endorsing and depositing checks made payable to AFA, Alzheimer's Foundation, and Alzheimer's Foundation of America, claiming unfair competition and trademark infringement. AFA FOF ¶ 119. AFA made certain statements in its filings suggesting that there was a likelihood of consumer confusion when the Association deposited checks made payable to "Alzheimer's Foundation." See FOF ¶ 106. Around this time, AFA also filed for the '014 Mark, the first mark incorporating the two-word name without "of America." FOF ¶ 153; Ex. 32. These filings were the first time AFA applied for any trademarks or registrations using "Alzheimer's Foundation." FOF ¶¶ 40-41. It was not until fall of 2010, six months after filing its lawsuit, that AFA made its first use in commerce of the '014 Mark. FOF ¶ 167.
The Association suggests that AFA took these actions and others to exploit consumer confusion they knew existed. As described at length above, AFA's SEO plan from 2010 onward included its so-called "competitors campaign," through which it purchased Association Marks as keyword. Exs. 49, 54, 55. AFA ran one ad with the word "association" in the text, although this was discontinued following complaint from the Association. FOF ¶ 72.
The Association emphasizes that AFA's patent counsel filed two false specimens of use as further evidence of AFA's bad faith in holding itself out as "Alzheimer's Foundation" in advertising despite evidence of confusion. In 2014, AFA filed for a standard character mark for "Alzheimer's Foundation" (the '324 Mark), which appears on the supplemental register. FOF ¶ 209; Ex. AM. In its application, AFA's patent counsel included as a specimen of use a page from its magazine in which the words "of America" had been obscured. Ex. 73 at 11-13. Kristen Traina, an associate working with Mark Ingber, AFA's trademark counsel, testified that she erroneously included this specimen and that she had highlighted "of America" with a yellow highlighter to mark it for exclusion. FOF ¶¶ 211-12. After the Association raised the issue of the false specimen, Ms. Traina filed press releases as substitute specimens of use, which were accepted by the USPTO. Tr. (Traina) 409:11-412:8, 416:21-417:11.
Then in March 2017, AFA filed a "section 8" declaration with the USPTO claiming ongoing use in commerce with respect to AFA's '014 composite mark. FOF ¶ 216. Again, AFA's trademark counsel filed one specimen of use in which "of America" was removed and that had not actually been used in commerce. Ex. 243. AFA filed an amendment after the Association raised the issue, and the USPTO accepted the amendment. Ex. 93; Tr. (Santana) 816:5-11, 818:10-23. Mark Ingber's associate, Emily Santana, testified that she had created a "mock up" of the AFA website to show AFA an example of the specimen they would need for filing and then mistakenly filed this mock up instead. FOF ¶¶ 218-19. However, none of the additional specimens of use in the amended Section 8 filing show the '014 Mark as distinct from the '067 Mark, which contains the words "of America." See Ex. 93; Tr. (Ingber) 838:16-20; FOF ¶ 227.
According to the Association, AFA had to create these false specimens of use because AFA almost never refers to itself by the two-word "Alzheimer's Foundation"
*282name except in the realm of contested keyword advertising. Yet, although the Section 8 filing contains the electronic signature of acting AFA CEO Charles Fuschillo, Mr. Ingber and Ms. Santana testified that neither Mr. Fuschillo nor Mr. Ingber had ever seen the false specimen of use. Tr. (Santana) 809:16-21; Tr. (Ingber) 828:22-830:10. Mr. Fuschillo had authorized the Ingber Law Firm to apply his signature to documents to be presented to the USPTO. Tr. (Fuschillo) 22:18-21. There was no evidence presented that AFA was aware of either of the false specimens of use.
H. Damages
Since the Court does not find for the Association on liability, there is no need to detail its theories of damages in any great length. Briefly, the Association offers three alternative theories of damages. Ex. 256 ¶¶ 42-48; Ex. 260.
The Association first presented an accounting expert, Marylee Robinson, CPA, who offered two calculations of damages. In Ms. Robinson's first scenario, she attempted to calculate revenues to AFA that were associated with its use of the name "Alzheimer's Foundation" in ads resulting from online searches by measuring all online revenues received within 30 days of the computer user seeing the phrase "Alzheimer's Foundation" in the heading of an online ad on a search engine results page. Ex. 260 ¶ 10; AFA FOF ¶ 249. Scenario 1 calculated damages from January 2010 through December 2016 to be $5,005,509. FOF ¶ 270.
In the second scenario, Robinson attempted to calculate revenues to AFA tied directly to its purchase of Association Marks and related terms as search keywords by measuring all online revenues received within 30 days after the computer user entered one of the Association Marks as a keyword. Ex. 260 ¶ 11; AFA FOF ¶ 250. Ms. Robinson's Scenario 2 calculated damages from January 2010 through December 2016 to be $1,220,264. FOF ¶ 271.
Robinson's damages calculations rest on a number of dubious assumptions. Ms. Robinson's calculations assume that 100% of the donors who entered one of the Association Marks as a keyword or that saw the two-word phrase were confused and would have donated to the Association but for the allegedly infringing behavior.8 Moreover, the donor is counted in both scenarios regardless of what else the donor sees online or offline or learns about AFA or the Association within the 30-day period; donors who may no longer be confused at the point of donation because their confusion was dispelled by the website are still included. Tr. (Robinson) 473:8-12, 475:6-19, 476:18-486:5.
The Association, through Harry Johns, its CEO, also offered a calculation focused on the increase in contributions to AFA from 2011 through 2016 as compared to a baseline amount established in 2006. Ex. 256 ¶¶ 43-47. Mr. Johns calculated damages of slightly over $11 million. Id. ¶ 46. Mr. Johns claims that this measures lost profits, that is, profits that would have otherwise gone to the Association. AFA FOF ¶ 247. His calculation is based on the core assumption that all donations received above the 2006 baseline were caused by the at-issue behavior, an assumption for which he offers no basis. His calculation includes donations from events, talks, print ads, telethon donors in a later year, direct mailings, and repeat, institutional, and corporate donors, and he did not separate online contributions from checks or grants.
*283Tr. (Johns) 226:14-227:14, 237:2-240:16, 248:7-249:15. Johns did not personally perform the calculations. Tr. (Johns) 208:11-13; 223:14-249:15.
AFA charges that none of the calculations establishes a causal link between AFA's allegedly infringing actions and the damages. As discussed briefly below, see infra Part III.C, the Court credits this argument, finding that both Robinson and Johns make a number of unwarranted assumptions.
III. Conclusions of Law
A. Lanham Act
The Association alleges claims against AFA under two similar sections of the Lanham Act, which protects trademark holders from unfair infringement.
First, Counts 1 and 2 of the Association's Counterclaims assert claims for trademark infringement under Section 32(a) and (b), respectively, of the Lanham Act. Amended Counterclaims ("Am. Count."), Dkt. No. 125, ¶¶ 48-58. Section 32 provides that:
(1) 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...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; or
(b) reproduce, counterfeit, copy, or colorably imitate a registered mark and apply such...to labels, signs, prints, packages, wrappers, receptacles or advertisements intended to be used in commerce upon or in connection with the sale...or advertising of goods or services on or in connection with which such use is likely to cause confusion, or to cause mistake, or to deceive, shall be liable in a civil action....
15 U.S.C. §§ 1114(1)(a)-(b).
Second, Counts 3 and 4 allege false designation of origin under sections 43(a)(1)(A) and (a)(1)(B), respectively, of the Lanham Act. Am. Count. ¶¶ 59-68. Section 43 prohibits the use in commerce of any word, term, name, symbol, device, or combination thereof, which:
(A) is likely to cause confusion, or to cause mistake, or to deceive as to the affiliation, connection, or association of such person with another person, or as to the origin, sponsorship, or approval of his or her goods, services, or commercial activities by another person, or
(B) in commercial advertising or promotion, misrepresents the nature, characteristics, qualities, or geographic origin of his or her or another person's goods, services, or commercial activities....
15 U.S.C. §§ 1125(a)(1)(A)-(B).
Despite differences in the statutory language, the same legal test applies to claims for trademark infringement and false designation under the Lanham Act. See Virgin Enters. Ltd. v. Nawab, 335 F.3d 141, 146 (2d Cir. 2003) ; The Sports Auth. Inc. v. Prime Hosp. Corp., 89 F.3d 955, 960 (2d Cir. 1996) ; Opal Fin. Grp., Inc. v. Opalesque Ltd., 634 Fed.Appx. 26, 27 (2d Cir. 2015).
As a threshold matter, in some situations, including that presented here, a plaintiff must allege that the defendant has made "use in commerce" of the plaintiff's trademark as the term is defined in 15 U.S.C. § 1127. In Rescuecom Corp. v. Google Inc., the Second Circuit held that the purchase of trademarks as keywords for internet advertising qualifies as using the *284mark in commerce, see 562 F.3d 123 (2d Cir. 2009), and AFA does not affirmatively dispute that the Association has established use in commerce. See Tr. 1039:11-13 (confirming that AFA does not dispute "use in commerce"). The Court concludes it is established here. See CJ Prods. LLC v. Snuggly Plushez LLC, 809 F.Supp.2d 127, 158 (E.D.N.Y. 2011) (finding a keyword purchase to be a valid "use in commerce"); accord LBF Travel, Inc. v. Fareportal, Inc., No. 13-CV-9143 (LAK)(GWG), 2014 WL 5671853, at *7 (S.D.N.Y. Nov. 5, 2014) ; Allied Interstate LLC v. Kimmel & Silverman P.C., No. 12-CV-4204 (LTS), 2013 WL 4245987, at *3 (S.D.N.Y. Aug. 12, 2013).
Second, in order to vindicate its claims, the Association must show that (1) it has a valid trademark entitled to protection; and (2) a likelihood that AFA's mark or use will cause confusion in the marketplace as to the source or sponsorship of the products at issue. See Playtex Prods., Inc. v. Georgia-Pac. Corp., 390 F.3d 158, 161 (2d Cir. 2004) ; Gruner + Jahr USA Publ'g v. Meredith Corp., 991 F.2d 1072, 1074 (2d Cir. 1993).
In the instant action, the Court is specifically presented with two allegedly infringing behaviors, which should be analyzed together, as they appear in the real world. As the Association describes it, the key question before the Court is "whether there is a likelihood of confusion arising from (1) AFA's use of the two-word name 'Alzheimer's Foundation' and (2) AFA's use of Association Marks and derivations of Association Marks in search keyword advertising and metatag keywords." Post-Trial Br. at 19 (emphasis added). The Association does not claim that AFA's use of "Alzheimer's Foundation of America," in any context, is infringing. Alzheimer's Found., 2015 WL 4033019, at *14.
It merits a brief pause to emphasize that a proper analysis under the Lanham Act takes these two actions into account simultaneously.9 Virtually no court has held that, on its own, a defendant's purchase of a plaintiff's mark as a keyword term is sufficient for liability. See, e.g., 1-800 Contacts, Inc. v. Lens.com, Inc., 755 F.Supp.2d 1151, 1174 (D. Utah 2010), aff'd in part, rev'd in part on other grounds, 722 F.3d 1229 (10th Cir. 2013), Gen. Steel Domestic Sales, LLC v. Chumley, No. 10-CV-1398, 2013 WL 1900562, at *10 (D. Colo. May 7, 2013), aff'd , 627 Fed.Appx. 682 (10th Cir. 2015) ; Hearts on Fire Co., LLC v. Blue Nile, Inc., 603 F.Supp.2d 274, 285 (D. Mass. 2009). However, as noted, the Court is tasked not with evaluating AFA's bidding on Association Marks in a vacuum, but rather the effect of the keyword purchases in conjunction with AFA's resulting advertisements. See Edible Arrangements, LLC v. Provide Commerce, Inc., No. 14-CV-250 (VLB), 2016 WL 4074121, at *11 (D. Conn. July 29, 2016). Thus, the crux of the likelihood of confusion issue is: Are AFA's actions in purchasing Association Marks as keywords and advertising itself as "Alzheimer's Foundation" likely to cause consumer confusion?
1) Validity of the Mark
In analyzing the Association's claims, the Court must consider whether the plaintiff has a valid trademark to begin with. Courts "look to see whether plaintiff's mark merits protection." Louis Vuitton Malletier v. Dooney & Bourke, Inc., 454 F.3d 108, 115 (2d Cir. 2006). To be protectable, the mark must be "distinctive" and not "generic."
*285Genesee Brewing Co. v. Stroh Brewing Co., 124 F.3d 137, 143 (2d Cir. 1997). A mark is distinctive if it "serves to identify a particular source." Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768, 112 S.Ct. 2753, 120 L.Ed.2d 615 (1992).
A registered mark enjoys a presumption of distinctiveness. See Alzheimer's Found., 2015 WL 4033019, at *6 (citing Dress for Success Worldwide v. Dress 4 Success, 589 F.Supp.2d 351, 358 (S.D.N.Y. 2008) ). The Association Marks are registered on the USPTO Principal Register and the registrations are owned by the Association. Exs. 1, 259 ¶¶ 3-5. For reasons outlined below in Part III.A.2.a, the Court concludes that the Association's Marks are not generic and do merit protection. The first prong is thus satisfied.
2) Likelihood of Confusion
Having established that the Association has valid and protectable marks, the Court then turns to "the crucial issue in an action for trademark infringement," that is, "whether there is any likelihood that an appreciable number of ordinarily prudent purchasers are likely to be misled, or indeed simply confused, as to the source of the goods in question." Mushroom Makers, Inc. v. R.G. Barry Corp., 580 F.2d 44, 47 (2d Cir. 1978). A plaintiff must prove "a probability of confusion, not a mere possibility." Gruner + Jahr, 991 F.2d at 1077.
"The Second Circuit has yet to adopt a particular approach for determining the issue of consumer confusion when a competitor uses a trademark to purchase keywords to advertise its products for sale on the Internet...." CJ Prods., 809 F.Supp.2d at 158. The Court will follow what is quickly becoming the majority approach in applying the eight-factor balancing test introduced by Judge Friendly in Polaroid Corp. v. Polarad Electronics Corp., 287 F.2d 492 (2d Cir. 1961), to this consideration. See, e.g., CJ Prods., 809 F.Supp.2d at 159 ; TheUniversal Church, Inc. v. Universal Life Church/ULC Monastery, No. 14-CV-5213(NRB), 2017 WL 3669625, at *11 (S.D.N.Y. Aug. 8, 2017) ; S & L Vitamins, Inc. v. Australian Gold, Inc., 521 F.Supp.2d 188, 205-06 (E.D.N.Y. 2007) ; Hearts on Fire, 603 F.Supp.2d at 289 (applying the First Circuit's equivalent). The Court's resolution of each separate factor is a finding of fact, while the balancing of the factors is a conclusion of law. See Star Indus. v. Bacardi & Co., Ltd., 412 F.3d 373, 384 (2d Cir. 2005).
The Court's analysis is not mechanical, but remains focused on the ultimate question of whether consumers are likely to be confused as a result of the allegedly infringing conduct. Id. The eight Polaroid factors are: (1) strength of the plaintiff's mark; (2) similarity of the marks; (3) proximity of the products and their competitiveness with one another; (4) evidence that the senior user may 'bridge the gap' by developing a product for sale in the market of the alleged infringer's product; (5) evidence of actual consumer confusion; (6) existence of bad faith on the part of the defendant; (7) the respective quality of the products; and (8) sophistication of consumers in the relevant market. Id.
Because the Polaroid factors are merely "a useful guide through a difficult quagmire," and each factor "must be evaluated in the context of how it bears on the ultimate question of likelihood of confusion," it is important to clarify upfront the specific type of consumer confusion alleged by the Association. Lois Sportswear, U.S.A., Inc. v. Levi Strauss & Co., 799 F.2d 867, 872 (2d Cir. 1986) ; accord Louis Vuitton Malletier v. Burlington Coat Factory Warehouse Corp., 426 F.3d 532, 532, n.2 (2d Cir. 2005) (stating that "the Polaroid factors must be applied with an eye toward each" type of alleged confusion).
*286At base, the Association alleges that AFA uses a name and practices likely to cause confusion "as to the affiliation, connection or association" of AFA with the Association, commonly referred to as "source confusion." See Am. Count. ¶ 60. However, source confusion may arise at various points in consumers' purchasing process and it is important, in analyzing the Polaroid factors, to remember when the consumers must be likely to be confused and about what they must be confused in order for the Association to prevail. See Lois Sportswear, 799 F.2d at 872.
Most commonly, a plaintiff will bring a claim based on point-of-sale confusion. With point-of-sale confusion, at the time consumers purchase a good or service or make a donation, they are confused as to the source of the product or service they are purchasing. "The classic case of direct confusion occurs when '[c]ustomers want to buy the [plaintiff's] product and because of the similarity of the marks, mistakenly buy the [defendant's] product instead." 1-800 Contacts, Inc., 722 F.3d at 1238-39 (quoting McCarthy on Trademarks § 23:10 at 23-70 (alterations in original) ).
Crucial here, courts also recognize a type of pre-sale confusion called "initial interest confusion," which occurs when "potential consumers initially are attracted to the junior user's mark by virtue of its similarity to the senior user's mark, even though these consumers are not actually confused at the time of purchase." Jordache Enters., Inc. v. Levi Strauss & Co., 841 F.Supp. 506, 514-15 (S.D.N.Y. 1993) ; see also Mobil Oil Corp. v. Pegasus Petroleum Corp., 818 F.2d 254, 260 (2d Cir. 1987). In the context of the internet, "the concern is that potential customers of one website will be diverted and distracted to a competing website." Bihari v. Gross, 119 F.Supp.2d 309, 319 (S.D.N.Y. 2000). However, "[e]ven if the customer quickly becomes aware of the competing source's actual identity and can rectify the mistake, the damage to the first user" may already have been done. BigStar Entmn't, Inc. v. Next Big Star, Inc., 105 F.Supp.2d 185, 207 (S.D.N.Y. 2000). The damage can occur in three ways:
[ (1) in] the original diversion of the prospective customers' interest; [ (2) in] the potential consequent effect of that diversion on the customer's ultimate decision whether or not to purchase caused by an erroneous impression that two sources of the produce maybe associated; and [ (3) in] the initial credibility which may be accorded by the interested buyer to the junior user's products-customer consideration that may otherwise be unwarranted and that may be built on the strength of the senior user's mark, reputation and goodwill.
Id. (citing Brookfield Commc'ns, Inc. v. W. Coast Entm't Corp., 174 F.3d 1036, 1063-64 (9th Cir. 1999) ). Nonetheless, because a user diverted to the wrong website can easily get "back on track," the Second Circuit has also expected a plaintiff prove "intentional deception" to prevail in an initial interest confusion case in the internet context. See Savin Corp. v. Savin Grp., 391 F.3d 439, 462 n.13 (2d Cir. 2004).
The task of distinguishing between situations that are and are not actionable as initial interest confusion may be best illustrated by two scenarios. The first is a pure bait and switch scenario, which is clearly actionable. For example, if AFA had purchased Association Marks as keywords and then advertised itself as "Alzheimer's Association," with nothing in its ad to distinguish itself from the Association aside from its URL, this would be a clear case of infringement. By contrast, comparison ads are not violations of the Lanham Act because it is clear to consumers that the plaintiff's mark is only being *287used as a point of contrast. For example, if Pepsi were to purchase the keyword "Coca-Cola" and then publish an ad with the headline "Try Pepsi-It Is Better than Coca-Cola," it would be the rare consumer confused by the source of the advertisement. The consumer searching for "Coca-Cola" may have been diverted, but she has not been confused or misled.10 The sine qua non of trademark infringement is consumer confusion caused by the infringing behavior, and behavior meant to harm a competitor does not necessarily entail infringement if a consumer is unlikely to be confused. For this reason, courts have repeatedly found that the purchase of a competitor's marks as keywords alone, without additional behavior that confuses consumers, is not actionable. See, e.g., 1-800 Contacts, Inc., 755 F.Supp.2d at 1174 ; Gen. Steel, 2013 WL 1900562, at *10 ; Hearts on Fire, 603 F.Supp.2d at 285. As one district court in Massachusetts framed the issue: "Mere diversion, without any hint of confusion, is not enough." Hearts on Fire, 603 F.Supp.2d at 286. This Court subscribes to that approach.
Thus, for the purpose of the Court's analysis-which principally focuses on initial interest confusion-the crucial question is whether AFA's actions in purchasing Association Marks as keywords and advertising itself under its two-word name are more akin to a bait-and-switch likely to confuse consumers, or to the simple act of offering consumers a choice.
Having described the specific confusion alleged here and having laid out how the Court views the ultimate question of likelihood of confusion in this context, the Court now turns to the task of evaluating each Polaroid factor in service of its final assessment.
a. Strength of Mark
The "strength of the mark" analysis focuses on the distinctiveness of the mark, and while registered marks enjoy a presumption of distinctiveness as discussed above, a registered mark may still be deemed weak. A mark's strength is measured by two factors: "(1) the degree to which it is inherently distinctive; and (2) the degree to which it is distinctive in the marketplace." W.W.W. Pharm. Co., Inc. v. Gillette Co., 984 F.2d 567, 572 (2d Cir. 1993) (internal quotation marks and citation omitted). These factors have also been labeled "conceptual strength" and "commercial strength." 2 McCarthy on Trademarks § 11:83 (5th ed.).
The Second Circuit has developed four categories to aid in determining the inherent distinctiveness or conceptual strength of a mark: (1) generic, (2) descriptive, (3) suggestive, and (4) arbitrary or fanciful. See Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F.2d 4, 9 (1976). These categories are defined as follows:
A generic mark is generally a common description of goods and is ineligible for trademark protection.... A descriptive mark describes a product's features, qualities or ingredients in ordinary language, and may be protected only if secondary meaning is established.... A suggestive mark employs terms which do not describe but merely suggest the features of the product, requiring the purchaser to use 'imagination, thought, and perception to reach a conclusion as to the nature of goods....' Fanciful or arbitrary marks are eligible for protection without proof of secondary meaning and 'with ease of establishing infringement.'
W.W.W. Pharm. Co., 984 F.2d at 572 (citations omitted).
*288The Association argues that the fact that its trademark is registered means the Court need not further analyze its inherent strength. Post-Trial Br. at 20 (citing Coty Inc. v. Excell Brands, LLC, 277 F. Supp. 3d 425, 441 (S.D.N.Y. 2017) ). While it is true that "an inconstestable registration-defined as a mark that 'has been in continuous use for five consecutive years' after the date of registration and is still being 'use[d] in commerce'-is conclusive evidence of a plaintiff's exclusive right to use the specific mark," Coty Inc., 277 F.Supp.3d at 441 (quoting 15 U.S.C. §§ 1065, 1115(b) ), a mark's inherent strength is still properly analyzed within the aegis of the first Polaroid factor. "The fact that a trademark is the subject of a federal registration that has ripened into incontestable status does not support the conclusion that the mark is 'strong.' " 2 McCarthy on Trademarks § 11:82 (5th ed.). In fact, in Coty, upon which the Association relies, the Court continued on to consider anew the strength of the mark at issue. See 277 F.Supp.3d at 445-46. "[W]hile the registration adds something on the scales, we must come to grips with an assessment of the mark itself." Zobmondo Entm't, LLC v. Falls Media, LLC, 602 F.3d 1108, 1115 (9th Cir. 2010).
The relative strength or weakness of the marks at issue is crucial to the Court's determination. In the present context of online advertising:
This factor is probative of confusion here because a consumer searching for a generic term is more likely to be searching for a product category. That consumer is more likely to expect to encounter links and advertisements from a variety of sources. By contrast, a user searching for a distinctive term is more likely to be looking for a particular product, and therefore could be more susceptible to confusion when sponsored links appear that advertise a similar product from a different source.
Network Automation, 638 F.3d at 1149 (citing Brookfield Commc'ns, 174 F.3d at 1058 n.19 ). As a result, evaluation of this first Polaroid factor is particularly crucial to the Court's overall determination and merits extended analysis. The Court considers the conceptual and commercial strength of the principal mark at issue-"Alzheimer's Association"-in turn.
Conceptually, the Court finds that the mark "Alzheimer's Association" is descriptive; it describes a charitable organization that does work related to the disease called Alzheimer's. A shorthand for the organization's formal name of Alzheimer's Disease and Related Disorders Association, the mark is comprised of two fairly common words. "Alzheimer's" is a descriptor for the disease the charitable organization focuses on, and is utilized by AFA in its name, and, unsurprisingly, by countless other charities engaged in the fight against Alzheimer's disease. "Association" is also a common descriptor applied to many charitable organizations, and not particularly distinct from other common charity names such as "foundation," "charity," "federation," "fund," or "society." Even taken together, the two words suggest nothing beyond forming a bare descriptor for an association related to Alzheimer's. Cf. Streetwise Maps, Inc. v. VanDam, Inc., 159 F.3d 739, 744 (2d Cir. 1998) (noting that several map producers use "street" in product names; thus plaintiff's mark using "street" was not particularly distinctive). The mark is conceptually weak.
While "the policies of trademark law disfavor according exclusive rights to marks that are descriptive," the compromise struck by the Lanham Act is to "allow[ ] descriptive marks registration and a minimal scope of protection, upon establishing that the mark had acquired a secondary meaning as a designator of source."
*289TCPIP Holding Co., Inc. v. Haar Commc'ns, Inc., 244 F.3d 88, 100 (2d Cir. 2001) (citations omitted). It is because "Alzheimer's Association" can establish secondary meaning that it still merits protection under trademark law as "distinctive" and not "generic." See supra Part III.A.1. Secondary meaning is established when "the mark comes to identify not only the goods, but the source of those goods." See 20th Century Wear, Inc. v. Sanmark-Stardust Inc., 815 F.2d 8, 10 (2d Cir. 1987) (internal quotation marks and citation omitted). This is a factual determination which "entails vigorous evidentiary requirements." Thompson Med. Co. v. Pfizer Inc., 753 F.2d 208, 217 (2d Cir. 1985) (internal quotation marks and citation omitted). The indicia commonly recognized as showing secondary meaning are (1) advertising expenditures; (2) consumers studies linking the name to a source; (3) sales success; (4) unsolicited media coverage; (5) attempts to plagiarize the mark; and (6) the length and exclusivity of the mark's use. Id. (citations omitted).
The Court's analysis of the mark's "secondary meaning" overlaps with the second overall factor of the mark's strength, its distinctiveness in the marketplace. See Streetwise Maps, 159 F.3d at 743. However, "secondary meaning" and "commercial strength" are distinct legal concepts in that the former is a threshold issue of validity, while the latter focuses on an infringement determination. 2 McCarthy on Trademarks §§ 11:82, 15:1, 15:25 (5th ed.). It is for this reason that a mark can simultaneously exceed the threshold for establishing "secondary meaning" and, thus, gain the protection of trademark law, while still being commercially weak. "Alzheimer's Association" presents this situation.
Although "secondary meaning" and "commercial strength" are separate concepts, "[w]hen determining the commercial or marketplace strength of a mark, the courts look to the same kind of evidence of real world recognition of the mark as is used to decide the presence or absence of secondary meaning." Id. § 11:82. Commercial strength of a mark is important because:
Widespread consumer recognition of a mark previously used in commerce increases the likelihood that consumers will assume it identifies the previously familiar user, and therefore increases the likelihood of consumer confusion if the new user is in fact not related to the first. A mark's fame also gives unscrupulous traders an incentive to seek to create consumer confusion by associating themselves in consumers' minds with a famous mark.
Virgin Enters., 335 F.3d at 148 (citation omitted).
The Association argues that its leading position in the Alzheimer's charity space, which is not in dispute, establishes the strength of its mark. The Association is the world's largest Alzheimer's-related non-profit and the world's largest non-governmental funder of Alzheimer's research, raising more than $160 million in contributions and spending more than $44 million on advertising or public awareness in 2016 alone. The Association had nearly 9 billion media impressions and more than 41 million website visits in 2016 and "Alzheimer's Association" has been used by the Association to identify itself since at least 1988. It would strain credulity to deny that AFA has purchased Association Marks as keywords in part due to the Association's size and presence and not only because of the descriptive nature of the words. See Tr. (Leone) 797:2-23 (opining that the competitor campaign was successful in part because of the size and recognition of the Alzheimer's Association).
However, consumer studies in evidence demonstrate that "Alzheimer's Association"
*290has not gained strong secondary meaning. "Consumer surveys can be the most persuasive evidence of secondary meaning because they are direct evidence of the relevant consumer groups' association of a service with a particular source." Rockland Exposition, Inc. v. All. of Auto. Serv. Providers of N.J., 894 F.Supp.2d 288, 320 (S.D.N.Y. 2012). The Copernicus Surveys show that when respondents are prompted for the first two health charity organizations that come to mind, only 3% mention the Association. See Ex. AF at ALZ0058612. This low number is understandable given that the question prompts only for the top two charities that come to mind, however, the response to the more probative question-asking respondents which organizations "involved in the fight against Alzheimer's disease" they have heard of-still yields only around 11% offering the Association. Id. at ALZ0058614. While undoubtedly higher than the 3% who mention AFA without prompting, the Court cannot conclude on the basis of these survey results that the Association has developed a strong secondary meaning or commercial strength associated with its standard character mark. Moreover, even amongst "Champions," a subgroup meant to designate the demographic groups that the Association targets with its advertising messages, unaided awareness of the Association among organizations "involved in the fight against Alzheimer's disease" has hovered between 10% and 20% for the past eight years. See id. at ALZ0058615; Tr. (Carson) 305:8-22 (explaining the Champions group).
In sum, the PTO's prior registration of the trademark constitutes prima facie evidence that the Association acquired the secondary meaning necessary for trademark protection of its '223 Mark, see 15 U.S.C. §§ 1057(b), 1115(a), and the Court finds no basis to challenge this presumption. However, "[d]etermination of whether there is infringement requires an evaluation of the strength of that trademark. Finding that a non-inherently distinctive designation has a secondary meaning is not the end of the full analysis of infringement, only the beginning." 2 McCarthy on Trademarks § 11:82 (5th ed.). The mark may have passed the threshold of recognition required to establish secondary meaning and warrant protection. But in light of the mark's descriptive nature and relatively low consumer awareness of the Association, the Court finds that the mark is weak both conceptually and commercially.
The Court's finding that "Alzheimer's Association" is a weak, descriptive mark helps drive the overall disposition of this case. As the Second Circuit stated, "[t]he narrower scope of protection commanded by weak, descriptive marks has two explanations[:]"
The first is that granting one merchant a monopoly over the use of descriptive marks tends unfairly to prevent competing merchants from employing appropriate descriptive terms in their marketing. The second has to do with the character of descriptive marks in relation to the likelihood of consumer confusion.
TCPIP, 244 F.3d at 100. As alluded to above, the second explanation is crucial here because "[l]ikelihood of confusion depends on consumer expectations." Id. at 101. If consumers expect similar marks to identify the same source, that similarity is likely to cause confusion; when consumers expect that similar marks were chosen because they describe similar products or services, the similarity is less likely to invite confusion that the products or services must come from the same source. Id.
Here, because of the weakness of the mark, it is not easy to disaggregate the consumers searching for the Association as a specific organization from those searching genetically for an Alzheimer's charity when they type "Alzheimer's Association"
*291or some similar derivation into a web browser. Relatedly, and as more fully explained below, the descriptive nature of both parties' marks also creates the challenge of disaggregating those consumers confused by AFA's actions from those confused simply by the similarity of the descriptive marks. The first factor thus weighs substantially in favor of AFA.
b. Similarity of the Marks
When evaluating the similarity of marks, "courts look to the overall impression created by the [marks] and the context in which they are found and consider the totality of factors that could cause confusion among prospective purchasers." Gruner + Jahr, 991 F.2d at 1078 ; accord Louis Vuitton Malletier, 454 F.3d at 117 ; Spring Mills v. Ultracashmere House, Ltd., 689 F.2d 1127, 1130 (2d Cir. 1982) ("[A]n inquiry into the degree of similarity between two marks does not end with a comparison of the marks themselves.").
With respect to the Association's claim regarding AFA's purchase of Association Marks as keywords and metatags, the proper comparison is between the resulting ads for AFA and Association advertisements or other search results. This is how consumer confusion would manifest itself. The mere fact of AFA's purchase of the Association Marks either verbatim, as with the keyword "alzheimer's association," or nearly verbatim, as with the keyword "alzheimer's association memory walk," see Ex. 97, cannot in and of itself cause confusion. See, e. g., 1-800 Contacts, Inc., 755 F.Supp.2d at 1174 ; Gen. Steel, 2013 WL 1900562, at *10 ; Hearts on Fire, 603 F.Supp.2d at 285 ; J.G. Wentworth, S.S.C. Ltd. P'ship v. Settlement Funding LLC, No. 06-CV-0597, 2007 WL 30115, at *6 (E.D. Pa. Jan. 4, 2007). In this sense, the analysis bleeds together into the Court's analysis of AFA's use of the two-word mark "Alzheimer's Foundation," the Association's second claim. "When examining the similarity of the marks in the AdWords context, the court must consider the degree of similarity between plaintiff's service mark and the [defendant's] advertisements appearing on the search-results page." CJ Prods., 809 F.Supp.2d at 159 (internal quotation marks, citation, and original alterations omitted).
Considering the offending ads in context then, the scenario is as follows: a consumer, probably searching for the Alzheimer's Association as an entity, but possibly searching for a generic Alzheimer's charity, types "alzheimer's association" into the Google or Bing search bar. Results show one or more sponsored ads at the top of the search results, followed by other content, such as the organization's websites, social media pages, Wikipedia pages, or other websites related to Alzheimer's.
As shown and described above, AFA's ad might show "Alzheimer's Foundation" in standard typeface, without any graphic, followed by the associated website URL that clicking on the ad would take you to, and a brief description or tagline. See, e.g., Ex. 71. Different trial exhibits showing various Google search results for "alzheimer's association" have the ads in a different order, or show different text used by each organization, but consumers are presented with the same basic context as shown in Exhibit 71 above. See, e.g., Ex. 90, 161, 212; see also Ex. 97 (listing the various headers and textual elements of AFA's advertisements).
The Association argues that AFA, by purchasing Association Marks as keywords, using the two-word name "Alzheimer's Foundation" in the header of its ads, and by sometimes using the word "association" in the text of the ads, has generated results that are strikingly similar for a consumer. See Post-Trial Br. at 23-24. The Association notes that in organic search results, AFA's website-appears as *292"Alzheimer's Foundation of America." Id. The Association suggests that AFA's choice to use its two-word name instead in the ads is purposeful and designed to increase the similarity with the Association's own ads. Id. at 24.
AFA responds that when viewing the ads in context, any similarity that exists between "Alzheimer's Foundation" and "Alzheimer's Association" is mitigated by the unique information and URL of its ads. AFA Post-Trial Br. at 20. AFA also uses no Association Mark in its ad. Id. at 20-21. AFA further relies on the unique qualities of its website, both visually and with respect to the website content, to argue against similarity, id. at 21, but this is misplaced, as the proper comparison is between the ads themselves and not their destinations. See Edible Arrangements, 2016 WL 4074121, at *13 (distinguishing confusion at this stage from "downstream" confusion (citing Hearts on Fire, 603 F.Supp.2d at 289 ) ). Moreover the URLs-www.alz.org and www.alzfdn.org-and ad text resulting from Google searches of purchased keywords in November 2017-"Honor a Loved One with a Tribute Donation. Support Research and Care." and "An Organization Providing Support for Your Loved One with Alzheimer's"-are not particularly dissimilar.
Viewed in somewhat of a vacuum, with consideration of the visual context of the ads but without consideration of the consumer's sophistication or purpose in entering the keyword and also without consideration of the differences in websites that the ads link to, the Court concludes that the ads at issue are similar in appearance and meaning. This factor weighs in favor of the Association.
c. Competitive Proximity
The third factor considers the extent to which the parties' products compete with each other for customers. Brennan's, Inc. v. Brennan's Rest., LLC, 360 F.3d 125, 134 (2d Cir. 2004). Courts examine "the nature of the products themselves and the structure of the relevant market," including "the manner in which the products are advertised, and the channels through which the goods are sold." Cadbury Beverages v. Cott Corp., 73 F.3d 474, 480 (2d Cir. 1996) (citation omitted). "When two users of a mark are operating in completely different areas of commerce, consumers are less likely to assume that their similarly branded products come from the same source." Virgin Enters., 335 F.3d at 150.
AFA attempts to downplay the competitive nature of the two organization's services and missions, noting that AFA is primarily focused on providing care services while the Association focuses on funding research, and attesting that the only competition is against the disease. AFA Post-Trial Br. at 21-22; Ex. 36 ¶¶ 15-16; Ex. DS ¶ 109. However, any fair understanding of the relevant consumer population and the genesis of the long history of litigation between these parties suggests that they are two charities performing similar work, occupying a similar space as Alzheimer's charities, and competing, to a certain extent, over the same donor pool. This factor weighs in favor of the Association.
d. Bridging the Gap
"Bridging the gap" concerns the likelihood that a plaintiff "not in direct competition with the defendant at the time a suit is brought will later expand the scope of its business so as to enter the defendant's market." Pretty Girl, Inc. v. Pretty Girl Fashions, Inc., 778 F.Supp.2d 261, 268 (E.D.N.Y. 2011). Because the parties already operate in the same market, see supra Part III.A.2.C, this factor is irrelevant. See Patsy's Brand, Inc. v. I.O.B. Realty, Inc., 317 F.3d 209, 218 (2d Cir. 2003) ;
*293U.S. Polo Ass'n v. PRL USA Holdings, Inc., 800 F.Supp.2d 515, 531 (S.D.N.Y. 2011).
e. Actual Confusion
The Second Circuit has made clear the prime importance of actual confusion as a factor, even if no one factor is dispositive. "It is self-evident that the existence of actual consumer confusion indicates a likelihood of consumer confusion. We have therefore deemed evidence of actual confusion 'particularly relevant' to the inquiry." Virgin Enters., 335 F.3d at 151 (citing Nabisco, Inc. v. PF Brands, Inc., 191 F.3d 208, 228 (2d Cir. 1999), and Streetwise Maps, 159 F.3d at 745 ). To be sure, actual confusion "is very difficult to prove and the [Lanham] Act requires only a likelihood of confusion as to source." Savin Corp., 391 F.3d at 459 (internal quotation marks and citation omitted). Proof of actual confusion is not necessary. See Centaur Commc'ns, Ltd. v. A/S/M Commc'ns, Inc., 830 F.2d 1217, 1227 (2d Cir. 1987). However, evidence of actual confusion is strongly probative of a likelihood of confusion. See Morningside Grp. Ltd. v. Morningside Capital Grp., LLC, 182 F.3d 133, 141 (2d Cir. 1999) ; Savin Corp., 391 F.3d at 459 (quoting World Carpets, Inc. v. Dick Littrell's New World Carpets, 438 F.2d 482, 489 (5th Cir. 1971) ("There can be no more positive or substantial proof of the likelihood of confusion than proof of actual confusion.") ).
However, to be useful to the Polaroid analysis, this evidence of confusion must be related to the actions or behaviors at issue. See, e.g., Rehrig Pac. Co. v. Norseman Plastics Ltd., Inc., No. 03-CV-470, 2003 WL 25667625, at *31 (C.D. Cal. Sept. 30, 2003) (noting that plaintiff "has not produced evidence sufficient to establish a causal link between [the consumer's] confusion and defendant's allegedly infringing behavior" on evidence of actual confusion prong). "Evidence of actual confusion must be viewed in its evidentiary context. Confusion may not be causally related to the use of similar marks at all." 2 McCarthy on Trademarks § 23:13 (5th ed.). In Judge Sweet's decision at the preliminary injunction stage, he concluded that "although some amount of actual confusion from the two-word designation has been conceded by the AFA, the Association has not established at this point that the scale of that confusion is significant. It has also failed to establish any actual confusion stemming from the AFA's internet search engine advertising practices...." 2015 WL 4033019, at *10 (emphasis added). Despite the Association's additional evidence proffered at trial, the Court finds the same result here. The question here is the extent to which any actual confusion can be attributed to AFA's internet advertising practices. See, e.g., The Sports Auth., 89 F.3d at 963 ("To show actual confusion, [Plaintiff] must demonstrate that [Defendant's] use 'could inflict commercial injury in the form of either a diversion of sales, damage to goodwill, or loss of control over reputation.' " (quoting Lang, 949 F.2d at 583 ) ); Champagne v. DiBlasi, 36 Fed.Appx. 15, 18 (2d Cir. 2002) (recounting the proffered evidence of actual confusion and noting that "[t]he defendants were not show to have done anything to cause confusion resulting in ticket purchases by these customers"); accord Streamline Prod. Sys., Inc. v. Streamline Mfg., Inc., 851 F.3d 440, 457 (5th Cir. 2017) ("However, not all confusion counts: evidence of actual confusion must show more than a fleeting mix-up of names; rather it must show that the confusion was caused by the trademarks and it swayed consumer purchase." (internal quotation marks and citation omitted) ).
As described above, the Association points to three basic types of evidence of actual confusion. First, it notes that over the years the organizations have received *294in excess of 11,000 checks made payable to the other. The Association notes that it receives more checks made out to "Alzheimer's Foundation," than to "AFA" or "Alzheimer's Foundation of America." FOF ¶¶ 111-112.2.
While probative of actual confusion generally, this evidence does not prove that this confusion is generated by any of AFA's actions. The Association offers one example of a donation check it received in 2014 accompanied by a printed copy of AFA's donation form from its website, but beyond this one example there is nothing to suggest that most, or even many, of these 11,000 checks reflect confusion born of AFA's online advertising practices. Rather, it could reflect, per the discussion above about the strength of the Association's mark, that "association" and "foundation" are synonyms and easily confused in the mind of a donor. The parties disagreed about the magnitude of the mistakenly addressed checks, but the Court will not wade into a numerical game of defining de minimis, especially when no evidence was presented of what percentage improperly addressed checks constituted of AFA's total receipts. See Tr. (Fuschillo) 63:9-81:6. Suffice to say that given the lack of proof connecting the mistakenly addressed checks to the advertising practices at issue, the first type of evidence of confusion is both limited and largely unpersuasive.
Second, the Association points to very limited instances of confusion in the media and to anecdotal reports of consumer confusion. Ex. 259 ¶ 12-14; Tr. (Leesment) 537:19-538:14. For example, Christine Foh, General Counsel for the Association, testified about a dispute concerning the estate of Sandra Lynch, in which a bequest to "Alzheimer's Foundation, Pennsylvania Chapter," was interpreted by a court as intended to benefit the Association's Delaware Valley Chapter. Ex. 259 ¶ 12. But there is no reason to believe this confusion is related to AFA's challenged actions. The Association presented evidence of NBC's Today show running a quote from the Association superimposed over AFA's logo. Post-Trial Br. at 27; Ex. 13.1; 14.1. However, the logo used by NBC-which shows "AFA" and "Alzheimer's Foundation of America"-was not one at issue in this litigation. No weight can be given to this evidence. Counterclaim Plaintiff also submitted a video clip from Celebrity Family Feud in which the host referred to one team as raising money for "Alzheimer's Foundation," when they were actually raising money for the Association. Ex. 12. As with the check evidence, the Association cannot credibly connect this single instance of confusion to AFA's practice of purchasing Association Marks as keywords or of advertising itself with the two-word name Alzheimer's Foundation.
In fact, these instances of confusion might be seen to undermine the Association's claims. In light of the descriptive nature of the marks, confusion between the unchallenged "Alzheimer's Foundation of America" mark and "Alzheimer's Association" supports the notion that it is the weakness of the marks and consumers' inattention, not AFA's specific disputed practices, that yields confusion.
Third, and most importantly, the Association presented the survey evidence and opinion testimony of Dr. Kent Van Liere, Ph.D., Vice President at NERA Economic Consulting. To be clear, survey evidence is not evidence of actual confusion, because surveys "do not measure the degree of actual confusion by real consumers;" instead, surveys provide circumstantial evidence of the likelihood of confusion. 2 McCarthy on Trademarks § 32:184 (5th ed.). Nonetheless, "[i]n borderline cases where evidence of confusion is not available or is not persuasive, the gap can *295sometimes be filled by a properly conducted survey of the relevant class of prospective customers of the goods or services at issue." Id. § 23.17. In order to be probative of the issue of actual confusion, however, the consumer surveys must "have been fairly prepared and [their] results directed to the relevant issues." See Sterling Drug, Inc. v. Bayer AG, 14 F.3d 733, 741 (2d Cir. 1994) (internal quotation marks and citation omitted). A properly designed survey can offer information about an artificial world from which the Court may draw inferences about the real world; accordingly, even if survey evidence is technically not evidence of actual confusion, the Court will analyze Dr. Van Liere's surveys within this factor of the Polaroid test. Accord Mobil Oil Corp., 818 F.2d at 259 (analyzing survey evidence under "actual confusion").
As discussed at length above, Van Liere and his firm conducted two studies, both of which were the subject of a Daubert motion brought by AFA to exclude Van Liere's testimony. The Court denied AFA's Daubert motion, deciding instead to "[take] into account all of the record evidence." See Tr. 996:24-997:2; see Conopco, Inc. v. Cosmair, Inc., 49 F.Supp.2d 242, 253 (S.D.N.Y. 1999) ("A survey is only inadmissible if its flaws destroy all of its relevance."). Nonetheless, valid criticisms of errors in methodology properly go to the weight of the expert's testimony. See Schering Corp. v. Pfizer Inc., 189 F.3d 218, 228 (2d Cir. 1999).
Dr. Van Liere's first study-a lineup or "Squirt" style study was meant to gauge the likelihood of confusion between the standard character marks "Alzheimer's Association" and "Alzheimer's Foundation." A lineup test is a relatively common test of confusion between marks. See 2 McCarthy on Trademarks § 32:173.50 (5th ed.) (discussing different courts' approaches to "Squirt" format evidence). Dr. Van Liere found a net confusion rate of 34% and consequently opined that "AFA's use of the name 'Alzheimer's Foundation' is likely to confuse a substantial proportion of consumers in the relevant population into believing that 'Alzheimer's Foundation' is the same as, or is affiliated with Alzheimer's Association." Ex. 262 ¶ 33.
While none of Mr. Poret's criticisms of Study 1 was sufficient to render it inadmissible, considered together, the flaws of Study 1 do suggest that it merits little weight. Most critically, Study 1 presents a comparison between the bare words "Alzheimer's Association" and "Alzheimer's Foundation," the simplicity of which only enhances the importance of choosing a proper control stimulus so as to control for the influence of factors not being tested-in this case, factors beyond the impact of AFA's use of the two-word mark "Alzheimer's Foundation." Dr. Van Liere's choice of a weak control in "Alzheimer's Trust" likely artificially inflated the net confusion levels. NERA staff's pre-tests showed a higher level of confusion generated by "Alzheimer's Charity" and "National Foundation for Alzheimer's Research." And while having a two-word control term may have been important, the choice of "Trust," a word more unique than "Foundation," more easily distinguished from "Association," and that denotes other meanings, is flawed. While the perfect control may not exist, the flaws in the chosen control undermine the weight the Court affords the study's results.
The second study Dr. Van Liere conducted was designed to replicate the process consumers go through in searching "Alzheimer's Association" and viewing the results on the search engine page. Based on its results, Dr. Van Liere concluded that AFA's keyword advertising campaign using Association Marks led to a likelihood of confusion at both the initial interest page and at the landing page.
*296As with Study 1, none of the flaws of the Internet Search Survey was sufficient to render it inadmissible, but considered together, the well-founded criticisms lodged by Mr. Poret's credible and persuasive testimony give the Court serious reservations in lending the results substantial weight. First, Study 2 presented a test condition with AFA's allegedly infringing ad as the first result. The order bias created by Dr. Van Liere's failure to rotate the condition may have inflated the net confusion rate. Second, Study 2 also used a control stimulus without any sponsored ads, thus depressing confusion in the control scenario (and inflating net confusion) because the first search result is the Association's website. By failing to control for the presence of sponsored ads altogether, Study 2 inflated net confusion. A better control would have only removed the allegedly infringing AFA ad, not all sponsored ads.
Further, unlike in both studies, when consumers enter "alzheimers association" as a search term in the real world, they may be using the term generically and not have a particular organization in mind when searching. In these situations, the consumers cannot be confused by AFA's ad or webpage, and this dynamic lessens the applicability of either study's results.
In sum, the Court's task in evaluating the three types of evidence of confusion in this case is complicated by the need to disaggregate confusion between the two charities' names in the colloquial sense from confusion born of the allegedly infringing behavior. Undoubtedly, the evidence of the misnamed checks and other proffered instances of mistaken labeling reveal the existence of some limited confusion, however the record does not support linking those isolated instances of confusion to the at-issue behaviors. See Streamline, 851 F.3d at 457 ("We have rejected anecdotal evidence of actual confusion when the proponent did not show that a misleading representation by the defendant, as opposed to some other source, caused a likelihood of confusion." (internal quotation, citation, and alteration omitted) ). The Association seems to ask the Court to infer that because there were mislabeled checks and AFA also advertised itself by purchasing Association Marks as keywords and holding itself out as the "Alzheimer's Foundation," that these two facts must be connected. See Tr. 946:17-947:25. There is scant evidence in the record supporting that inference. The Association has failed to connect the check evidence or other anecdotal and media instances of confusion to AFA's allegedly infringing actions.
Dr. Van Liere's surveys might have filled in that gap in the evidence, allowing the Court to make informed inferences about the existence of actual confusion, but the problems with the studies' controls, in particular, were too severe for the Court to place great weight in the results. In fact, the very necessity of a well-designed survey control is "to pin down causation." 2 McCarthy on Trademarks § 32:187 (5th ed.) (citation omitted). "The control question or questions should use a mark similar enough to the actual mark that it provides an accurate measure of the confusion created by the accused mark, not by some other similarity." Id. (citation omitted). Although not meriting the exclusion of Dr. Van Liere's report altogether, the flaws in Dr. Van Liere's two surveys outlined above indicate that the opinions and conclusions formed based on the studies should be accorded minimal weight. Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 596, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993) ("Vigorous cross-examination [and] presentation of contrary evidence...are the traditional and appropriate means of attacking shaky but admissible evidence").
*297In total, due to the limited quantum of confusion, none of which can be credibly tied to AFA's Internet advertising practices, this factor weighs in favor of AFA.
f. Evidence of Bad Faith / Defendant's Intent
Bad faith generally refers to an attempt by a defendant to adopt a mark "with the intention of capitalizing on plaintiff's reputation and goodwill and on any confusion between his and the senior user's product." Savin Corp., 391 F.3d at 460 (internal quotation marks, citation, and alterations omitted); Lang v. Ret. Living Publ'g Co. , 949 F.2d 576, 583 (2d Cir. 1991). In some cases, however, even where the defendant knew of the earlier mark, the Second Circuit has found the defendant to be acting in good faith "in the absence of additional evidence indicating an intent to promote confusion or exploit goodwill or reputation." Star Indus., 412 F.3d at 388. A finding of bad faith "does not bear directly on whether consumers are likely to be confused." Virgin Enters., 335 F.3d at 151.
As noted above, the Second Circuit, in Savin Corp., suggested that in the context of internet initial interest cases, "[b]ecause consumers diverted" can more readily "get back on track," "intentional deception" is a requirement. 391 F.3d at 462 n.13. AFA argued at trial that the absence of good faith and the presence of intentional deception should be separate analyses. Tr. 1040:8-15. In this case, the Court conducts the analysis together for the sake of simplicity and because, in this case, a finding in favor of AFA on the bad faith factor automatically yields a finding of no intentional deception. In other words, a finding for Plaintiff on this particular Polaroid factor is a necessary, but not sufficient, condition for finding for Plaintiff overall.
The Association tells a compelling story about AFA's motivations, based largely on a timeline of events presented above. See supra Part II.G. According to the Association, AFA complained of the confusion created by the similarity in the organization's names back in 2004, but, despite its own complaints and lawsuits, then proceeded to employ Association Marks in its metatags and as keywords. The Association claims that AFA "undertook activities specifically to profit from the Association's goodwill and to generate, exacerbate and profit from confusion between the two entities." Post-Trial Br. at 34; see also Tr. 926:24-927:6. According to the Association, these activities included first filing for registration of the two-word name "Alzheimer's Foundation" in late 2009 and 2010 after litigation between the parties commenced, using the two-word name in online advertising, alongside, at times, the word "association," and twice submitting false specimens of use to the USPTO cover up for the fact that AFA generally did not use the two-word name outside of the context of contested keyword advertising. Post-Trial Br. at 33-34.
Unsurprisingly, AFA presents a different story and timeline of its behavior over the years. AFA insists that it was both purchasing Association Marks as keywords and using the two-word name Alzheimer's Foundation back in 2004 in its internet advertising. See Exs. 97, BR-1. AFA was unaware of the false specimens of use, and argues that bad faith cannot be imputed to AFA from mistakes made by two different junior associates for its outside counsel. The Court did not find Mr. Ingber, AFA's trademark counsel, particularly credible in his explanation of the false filings, but circumstantial evidence about AFA's trademark applications does not itself suggest an intention to capitalize on the Association's goodwill or to exploit confusion.
Moreover, AFA cries, the Association was doing the same, purchasing "Alzheimer's Foundation" as a keyword prior to *2982010, see Ex. CB (showing Association's decision to remove all "foundation"-related keywords from its keyword campaigns), and co-existing with AFA for six years prior to this lawsuit while both organizations were engaging in these practices. When the Association complained about AFA's use of the word "association" in its ad text in 2014, AFA removed the offending ads, while claiming it did not use the word out of an intent to deceive. Exs. DS ¶ 61, BX, I. Consequently, AFA argues that it had a good faith belief that its actions were not infringing because it had used the two-word mark for six years without incident, the Association had no rights to the mark, none of the ads were false nor did they contain any Association Marks; Google and Bing allowed their ad practices, and the Association was engaged in similar behavior.
The Court finds this factor favors AFA. The Association makes numerous assumptions in ascribing intent to AFA's actions that are unsupported by the evidence it presents. In many respects, the lack of direct evidence of actual confusion undermines the Association's attempts to argue that AFA acted with the intention of exploiting consumer confusion. It is the rare piece of evidence that suggests AFA believed the practices at issue here-its purchase of Association Marks as keywords and its use of the two-word name-were causing consumer confusion. The Court's comparison of the circumstantial evidence the Association offers here with the evidence of bad faith in cases like CJ Products and Edible Arrangements , see infra Part III.A.2.i, places its limitations in stark relief.
Additionally, because of the Court's finding on this factor, Plaintiff also fails to establish the "intentional deception" the Second Circuit suggests is needed for proving internet-related initial interest confusion. Savin Corp., 391 F.3d at 462 n.13.
g. Quality of the Product
There is nothing in the record to show that AFA's product is inferior to the Association's. The Association attempts to re-frame this factor as focusing on the harm to their brand, asserting that "[a] senior user is not required to put its reputation in a junior user's hands, no matter how capable those hands may be." Post-Trial Br. at 34 (quoting U.S. Polo Ass'n, Inc., 800 F.Supp.2d at 537 ). But even if the Court were to accept the Association's framing, "[t]he issue of the quality of the secondary user's product goes more to the harm that confusion can cause the plaintiff's mark and reputation than to the likelihood of confusion," Virgin Enters., 335 F.3d at 152 (citation omitted), making it less important here.
In fact, for the purposes of likelihood of confusion, the factor has the opposite effect. "[A] very marked difference in quality between the two products would militate against finding a likelihood of confusion as consumers 'will be less likely to assume that the senior user whose product is high-quality will have produced the lesser-quality products of the junior user.' " Star Indus., 412 F.3d at 389 (quoting Savin Corp., 391 F.3d at 461 ). While receiving little weight in the overall calculus, because there is little evidence suggesting such a difference, this factor weighs in favor of AFA.
h. Sophistication of the Consumer
This factor examines a buyer's attentiveness in evaluating a product or service. See CJ Prods., 809 F.Supp.2d at 155-56. "The more sophisticated the consumers, the less likely they are to be misled by similarity in marks." TCPIP, 244 F.3d at 102 (citing Cadbury Beverages, Inc., 73 F.3d at 480 ). The Association *299points to the fact that a high percentage of online donors are first-time donors in arguing that the relevant population is not particularly sophisticated. Post-Trial Br. at 35 (citing Exs. 60, 79). AFA counters that consumers on the internet are more sophisticated than your average consumer. AFA Post-Trial Br. at 38 (citing Network Automation, 638 F.3d at 1152 ). AFA claims that internet users who access, read and understand internet searches and follow multi-step processes to donate to charities online are even more sophisticated. Id. Additionally, many of the donors to the organizations are foundations, corporations, repeat donors, and other sophisticated persons or entities. Ex. DS ¶¶ 67-68.
AFA's argument largely misses the point. The Association does not claim that foundations or corporations are likely to be confused by AFA's Google or Bing ads, nor that repeat donors have repeatedly been confused about which organization they are donating to.11 Rather, the relevant population is the average consumer searching for the Association online and seeing AFA's ad, who are less likely to be institutional donors and many of whom are first-time donors. See Ex. 60 at AFA0021415; FOF ¶¶ 262-63. With respect to initial interest confusion, these consumers are not particularly sophisticated. Moreover, while it takes more effort to navigate the website to reach the donation page-a consideration taken into account with respect to the overall likelihood of AFA's actions to confuse consumers at the point of sale-in this day and age, the ability to complete a form on a website does not itself make consumers particularly sophisticated. This factor weighs in favor of the Association.
i. Weighing the Factors
To sum up, four factors favored AFA (strength of the mark, actual confusion, bad faith, and quality of the product), three factors favored the Association (similarity of marks, competitive proximity, and sophistication of the consumers), and one factor was neutral (bridging the gap). However, the Court "must focus on the ultimate issue of whether consumers are likely to be confused, rather than on a numerical break-down of the factors." WE Media, Inc. v. Cablevision Sys. Corp., 94 Fed.Appx. 29, 33 (2d Cir. 2004). For the question presented here-whether consumers typing in an Association Mark and seeing an AFA ad using the two-word name "Alzheimer's Foundation" are likely to be confused-the Court finds that the most important factors are strength of the mark, similarity of the marks, and evidence of actual confusion. Two of those three factors strongly favor AFA.
Finally, there is one additional consideration with internet advertising that does not neatly fit within one of the eight factors analyzed above: the labeling and segregation of online advertising. Google and Bing both label sponsored advertising results with the word "ad" next to the listing, and Bing previously displayed the ads on a sidebar, separated from the organic results. The presence of this "ad" signifier heightens consumers care and attention in clicking on the links, and further diminishes the likelihood of initial interest confusion. See Network Automation, 638 F.3d at 1153-54 ; 1-800 Contacts, Inc., 722 F.3d at 1245.
In determining whether these factors end up adding up to a likelihood of confusion, one of the persistent difficulties for the Court is its attempt to determine how many consumers fall into each of the following relevant subpopulations. First, there are those consumers not confused by AFA's practices. These consumers may not *300be confused either because in searching for "Alzheimer's Association," they use the phrase in a generic sense and so are not looking for the Association in particular, or because although they are searching for the Association in particular, upon viewing the content of the search results, they understand the difference between the two organizations.
Second, there are those consumers who click on AFA's ad by mistake, and are diverted because of AFA's purchase of Association Marks as keywords and its use of the two-word name "Alzheimer's Foundation." These consumers would not have been diverted if AFA had used its full name or had not purchased Association Marks as keywords. The consumers in this subgroup form examples of "initial interest confusion." A subset of these consumers remain confused even after viewing AFA's website, with its distinctive color, logo, name, URL, and content, and proceed to make a donation to AFA believing they were donating to the Association. This subset constitutes traditional "point-of-sale" consumer confusion.
Third and finally is the group who is mistakenly diverted for reasons unrelated to AFA's at-issue actions, such as because of the inherently weak and descriptive nature of the parties' marks. These consumers are "confused" in the colloquial sense, but would have been confused even if they searched the word "Alzheimer's" alone or even if AFA solely utilized its full name.
Based on the evidence presented at trial, especially in light of the flaws of the surveys noted above, it is difficult to conclude with any reasonable certainty the proportion of consumers that fall into each group. The surveys could have assisted the Court's task in disaggregating the confusion caused by the descriptive nature of the marks from confusion caused by AFA's actions, but the controls Dr. Van Liere used did not help control for that confounding variable. Is it possible that some consumers searching specifically for the Alzheimer's Association are diverted to AFA's website because it drops "of America" from its ad text? Yes. However, as with any alleged trademark violation, plaintiffs must prove "a probability of confusion, not a mere possibility." Gruner + Jahr, 991 F.2d at 1077 (emphasis added). The extensive trial record does not support a finding of such a probability. That some confusion may result from the similarity between "Alzheimer's Association" and "Alzheimer's Foundation" is not sufficient to conclude that AFA's actions are likely to mislead "an appreciable number of ordinarily prudent purchasers." Mushroom Makers, 580 F.2d at 47.
Comparison of the facts presented here to cases in which other courts have found a likelihood of consumer confusion-an exercise that the Association invites in its post-trial reply brief, see Post-Trial Reply at 8-9-is instructive. Two such cases that serve as helpful comparators are CJ Products and Edible Arrangements. The difference between the present context and record evidence and the facts before the courts in those cases only helps to confirm the Court's evaluation.
In CJ Products, which is cited three times by the Association, the plaintiffs developed, produced, sold, and distributed a plush toy in the shape of an animal that unfolds into a flat pillow called a "Pillow Pet." 809 F.Supp.2d at 139-40. Plaintiffs had also extensively advertised on television, and their product was a top selling toy during the recent holiday season. Id. at 147, 153. Among plaintiffs' registered trademarks was the standard character mark "My Pillow Pets," they were in the process of registering "Pillow Pets," and their website was www.mypillowpets.co. Id. at 139-40. Defendants purchased the phrases "Pillow Pets" and "My Pillow *301Pets" through Google Adwords, and then advertised through sponsored posts with phrases like "Official PillowPets.Co," which linked directly to defendants' website, pillowpets.co. Id. at 157-59. The defendants' website used the same color, format, fonts, and phrases- including the mark "Pillow Pets"-as plaintiffs'. Id. at 159. Additionally, there was clear evidence of actual confusion caused by the allegedly infringing behavior in the form of consumer complaints and website traffic data, which showed the traffic to defendants' pillowpets.co precipitously rising at the same time they began to use plaintiffs' mark. Id. at 155, 160. Finally, there was clear evidence of defendants' bad faith in their decision to abandon their previously-used URL "plushez.com," and to falsely write "As Seen on TV" on their website and on the product tags. Id. at 147-48, 160.
Accordingly, CJ Products is factually inapposite. The evidence in that case showed stronger marks at issue, clear evidence of actual confusion caused by the defendants' conduct, and bad faith. Defendants' conduct was so egregious that the court granted plaintiffs a preliminary injunction with respect to their copyright infringement claims, false advertising claims, and false designation of origin claim, and enjoined defendants' use of the plaintiffs' marks in the Google AdWords program: By contrast, here the Court encounters relatively weak marks that do not appear in Defendant's sponsored ads, limited evidence of actual confusion caused by Defendant's actions, and limited, circumstantial evidence of bad faith.
In Edible Arrangements, plaintiff Edible Arrangements, a leading seller of "artfully designed fresh fruits...arranged to resemble floral arrangements" sued a competitor who mostly sold "coated fruit products" and who did not sell fruit packaged to resemble floral arrangements. 2016 WL 4074121, at *1-3. Nevertheless, the defendant not only purchased "edible arrangements" as a keyword, but also used the phrase "edible fruit arrangements" in the header of its sponsored ads. Id. at *2-3. Despite limited evidence of actual confusion, the court focused on the unsuitability of the phrase "edible fruit arrangements" as a descriptor for hand-dipped fruit and held that a "reasonable trier of fact could conclude that a consumer searching for 'edible arrangements' is looking for a distinct product line of aesthetically shaped fruit and not merely for any and all gifts containing boxes of edible fruits and berries." Id. at *13. Defendant's ad, in which the seller is only identified in the small print of the URL, could result in a likelihood of confusion within the broader context. Id. While the district court was only denying defendant's motion for summary judgment, the court analyzed all of the Polaroid factors and found that the strength of plaintiff's mark, the similarity of the marks and products, and defendant's bad faith "strongly suggest a likelihood of confusion from [defendant's] use of [plaintiff's] mark." Id. at *11.
By contrast, the comparative weakness of the keyword and mark "Alzheimer's Association," Coupled with AFA's use of "Alzheimer's Foundation" in its ads, does not comprise the same clear exploitation of the plaintiff's brand or of consumers' clear intentions in searching that keyword as seen in Edible Arrangements. A descriptor such as "chocolate-dipped berries" was clearly more accurate than "edible fruit arrangements" for defendant's products, and while AFA could have referred to itself as "Alzheimer's Foundation of America" in its ads, its use of the two-word phrase "Alzheimer's Foundation," does not clearly traffic off the reputation of the Association in the same way.
"[T]he crucial issue in an action for trademark infringement... is whether *302there is any likelihood that an appreciable number of ordinarily prudent purchasers are likely to be misled, or indeed simply confused, as to the source of the goods in question." Mushroom Makers, Inc., 580 F.2d at 47. In the above cases, the answer was yes. Here, the Court does not conclude based on the extensive trial record that an appreciable number of prudent consumers would be confused based on the allegedly infringing actions of the Counterclaim Defendant.12
B. Other Causes of Action
Given the Court's findings on the Lanham Act claims, the Association's remaining causes of action all fail as well.
First, the Association pleaded unfair competition under New York common law. Am. Count. ¶¶ 96-99. To prevail, a plaintiff must establish the same elements as under the Lanham Act, and demonstrate the defendant's bad faith. See Luv n' Care, Ltd. v. Mayborn USA, Inc., 898 F.Supp.2d 634, 643 (S.D.N.Y. 2012) ; Jeffrey Milstein, Inc. v. Greger, Lawlor, Roth, Inc., 58 F.3d 27, 34-35 (2d Cir. 1995) ; Tiffany (NJ) Inc. v. eBay, Inc., 576 F.Supp.2d 463, 519 (S.D.N.Y. 2008), aff'd in part, rev'd in part on other grounds, 600 F.3d 93 (2d Cir. 2010) ; U.S. Polo Ass'n, 800 F.Supp.2d at 538. Per the Court's analysis above, the Association fails to establish unfair competition.
Second, the Association claims AFA committed deceptive acts and practices under New York General Business Law § 349. Am. Count. ¶¶ 101-03. To prevail under G.B.L. § 349, the standards are substantially the same as those applied to false designation claims under the Lanham Act. See Naked Cowboy v. CBS, 844 F.Supp.2d 510, 518 (S.D.N.Y. 2012) ; Van Praagh v. Gratton, 993 F.Supp.2d 293, 301 (E.D.N.Y. 2014). In fact, the standard for a § 349 claim is actually higher, as there must be "specific and substantial injury to the public interest over and above the ordinary trademark infringement...." Nomination Di Antonio E Paolo Gensini S.N.C. v. H.E.R. Accessories Ltd., No. 07-CV-6959 (DAB), 2009 WL 4857605 at *8 (S.D.N.Y. Dec. 14, 2009) ; see also TheUniversal Church, Inc., 2017 WL 3669625, at *17 ; LBF Travel, Inc., 2014 WL 5671853, at *11. Regardless, the Association fails to establish this claim.
Third, and finally, the Association seeks the cancellation of AFA's '014 Mark on grounds that it creates a likelihood of confusion with the Association's '223 Mark, Am. Count. ¶¶ 104-05, and also on the basis of abandonment. Post-Trial Br. at 37. For many of the same reasons stated above, the Court rejects the first argument for cancellation. Regarding abandonment, the Association claims that AFA has not used the '014 Mark in commerce since 2011. Failure to use a mark for three years creates a presumption of abandonment. 15 U.S.C. § 1127.
"The party asserting abandonment bears the burden of persuasion with respect to two facts: (1) non-use of the mark by the legal owner, and (2) lack of intent by that owner to resume use of the mark in the reasonably foreseeable future." ITC Ltd. v. Punchgini, Inc., 482 F.3d 135, 147 (2d Cir. 2007), "Because it constitutes a forfeiture of a property right, abandonment of a mark must be proven by clear and convincing evidence." Empresa Cubana Del Tabaco v. Culbro Corp., 213 F.R.D. 151, 156 (S.D.N.Y. 2003). Here, AFA claims that its Section 8 maintenance application filed and accepted by the USPTO
*303in 2017 shows that it did not abandon use of the '014 Mark. AFA Post-Trial Br. at 47 (citing Ex. 93). As discussed above, the Court does not see where the '014 Mark, as opposed to the '067 Mark, is included as a specimen of use in this filing. Nonetheless, putting the specimens of use aside, the filing of the application alone demonstrates AFA's intent to use the '014 Mark. Given the high burden placed on the Association to establish abandonment, and the limited evidence adduced, the Court cannot find abandonment.
C. Damages
In light of the Court's findings on liability, the Court need not reach the merits of the parties' disputes over damages, but briefly discusses them to further illustrate the weak causal link between AFA's allegedly infringing actions and any consumer confusion. "When seeking money damages, a plaintiff 'must prove both actual damages and a causal link between defendant's violation and those damages.' " Aviva Sports, Inc. v. Fingerhut Direct Marketing, Inc., 829 F.Supp.2d 802, 815 (D. Minn. 2011) (quoting Rhone-Poulenc Rorer Pharm., Inc. v. Marion Merrell Dow, Inc., 93 F.3d 511, 515 (8th Cir. 1996) ).
The Association's damages expert, Marylee Robinson, made her calculations based on the assumption that the Court would find that every instance of AFA using "Alzheimer's Foundation" in an ad would be infringing. Nonetheless, even operating under that assumption, it does not necessarily follow that 100% of computer users who donate online to AFA within thirty days of seeing "Alzheimer's Foundation" on a search engine results page were confused. While any damages calculation based on incomplete information is inherently an estimate, Ms. Robinson's calculation is clearly over-inclusive. She does not analyze or control for the many factors that would influence someone's decision to donate outside of the infringement, including that they were searching for AFA in particular from the start or that they read the content of AFA's website or otherwise investigated or compared Alzheimer's charities before donating. See supra Part II.H. To assume that any donation within the 30-day period is caused by the allegedly infringing behaviors is unwarranted and unsupported by the evidence.
The "baseline theory" of damages proffered by Mr. Johns is even less credible. His opinion was offered despite not having personally performed the calculations. Tr. (Johns) 208:11-13; 223:14-249:15. His opinion was based on the assumption that all of AFA's revenues from 2011 onwards, above the baseline of its 2006 revenues, would have gone to the Association but for AFA's allegedly infringing behavior, even revenue accrued through grants, events, talks, mailings, corporate donations, or other donations not made online. Tr. (Johns) 229:10-14; 233:23-240:16. The calculation does not even attempt to link the "lost profits" to AFA's actions.
Thus, even if liability had been established, the Association would not be entitled to actual damages. See Gucci Am., Inc. v. Guess?, Inc., 868 F.Supp.2d 207, 243 (S.D.N.Y. 2012) ("In order to obtain monetary relief for damages stemming from trademark infringement, the owner of a trademark normally must prove actual consumer confusion or deception resulting from the violation of the Lanham Act." (internal quotation marks, citation, and alterations omitted) ).
D. Attorney's Fees
The Lanham Act permits the Court to award reasonable attorney's fees "in exceptional cases," 15 U.S.C. § 1117, which includes instances "of fraud or bad faith."
*304Twin Peaks Prods., Inc. v. Publ'ns Int'l, Ltd., 996 F.2d 1366, 1383 (2d Cir. 1993) (internal quotation marks and citation omitted). "A finding of bad faith is warranted when a plaintiff's trademark infringement claim is entirely without merit and was asserted for improper purposes such as harassment or delay." Contractual Obligation Prods., LLC v. AMC Networks, Inc., 546 F.Supp.2d 120, 129 (S.D.N.Y. 2008). The Court does not find that the Association was motivated by an "improper purpose" in filing its counterclaims, and does not award attorney's fees.
IV. Conclusion
For the foregoing reasons, the Court finds in favor of Counterclaim Defendant Alzheimer's Foundation of America and dismisses Counterclaim Plaintiff's claims in the entirety. The Court does not award attorney's fees.
The Clerk of Court is respectfully directed to enter judgment and to close the case.
SO ORDERED.
Unless otherwise specified, all docket numbers refer to the docket in 10-CV-3314, and not to the related action.
Judge Sweet dismissed AFA's Fourth Amended Complaint on the grounds that AFA lacked standing as the named payee on checks received and deposited by the Association, and that AFA failed to show that the Association had used AFA trademarks in commerce. See Dkt. No. 170. AFA then moved for and was granted reargument, but the Court upheld its prior decision. Alzheimer's Found., 2014 WL 4290455.
Because only the trademark counterclaims remain, all following references to the parties will be as Counterclaim Plaintiff and Counterclaim Defendant.
See Dkt. Nos. 344-50. Both parties submitted post-trial proposed findings of fact and were provided an opportunity to respond to each other's proposed findings of facts with opposing record citations. See Dkt. Nos. 344, 346, 347 & 349. The abbreviation "FOF" represents Dkt. No. 347 which incorporates the Association's proposed findings of facts and AFA's responses. The abbreviation "AFA FOF" represents Dkt. No. 349, which incorporates AFA's additional proposed findings of fact and the Association's responses. If a party did not oppose a proposed finding of fact with an appropriate record citation, and if that unopposed proposed finding of fact was supported by an appropriate record citation, the Court deemed that fact admitted and incorporates such unopposed facts as factual findings of the Court.
To the extent that any finding of fact reflects a legal conclusion, it shall to that extent be deemed a conclusion of law, and vice versa.
Exhibits offered by Counterclaim Plaintiff Alzheimer's Association are denoted numerically. Exhibits offered by Counterclaim Defendant Alzheimer's Foundation of America are denoted alphabetically.
See Winnie Hung, Limiting Initial Interest Confusion Claims in Keyword Advertising, 27 Berkeley Tech. L.J. 647, 666 (2012) (stating that statistics show ad rank has a significant effect on consumers' confusion).
Accordingly, Robinson conceded that if the Court were to determine that AFA's use of the term "Alzheimer's Foundation" on its own was not likely to cause confusion, there would be no damages in scenario one. Tr. (Robinson) 459:13-16.
The Association has argued that AFA's use of the two-word name "Alzheimer's Foundation" can be analyzed alone. See Tr. 1008:8-18. Regardless, because the Court does not find the compound effect of the keywords and two-word name likely to confuse consumers, there is no need to consider the two-word name separately.
The Association recognized this distinction at trial. See Tr. 945:10-25.
Although, as recognized below, the failure to account for repeat donors does speak to the flaws in the Association's damages calculations. See infra Part III.C.
Because the Court finds in favor of AFA, it declines to reach AFA's affirmative defenses of laches and of unclean hands.