BanxCorp v. Costco Wholesale Corp. (978 F. Supp. 2d 280, 2013)

OPINION AND ORDER [REDACTED]

KENNETH M. KARAS, District Judge.

Consider the percentage “3.95%.” It seems to be a totally ordinary percentage. It is the amount by which Eastern Michigan University increased its tuition and fees for the 2012-13 school year relative to the previous one. It is how much the Mayor of Poughkeepsie proposes to increase the city tax levy for 2014. It is the amount by which sugar prices rose in India one day in November 2012. And, according to Plaintiff Banxcorp, it was the United States national average interest rate for five-year certificates of deposit as of December 21, 2005.

For Plaintiff, then, 3.95% is not such an ordinary percentage. Rather, Plaintiff initiated this lawsuit in part because it claims it has a valid federal copyright in that particular percentage — or, at least, that it has a copyright in its series of percentages of national average interest rates, of which 3.95% on December 21, 2005 is one part. And it claims that it is entitled to substantial money damages because Defendants Costco and Capital One — a large retailer and a large bank, respectively — unlawfully copied those percentages in a series of individual advertisements touting how much higher their particular deposit rates were than the national average, as reported by Plaintiff. Defendants’ copying of individual averages is conceded; at issue for the copyright claim in this case is whether the percentages themselves are entitled to federal protection under the Copyright Act.

The Court previously determined, on Defendants’ motion to dismiss, that Plaintiff plausibly had alleged that its works of authorship had certain features that could, drawing all inferences in Plaintiffs favor, lead to the conclusion that its works of authorship were entitled to copyright protection. See BanxCorp v. Costco Wholesale Corp., 723 F.Supp.2d 596, 601-09 (S.D.N.Y.2010). But now the evidence is in, and, on cross-motions for summary judgment, the Court determines that, even drawing all reasonable inferences from the evidénce in Plaintiffs favor, the averages are unprotectable because they are uncopyrightable facts, because they are too short to be copyrighted, and because the so-called merger doctrine — which applies where there is “only one ... or so few ways of expressing an idea, that protection of the expression would effectively accord protection to the idea itself,” id. at 608 (internal quotation marks and alterations omitted) — bars copyright protection.

But that is not the only claim in this case. Plaintiff also contends that Defendant Capital One exceeded the scope of a License Agreement it signed that allowed it to use Plaintiffs data for certain marketing purposes. The Court finds that the contract is ambiguous in relevant part and that a reasonable jury could decide in favor of either Party on this claim. Accordingly, summary judgment is not appropriate for either party on the contract claim.

I. Background

A. Factual Background

1. The Parties

Plaintiff Banxcorp is a Delaware corporation that does business under the name “Banxquote.” (Pi’s. Resp. to DSUF ¶ 131 (citing Lipkis Decl. Ex. 76).) Plaintiff touts online that it “provides a family of widely followed indices and benchmarks that measure the rates and performance of banking, depository, mortgage, home equity and consumer loan markets.” (Mertzel Decl. Ex. 0, at BX 0048.) In other words, Plaintiff regularly surveys the interest rates or other prices offered by particular financial institutions across the country, and then compiles this data into various indices that represent “national averages” of the rates.

Defendant Capital One Financial Corporation is a Delaware corporation that is the parent company of co-Defendants Capital One Bank (USA), N.A., and Capital One, N.A., which are nationally chartered banks with principal places of business in Virginia. (DSUF ¶¶ 1-4.) The Court refers to these entities collectively as “Capital One” except where expressly noted. Capital One, a well-known national bank, provides so-called national direct banking products and services directly to consumers from its national headquarters. (DSUF ¶¶ 5, 8.)

Defendant Costco Wholesale Corporation, a Washington corporation, is the second largest retailer in the United States. (DSUF ¶ 11.) Costco operates over 600 warehouse-style retail stores worldwide and has approximately 66.5 million cardholders. (DSUF ¶ 12.) In addition to the products sold at its warehouses, Costco markets a variety of services to its members. (DSUF ¶ 16.) Nearly all of these services are provided by third parties that have marketing agreements with Costco. (DSUF ¶ 17.)

2. The Use of Plaintiffs Data in Capital One and Co-Branded Advertisements and Marketing Materials

Capital One markets its banking products nationally. During the time period relevant to this case, its marketing materials frequently provided the Capital One rate being offered for a particular financial product alongside one or more comparison rates, such as a competing bank’s rates or a national average rate. (DSUF ¶¶ 65, 67.) Capital One used comparison rates in many ads because it found that consumers often responded favorably to advertisements that provided a point of reference. (DSUF ¶ 71.)

Beginning in May 2003, Costco and Capital One entered into a series of marketing agreements. (DSUF ¶¶22, 24.) Costco and Capital One referred to this relationship as a “partnership,” whereby Costco would facilitate the marketing of Capital One products and services to Costco members, and Capital One would provide Costco members with certain financial products and services at a “premium” rate. (DSUF ¶ 36.)

Prior to January 2004, Capital One had been using national averages provided by a company called Bankrate in many of its advertisements. (DSUF ¶ 76.) But, for a variety of reasons — including the fact that Plaintiff published its rates for free online, which allowed potential consumers to verify the accuracy of the national averages, (DSUF ¶¶ 80, 81) — Capital One decided to switch to Plaintiffs averages. (DSUF ¶ 84.) On January 28, 2004, Capital One entered into a license agreement with Plaintiff to use Plaintiffs savings and jumbo CD averages, as well as its savings and jumbo money market averages, in many of its marketing materials, both online and in print. (DSUF ¶¶94, 115.) Capital One agreed to pay $6,000 per year for this privilege. (DSUF ¶ 115.) During the course of the agreement, Capital One obtained the national averages by copying the relevant data directly from Plaintiffs website. (DSUF ¶ 99.)

Soon after the license agreement became effective, Capital One began using Plaintiffs data in its standard national marketing materials. (DSUF ¶ 97.) Later, Capital One began using Plaintiffs averages in marketing materials, both online and in print, that were created and distributed as part of the partnership agreement with Costco. (DSUF ¶ 98.)

The record contains many examples of these partnership advertisements. An entirely typical one from 2006 states at the top: “Earn more with exclusive rates for Costco members!” (Decl. of Michael Kiernan, Ex. C, at COB0000194.) On the left side of the ad, there are several bullet points touting features of the account, and an offer stating that “Costco Executive Members receive $25 credited to their first new account opened.” (Id.) On the right side are two bar graphs. The first says “Money Market Account ($5,000 account balance),” and below that are two bars of different heights. (Id.) The left bar, in large numbering, states that Capital One’s rate is 4.26%, and, in smaller print to the right of this, the ad notifies the reader that 4.26% is the “Annual Percentage Yield,” or “APY,” and there is a single asterisk next to that definition. (Id.) The right bar is much lower, and, above it in slightly smaller lettering and numbering, the ad states that the “National Average” is 1.20% APY, and there are two asterisks next to “APY.” (Id.) The second bar graph, which is reproduced just below, is similar to the first, except the second graph gives the Capital One and national average rate for a “Certificate of Deposit ($5,000 deposit, 5-year term).” (Id.) In this graph, the Capital One rate is 5.16% APY, and the National Average is 3.95% APY. (Id.) The comparative height of the bars is adjusted accordingly.

The single asterisk and the double asterisk are defined in small print on the left side of the page. (Id.) The text following the single asterisk gives further details of the offer. (Id.) It is typical of the fine print that many people have encountered in the industry: the minimum daily balance requirement, minimal initial deposits, and the obligatory disclosures that the “terms and conditions of this offer” and the “rates” advertised are “subject to change without notice.” (Id.) Meanwhile, more relevant for purposes of this case, the text following the double asterisk contains the source of the national average representation. It reads, in full: “National average of APYs for CDs and money market accounts as published by Banxquote.com as of 12/21 /05.” (Id.)

Below the two graphs on the right side of the ad is marketing copy. “I love the exclusive perks Capital One offers Costco Executive Members, like the $25 I received when I opened my account,” says “Jeffrey S.” who is, presumably, a satisfied customer. (Id.) On the left side of the page, the ad implores the reader that he or she should “Open an account today!”, and it instructs the reader either to visit costco.com or call a toll-free number to do so. (Id.) The logos of both Capital One and Costco are featured, and there are additional disclosures, fine print — i.e., “Member FDIC” — -and even a copyright invocation by “Capital One Services, Inc.” (Id.)

Defendants used the national average data reported by Plaintiff frequently and essentially continuously during the 2004-OS period that is at issue in this suit. (PSUF ¶¶ 96, 100.) In particular, the then-current Capital One interest rate was continually displayed next to a relevant national average rate from Plaintiff on a co-branded website, and Defendants regularly distributed brochures and marketing campaigns similar to the advertisement described above during the relevant time period. This co-branded website, which Defendants acknowledge was initially subject to Costco’s approval, was advertised to the public as being accessible solely by visiting Costco’s website at costco.com and clicking on “Services.” (Defs.’ Resp. to PSUF, ¶¶ 89, 97.)

B. Procedural History

1. Prior Determinations

Originally, Plaintiffs CEO Norbert Mehl was also a Plaintiff in this case, and, proceeding pro se, Plaintiffs filed their Complaint on February 25, 2009. BanxCorp, 723 F.Supp.2d at 600. After retaining counsel, Plaintiffs filed the SAC on September 2, 2009. Id.

The SAC alleges seven causes of action. Id. There are two federal causes of action: Count One, which alleges copyright infringement based upon Defendants’ improper use of the BanxQuote Indices, (id. ¶¶ 106-16); and Count Three which alleges violation of the Digital Millennium Copyright Act (“DMCA”), based on allegations that when Defendants copied the BanxQuote Indices they altered or removed the copyright management information BanxCorp. had associated with the data, (id. ¶¶ 126-33). The remaining five causes of action arise under New York law: Count Two alleges hot news misappropriation of the time-sensitive data contained in the BanxQuote Indices, (id. ¶¶ 117-25); Count Four alleges fraud based on allegations that Defendants materially misrepresented their intentions with respect to their use of the BanxQuote Indices pursuant to the license agreement, (id. ¶¶ 134^13); Count Five alleges breach of contract against Capital One only, based on the alleged distribution to, and use of the BanxQuote Indices by, Costco in violation of the License Agreement, (id. ¶¶ 144-51); Count Six alleges unfair competition based on allegations that Defendants’ use of the BanxQuote Indices gave Defendants an unfair competitive advantage both in terms of decreased web traffic at Plaintiffs’ websites and in terms of direct competition in providing savings accounts and CDs, (id. ¶¶ 121, 152-57); and Count Seven alleges unjust enrichment based on allegations that Defendants received value due to their wrongful use of the BanxQuote Indices, (id. ¶¶ 158-61).

Defendants moved to dismiss each claim for failure to state a claim, and the Court granted the motion in part and denied the motion in part. In particular, the Court dismissed as preempted by the Copyright Act Count Four, alleging fraud; Count Six, alleging unfair competition; and Count Seven, alleging unjust enrichment. BanxCorp, 723 F.Supp.2d at 617-20. The Court also dismissed Mehl personally as a Plaintiff, because Mehl conceded he lacked standing. Id. at 621.

On July 8, 2011, the Parties stipulated that Count Two, alleging hot news misappropriation, and Count Three, alleging the DMCA violation, would be dismissed with prejudice. (Dkt. No. 68.) Thus, two claims now remain in the case: Count One, the federal claim for copyright infringement; and Count Five, the state claim for breach of contract against Capital One only.

2. Copyright Registrations

Plaintiffs copyrights were unregistered during the time of Defendants’ allegedly infringing activity. On March 5, 2009, after Plaintiff filed this lawsuit, Mehl submitted to the Register of Copyrights twenty applications for a federal copyright in the averages. (Mertzel Decl. Ex. T.) Each individual application covers a three-month span from January 1, 2004 to December 81, 2008. (Id.; id. at Ex. U, at 1.) Each “work” consists of five tables of rates for various financial products, totaling approximately 400 different rates. (Id. Ex. S, Ex. U, at 1.) Mehl described the set of weekly tables that comprise each individual registration as a “[g]roup registration for database titled BANXQUOTE INDEX.” (Id. at Ex. S, at BX002117.) He identified each quarterly group of tables as a derivative work, and in the space where a registrant is required to identify any “preexisting work or works that this work is based on incorporates,” he wrote “Previously published database.” (Id. at BX002118.) Where he was asked to “give a brief general statement of the material that has been added to this work and in which copyright is claimed,” he typed “Weekly updates.” (Id.) Later, a representative of the Copyright Office notified Mehl that “the application does not clearly describe the new material on which the claim may be based.” (Id. at Ex. U.) The representative suggested that an appropriate statement of the new material in the work would be “ ‘revised compilation,’ ” and Mehl agreed. (Id.) The twenty works were then registered. (Id. at Ex. V.)

C. The Instant Motions

The Parties conducted discovery on the remaining claims. The Parties have now cross-moved for summary judgment on both claims. Plaintiff also submitted objections under Federal Rule of Civil Procedure 56(c)(2) to the admissibility into evidence of certain materials, and it moved for sanctions against Defendants for violation of the discovery rules. The Court held oral argument on all outstanding motions on September 17, 2018.

II. Discussion

A. Standard of Review

Before the Court are cross-motions for summary judgment. Summary judgment shall be granted where the movant shows that there is “no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” Fed. R.Civ.P. 56(a); see also Celotex Corp. v. Catrett, 477 U.S. 317, 322-23, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986). “When ruling on a summary judgment motion, the district court must construe the facts in the light most favorable to the non-moving party and must resolve all ambiguities and draw all reasonable inferences against the movant.” Dall. Aerospace, Inc. v. CIS Air Corp., 352 F.3d 775, 780 (2d Cir.2003); see also Tufariello v. Long Island R.R. Co., 458 F.3d 80, 85 (2d Cir.2006) (noting that a court must draw all reasonable inferences in the nonmovant’s favor).

A party seeking summary judgment bears the burden of establishing that no genuine issue of material fact exists. See Atl. Mut. Ins. Co. v. CSX Lines, L.L.C., 432 F.3d 428, 433 (2d Cir.2005). “When the burden of proof at trial would fall on the nonmoving party, it ordinarily is sufficient for the movant to point to a lack of evidence to go to the trier of fact on an essential element of the nonmovant’s claim. In that event, the nonmoving party must come forward with admissible evidence sufficient to raise a genuine issue of fact for trial in order to avoid summary judgment.” Jaramillo v. Weyerhaeuser Co., 536 F.3d 140, 145 (2d Cir.2008) (citations omitted).

Importantly for this case, “[w]hen the moving party has carried its burden under Rule 56(c), its opponent must do more than simply show that there is some metaphysical doubt as to the material facts.” Matsushita Electric Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 586, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986) (footnote omitted); see also Wrobel v. Cnty. of Erie, 692 F.3d 22, 30 (2d Cir.2012) (“To survive a motion under Rule 56(c), [plaintiff] need[s] to create more than a metaphysical possibility that his allegations were correct; he need[s] to come forward with specific facts showing that there is a genuine issue for trial.” (internal quotation marks and emphasis omitted)). A fact is material when “it might affect the outcome of the suit under governing law.” McCarthy v. Dun & Bradstreet Corp., 482 F.3d 184, 202 (2d Cir.2007) (internal quotation marks omitted). At summary judgment, “[t]he role of the court is not to resolve disputed issues of fact but to assess whether there are any factual issues to be tried.” See Brod v. Omya, Inc., 653 F.3d 156, 164 (2d Cir.2011) (internal quotation marks omitted). Thus, a court’s goal should be to “isolate and dispose of factually unsupported claims.” Celotex, 477 U.S. at 323-24, 106 S.Ct. 2548.

At the summary judgment stage, it is the “duty of district courts not to weigh the credibility of the parties.” Jeffreys v. City of N.Y., 426 F.3d 549, 554 (2d Cir.2005). Thus, even when a plaintiff has relied exclusively on his own testimony, courts have denied summary judgment— but only as long as the plaintiffs “testimony was not contradictory or rife with inconsistencies such that it was facially implausible.” Fincher v. Depository Trust & Clearing Corp., 604 F.3d 712, 726 (2d Cir.2010); see also Bridgewater v. Taylor, 832 F.Supp.2d 337, 345 (S.D.N.Y.2011) (denying summary judgment for plaintiff where defendant’s evidence consisted “solely of his own testimony,” but this testimony offered “a plausible alternate version of events”); Bennett v. Vaccaro, No. 08-CV-4028, 2011 WL 1900185, at *7-8 (S.D.N.Y. Apr. 11, 2011) (denying summary judgment where defendants did not establish that plaintiffs “testimony is, either on its face or in light of any other statements he has made, so self-contradictory or implausible as to rule out crediting it,” and there was no evidence that plaintiff “ever contradicted his current version of [events]”).

B. Copyright Infringement Claim

1. Overview

“ ‘To prevail on a claim of copyright infringement, the plaintiff must demonstrate both (1) ownership of a valid copyright and (2) infringement of the copyright by the defendant.’ ” Cameron Indus., Inc. v. Caravan, Ltd., 676 F.Supp.2d 280, 283-84 (S.D.N.Y.2009) (quoting Yurman Design, Inc. v. PAJ, Inc., 262 F.3d 101, 109-10 (2d Cir.2001)); see also Porto v. Guirgis, 659 F.Supp.2d 597, 608 (S.D.N.Y.2009) (requiring “ ‘ownership of a valid copyright, and [] copying of constituent elements of the work that are original’ ” (quoting Williams v. Crichton, 84 F.3d 581, 587 (2d Cir.1996))). It is undisputed that Defendants actually copied Plaintiffs individual averages. But Defendants vigorously dispute that they have copied anything protectable under federal copyright laws, because, among other arguments, the individual averages are unprotectable, discovered facts; they are uncopyrightable short phrases; and, even assuming the final values are in some sense “expressions,” the merger doctrine precludes their protection.

In resolving these issues, the Court first determines what material facts are in genuine dispute. Then, the Court surveys the law of copyright in factual material. Next, taking the facts in the light most favorable to the non-moving party, the Court explains why the averages are uncopyrightable facts. Finally, the Court explains additional why the averages are uncopyrightable under various other doctrines.

2. Plaintiff’s Products

Because the legal lines are so carefully drawn in this area, it is vital to understand in detail the nature of how Plaintiffs averages are computed, how they are presented to the public, and how they are used. Despite Plaintiffs efforts to muddy some of the waters, few material facts are in genuine dispute.

a. The Computation of Plaintiff s National Average Rates

According to Plaintiffs own website, Plaintiff “provides a family of widely followed indices and benchmarks that measure the rates and performance of banking, depository, mortgage, home equity and consumer loan markets.” (Mertzel Decl. Ex. 0, at BX 0048.) In other words, Plaintiff regularly surveys the interest rates or other prices offered by particular financial institutions across the country, and then compiles this data into various indices that represent “averages” or other important financial benchmarks.

Descriptive Statement - Subject and origins of the data:

U.S. national average rates quoted by the largest banks in all 50 states and Washington DC. as provided to BanxQuote and published continuously online at www.banxquote.com.

Approximate number of total data records: 400

Nature and frequency of changes: weekly updates

(Martial Deck Ex. S, at BX002125.)

Plaintiff compiles tables of averages organized by date, such as the one at the top of the following page. A variety of industry and general news publications described Plaintiffs product in a manner similar to that in which Plaintiff presented its own data. For instance, the record reveals that, in 2001, the Wall Street Journal’ s “Banxquote Banking Center” reported that Plaintiff “provides benchmark rates and pricing information of financial institutions throughout the United States.” (Mertzel Decl. Ex. N, at BX0091.) Newsweek, in an article on savings rates around the country, noted that Plaintiffs CEO Norbert Mehl “surveys rates on savings deposits nationwide.” (Mertzel Decl. Ex. P., at BX0105.) American Banker reported that Banxcorp is a firm that “monitors CD [i.e., certificate of deposit] rates,” and featured in its front-page story a graph representing “[y]ields on 6-month CDs” from January to September of 1989, crediting “Banxquote” as the data’s source. (Mertzel Decl. Ex. Q, at BX00107.) Indeed, the Wall Street Journal regularly included in its print edition a table of benchmark rates provided by Banxquote, such as one in the record titled “Banxquote Money Markets,” which has a subheading stating “Average Yields of Major Banks.” (Mertzel Decl. Ex. Q, at BX0065.)

The way the Banxquote indices are produced reflects their stated purpose: They are mathematical averages of the rates advertised by certain major financial institutions, updated at least weekly. Thus, a former software developer at Banxcorp named Abu Thomas testified that “if there are five banks,” then, to calculate its average rate, Banxcorp would “take the average of five banks.” (Lipkis Decl. Ex. 11, at 41.) The deposition continued:

Q: So you take the rate that each of the five banks is paying on money markets, add it up, and divide?

A: Yes.

Q: Simple mathematical average?

A: Yes.

Q: Is there any weighting of the banks included in the national average?

A: No.

(Id.) This method was independently confirmed by Defendants’ expert Bruce Webster, a computer scientist who examined Plaintiffs source code. (See Webster Decl. Ex. A, at 1.) Webster noted that the national average values copied by Defendants “are simple mathematical averages of reported rates, with no weighting or other calculations involved.” (Id.) In fact, only one actual computational function is used: a built-in database function called “AVGO,” which adds up the total of the values and divides by the number of entries. (Id. at 20.) Plaintiff has proffered no evidence that would show that the computation process is any more complex.

So the computational process is uncomplicated, and the output is but a single number on any given date. It turns out the inputs are equally straightforward: Plaintiff maintains a database into which someone inputs the interest rate or other relevant, publicly available financial information from one big bank in each state, plus one in Washington, DC, and then the software calculates an average. (Id. at 27.) From November 2002 to May 2007, for instance, the set of banks remained entirely consistent for the 5-year national average CD rate. (Id. at 28.) That, is, the input is simply the 5-year CD rate from “exactly the same set of banks, week after week, for 233 weeks.” (Id.) The story is nearly the same for the other average relevant to this case, the national money market rate. For that average, Plaintiff made only three changes to the input banks from November 2002 to May 2007. (Id.; see also Mertzel Reply Decl. Ex. K, at 1 (Mehl, in an email to Capital One representatives, stating that “BanxQuote calculates the average rates based on the largest banks in each of the 50 states and DC”); Mertzel Deck Ex. S, at BX002125 (Plaintiff’s submission to the Copyright Office stating that the data consisted of the “U.S. national average rates quoted by the largest banks in all 50 states and Washington DC”).)

The evidence supporting many of the facts above comes primarily from Plaintiffs own website, the deposition of its own former employee, and the sole expert report submitted on this issue. Plaintiff in its 56.1 statement and Mehl in his deposition dispute some aspects of this account, but — in addition to being extremely confusing and rife with legal propositions couched as factual differences — Plaintiffs factual account, to the extent it differs from anything discussed above, is “contradictory [and] rife with inconsistencies such that it [is] facially implausible.” Fincher, 604 F.3d at 726. In other words, Plaintiffs deposition testimony fails to create any genuine dispute regarding the facts of how the averages are created, or of their perception as factual representations of the national average rates by the financial and general interest media. See Fed. R.CivJP. 56(a).

Defendants offered the following as a statement of undisputed material fact: “Banxquote provides benchmark rates and pricing information on financial institutions throughout the United States.” (DSUF ¶ 132.) As explained above, that statement is amply supported by the record. But Plaintiff attempts to create some factual dispute over that fact:

BanxCorp does not dispute this statement, to the extent that the BanxQuote indices pertain to the creation of nonbinding indices used to predict or estimate the performance of bank money market savings and CD rates in the United States, rather than the discovery of facts or actual national average bank rates in a literal sense. The phrase “national average bank rates” or “benchmark rates” is a paradoxical colloquial or figurative expression or arguably an oxymoron since there are thousands of banks in the United States. In addition, the disclosure of individual bank rates or recording of national average interest rates paid by banks are not compulsory or required by law, [and] there is a broad range of numerous possible variations not based on the same or substantially similar underlying market facts or singular form of expression. BanxCorp further refers to its response to [16 other factual statements]. [Citing its Response to Request for Admission and its Copyright Registrations.]

(Ph’s Resp. to DSUF ¶ 132.)

The actual disputed facts are mostly obscured in that confusing response, and the explanation Mehl gives in his deposition is no more straightforward. Consider the following important exchange. The question that triggers the response that Mehl provides at the beginning of the quotation below pertains to what he thinks it means when Banxquote reports that, for instance, “.23 percent” is the national average interest rate for money market accounts on a given date.

A: First of all, we just put numbers there. People use it for different purposes. So what they try to measure or how they try to measure, we cannot control. So we use a certain system to show their performance over time.

Q: What is .23 telling me? What are you trying to tell me?

A: I don’t know what it tells you. I know what it tells me.

Q: What are you saying when you say .23? What is .23?

A: It shows an index based on certain banks that we track and certain indices that we track that are in our system. And then it uses — and it shows it over time.

Q: But today the .23 is not over time. I’m asking what the .23 is telling me. What are you telling me is .23?

A: It could be what it would have been if you had left the money invested for a year.

Q: Is the average that banks are paying on money markets something that can be measured?

A: Theoretically, it could be____ First of all, it depends on what the meaning of the word “average” is, okay. So when it comes to the average for money markets, the word “average” has been used as a colloquial term by various publishers and media, including us. And each one attempts to show some index of certain banks that they track. So that’s why the word “average” in that context has a different meaning than what the word “average” would have in another context....

So that’s one qualification. The second is that in order to compute an average, you would have to track the interest rate of every bank in the United States. So in theory, it may be possible, but in reality, it does not seem to be possible, because of the number of banks and the number of — and the variety of money market products. So there is no set convention that would allow somebody to measure an actual average. So you could measure an estimate at best.

(Lipkis Decl. Ex. 6. at 88-91.)

Plaintiff thus appears to be putting forth a contrary factual account of how Plaintiffs averages were described by Plaintiff, by Mehl, and by the media. But Mehl’s testimony on this point has no actual support in the record, and indeed is directly contradicted by a variety of documentary evidence. For instance, Mehl’s statement that Banxcorp “just puffs] numbers there” is patently wrong: as explained above, Plaintiff published not only “numbers”— Plaintiffs tables of average rates — but Plaintiffs own website contains a variety of detailed explanations of what the data represents. (Mertzel Decl. Ex. 0, at BX0048.) And its standard License Agreement, which both Plaintiff and Capital One signed, states that “Banxquote agrees to take all reasonable actions necessary to keep BanxQuote data current, accurate, true and complete, and to notify [Capital One] of any errors or omissions.” (Lipkis Decl. Ex. 1-A, at 2.) Thus, Mehl’s testimony that he held out the BanxCorp averages as only estimates or predictions is belied by his own promise on a document he personally signed to keep his “data current, accurate, true and complete.” No reasonable jury could believe Plaintiffs testimony on this point.

Also facially implausible is the statement in Plaintiff’s Response to Defendants’ Statement of Undisputed Material Facts that the “phrase ‘national average bank rates’ or ‘benchmark rates’ is a paradoxical colloquial or figurative expression or arguably an oxymoron since there are thousands of banks in the United States.” A mountain of incontrovertible documentary evidence in the record shows that these phrases are not “paradoxical colloquials”— whatever that means — but are instead the exact phrases used by Banxcorp itself, by Mehl, and by the entire financial media to formally describe the averages. (Mertzel Deck Ex. 0, at BX0048); see Dzanoucakis v. Chase Manhattan Bank, USA, No. 06-CV-5673, 2009 WL 910691, at *8 (E.D.N.Y. Mar. 31, 2009) (where the “uncontroverted record clearly supports a [particular] finding,” then “[p]laintiff s own self-serving declaration to the contrary is insufficient, under the circumstances, to raise a triable issue of fact”).

In addition, Plaintiff and Mehl get quite philosophical about the nature of an “average.” But the point they are making about the “hypothetical” nature of an average is both factually wrong and logically fallacious; this testimony, too, fails to create a genuine dispute of material fact about Plaintiffs product. For one thing, Plaintiffs statement is facially implausible as a matter of describing the way the term “average” was actually used, as the record shows that both Plaintiff as a company, Mehl himself, and the media used the term “average” in more than the “colloquial” sense; rather, every relevant piece of documentary evidence in the record other than Mehl’s self-serving deposition testimony reveals that, to those well-versed in financial markets, Plaintiffs data was treated as an “average” in the non-colloquial, mathematical sense. (E.g., Mertzel Decl. Ex. N, at BX0091.) And that media treatment was entirely justified, because Mehl’s statement is fallacious as a matter of mathematics. After all, the numbers that Plaintiff produced are unquestionably averages in the technical sense: any given data point reported by Plaintiff is an exact, mathematical average of the rates that form the input into the database, computed using the database function “AVG().” (Webster Deck Ex. A, at 1, 20.) Indeed, there is no testimony that contradicts Defendant’s factual account of how the averages were calculated. True, there is an entirely separate question of whether Plaintiffs particular average, taken from a set of large, geographically diverse banks, is a good approximation of what Mehl refers to as the “actual average” of “every bank in the United States.” But the answer to that question in no way affects the mathematical reality and the incontrovertible fact that Banxcorp’s data points are simple averages. Thus, no reasonable jury could conclude that each data point is anything other than a mathematical average. The legal consequences of the fact that Plaintiffs input does not include data from every single bank in the United States are discussed in the next sub-section.

Plaintiffs assertion that somehow the data are predictions of future performance — that, in Mehl’s words, an individual average of “.23%” in fact “shows performance over time” — is also without support. As explained, all the evidence in the record is that “,23%” is the average of the roughly current rate that certain financial institutions are paying on their deposits. (E.g., Mertzel Deck Ex. Q, at BX0065 (showing the “Banxquote Money Markets” table of average yields in the Wall Street Jour nal).) Plaintiff attempts to portray each number as somehow containing an uncertain temporal component, apparently because an interest rate is something that is ultimately compounded and the resulting money is paid out at a future date. But the fact that yesterday’s average rate was .23% has ramifications for depositors toT day, tomorrow, and in five years does not mean that yesterday’s rate of .23% is not a simple historical fact, fixed forever at that particular point in time. No reasonable jury could find that an individual data point represents “performance over time” in any relevant sense, because no future financial upheaval could ever change the fact that the average rate being offered as of a particular date was a particular number.

As a final matter with regard to the factual characterization of the data, Plaintiff makes much of the fact that competing providers of data — namely, those provided by the companies Informa or Bankrate, Plaintiffs competitors — produced national average rates that differ somewhat from Plaintiffs data. Plaintiff includes an expert report on damages containing an exhibit showing “Differences between Banxquote, Bankrate, and Informa Indices.” (Lipkis Decl. Ex. 77, at Ex. C.) In this section, there are a few exhibits that do indeed show some variation in reported rates: On one particular date in August 2008, the three indices diverged by anywhere from approximately .20 to .59 percentage points, depending on the index. (Id. at COB0014998.)

The Court, of course, takes this evidence at face value, and the Court construes the divergence to be as large as Plaintiffs evidence permits. But it should be clear that any comparison with other averages does not change the nature of Plaintiff’s calculation at all: that is, how others might calculate a national average CD rate does not change the nature of the evidence regarding how Plaintiff actually produces its data.

b. Summary: Undisputed Material Facts Regarding Plaintiffs Averages

In sum, the record shows that the following material facts regarding Banxquote’s national averages for the money market and CD data are not in genuine dispute:

1. Plaintiffs input to the averages was, for over four years, the most recently published rates of major banks. These rates were always publicly available and are objectively verifiable facts about what interest rate a given bank is offering at a particular moment in time. These rates are inputted into Banxquote’s database approximately weekly. The particular banks used as inputs changed infrequently, if ever, during the period at issue. Some essentially trivial conversion of published bank rates may have been necessary to ensure that the rates were entered in a standard format.

2. Once the rates were inputted and standardized, the software calculated a simple mathematical average of the rates. No weighing or any other, sophisticated calculation or algorithm was used. The built-in “AVG( )” function was the only meaningful computer function used. No financial data was used as part of the input to the calculation except the particular interest rate of a particular financial institution as of a particular date.

3. The output of the calculation was a single number that is the exact mathematical average of the inputted rates as of a particular date. These outputs could then be compiled into a table, organized by date.

4. Plaintiff represented to consumers, customers, and the financial media that the averages were objective facts about average national interest rates as of a particular date. Plaintiff promised licensees of its data that it would ensure the data was “current,” “accurate,” and “true.” The averages were called “benchmark rates,” “national average rates,” or other similar terms by Plaintiff, by Mehl, and by the financial and general interest press.

As described, the only evidence that might contradict these facts comes entirely from Mehl’s own affidavits or deposition testimony. “But a self-serving, contradictory affidavit fails to raise a triable issue of fact when it conflicts with documentary evidence.” Christiana Bank & Trust Co. v. Dalton, No. 06-CV-3206, 2009 WL 4016507, at *4 (E.D.N.Y. Nov. 17, 2009); see also Dzanoucakis, 2009 WL 910691 at *8 (where the “uncontroverted record clearly supports a [particular] finding,” then “[plaintiffs own self-serving declaration to the contrary is insufficient, under the circumstances, to raise a triable issue of fact”). There is no triable issue of fact regarding how Plaintiffs data is computed or how it has been presented to and understood by the public. Like many other cases implicating copyright in largely factual material, this case is ripe for decision at summary judgment. See New York Mercantile Exchange, Inc. v. IntercontinentalExchange, Inc., 497 F.3d 109, 119 (2d Cir.2007) (deciding on summary judgment a case presenting the question whether settlement prices for futures contracts are entitled to copyright protection); RBC Nice Bearings, Inc. v. Peer Bearing Co., 676 F.Supp.2d 9, 24 (D.Conn.2009) (deciding on summary judgment a case presenting the question whether load ratings for ball bearings are entitled to copyright protection).

3. The Legal Landscape of Copyright in Factual Material

With the details of how Plaintiffs averages are computed and perceived in mind, it is time to turn to the legal question of whether Plaintiffs averages are protectable under the Copyright Act. The Court discussed the general legal landscape regarding copyright protection of factual material at some length in its earlier decision in this case. BanxCorp., 723 F.Supp.2d at 601-09. The discussion here tracks that discussion in many ways but also adds to it to include recent developments and additional analysis now that the issues in the case have been crystallized.

The key legal starting point is the deceptively simple proposition that “facts are not copyrightable.” Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., Inc., 499 U.S. 340, 344, 111 S.Ct. 1282, 113 L.Ed.2d 358 (1991). In Feist, the seminal Supreme Court precedent regarding the copyright-ability of factual material, the Court theorized that no copyright can exist in facts “because facts do not owe their origin to an act of authorship[,] ... [and are] not created[,] ... [but] merely discovered....” Id. at 347, 111 S.Ct. 1282. But while originality “remains the sine qua non of copyright,” “[fjactual compilations ... may possess the requisite originality ... [where] [t]he[ ] choices as to selection and arrangement ... are made independently by the compiler and entail a minimum degree of creativity.” Id. at 348, 111 S.Ct. 1282. So the key line the Court drew in Feist is the one between uncopyrightable facts themselves and their arrangement, which may contain some elements of protectable originality. As applied by the Court in Feist, the white pages in a telephone directory contained only unprotectable facts arranged in an entirely unoriginal way. See Feist, 499 U.S. at 362, 111 S.Ct. 1282. The defendant was found not liable for copying at least 1,309 entries from the plaintiffs telephone directory and incorporating those entries into its own directory. See id.

The cases applying Feist have made clear that some propositions are less obviously factual than the indisputably factual proposition that, say, the President of the United States resides at 1600 Pennsylvania Avenue in Washington, D.C. For instance, in New York Mercantile Exchange, Inc. v. IntercontinentalExchange, Inc., 497 F.3d 109 (2d Cir.2007), the Second Circuit was presented with the question of whether settlement prices for futures contracts were uncopyrightable facts. Id. at 114. While it might seem that the answer should be “yes, of course settlement prices are facts,” it turns out that the determination of the settlement prices by the plaintiff exchange is fairly complicated. According to the Second Circuit,

A futures contract requires the delivery of a commodity at a specified price at a specified future time, though most contracts are liquidated before physical delivery occurs.... The settlement prices are used to value the open positions---Unlike on a securities exchange, the settlement price may not be the final trade, for two reasons. First, because of the nature of trading, it is not always clear which trade was the closing trade.... Second, ... [flor the “outer” months, those further from the trading date, there is often little or no trading on a particular day---- For high-volume months, settlement prices are based on a formula: “a weighted average of all trades done within the closing range.” ... For low-volume months, the extent of the ... creative judgment is disputed.

Id. at 110-11 (footnotes omitted). The Second Circuit ultimately decided the case on the alternative ground that the merger doctrine barred copyright protection, id. at 115, but the Second Circuit stated in well-considered dicta that “there [wa]s a strong argument” that the settlement prices were unprotectable facts, id. at 114, though that conclusion was less certain for the low-volume months, id. at 116. For those low-volume months, the Court stated that, because “there is no real market to speak of,” the settlement prices “appear[ ] closer to creation, to making predictions of expected values.” Id. (internal quotation and ellipsis omitted). By contrast,

For high-volume months, settlement prices are determinations of how the market values a particular futures contract ... [,] not how the market should value them or will value them. Under this view, the market is an empirical reality, an economic fact about the world.... So characterized, there is one proper settlement price; other seemingly-accurate prices are mistakes which actually overvalue or undervalue the futures contract.

Id. at 115 (emphasis in original). Therefore, consistent with the dicta in New York Mercantile, when confronted with raw data that have been converted into a final value through the use of a formula, courts should put significant weight on the degree of consensus and objectivity that attaches to the formula to determine whether the final value is fundamentally a “fact.” See Columbia Broad. Sys., Inc. v. Am. Soc. of Composers, Authors & Publishers, 620 F.2d 930, 935 (2d Cir.1980) (“[Appellate courts ... have an entirely legitimate function of elucidating principles of law, fairly raised by litigation, even if the resulting pronouncements are not absolutely required for the precise decision reached. Appellate guidance is not valueless because it is dictum.”).

In New York Mercantile, the Second Circuit contrasted the settlement prices at issue with a “compilation of estimated projections for used car prices” that the Second Circuit had previously held merited copyright protection as a group in CCC Information Services, Inc. v. Maclean Hunter Market Reports, 44 F.3d 61 (2d Cir.1994). See N.Y. Merc., 497 F.3d at 115 n. 5. The crucial distinction between the two cases was that “[t]he values [in Mac-lean Hunter ] were based on assumptions about ‘average’ cars; as these cars did not exist, there could be no actual market to discover.” Id. By contrast, “settlement prices can be seen as ‘pre-existing facts’ about the outside world which are discovered from actual market activity.” Id.

New York Mercantile and Maclean Hunter together provide helpful guideposts in determining the copyright status of price data. If the data purports to represent actual objective prices of actual things in the world — the actual price of an actual settlement contract on a particular day — it is an unprotectable fact; if the data purports to represent an estimated price of a kind of idealized object — for instance, what a hypothetical, mint condition 2003 Ford Taurus with approximately 60,000 miles might be worth — then the hypothetical price may be eligible for some form of copyright protection in the right circumstances. See Maclean Hunter, 44 F.3d at 71 (distinguishing between “building-block” ideas “that undertake to advance the understanding of phenomena or the solution of problems ... and those ... that do not undertake to explain phenomena or furnish solutions, but are infused with the author’s taste or opinion”).

To illustrate this distinction, this Court gave the following example in its previous opinion. If a scientist knew an object’s mass and the force acting upon the object, this raw data could be converted into the object’s acceleration due to that force by using the “formula” known as Newton’s Second Law of Motion. This use of a formula would merely discover an “empirical reality,” and therefore the result would be uncopyrightable. This is true even if the resulting output is not completely accurate, so long as the formula used is generally accepted and quintessentially objective. Thus, the output data generated by using Newton’s Second Law of Motion — force equals mass times acceleration, or “F=ma” — would be a series of uncopyrightable facts, even though the output is in some sense an estimation because Newton’s formula fails does not consider relativistic effects. See Albert Einstein & the Theory of Relativity, http://csepl0.phys. utk.edu/astrl61/lect/history/einstein.html (last visited September 26, 2013) (lecture from “Astronomy 161” course at the University of Tennessee, Knoxville).

Since New York Mercantile, there has been one published opinion from a district court in the Second Circuit that has attempted to navigate these tricky waters. In RBC Nice Bearings, Inc. v. Peer Bearing Co., 676 F.Supp.2d 9, 21 (D.Conn. 2009), the court considered the copyright-ability of “load ratings” of ball bearings, which are measures of the “radial force a particular bearing having known geometric and physical attributes, such as size and quantity of balls, can withstand.” Id. at 16 (internal quotation marks omitted). The exact values of the load ratings were “mainly a function of the geometry of the bearing and material, [but also accounted for] certain other ‘life factors’ enumerated in published industry guidelines ... [such as] tolerances, material cleanliness, lubrication, hardness, and operating temperature.” Id. The plaintiffs strongest argument that the load ratings were not mere uncopyrightable facts was that “creativity [wa]s used in developing the load ratings ... [, because] certain bearing manufacturers use the various ‘life factors’ ... to adjust their load rating calculations from a standard calculation based only upon the geometrical features of the bearings.” Id. at 22. However, the court found that argument unpersuasive in the light of New York Mercantile, because

[w]hile there may be some level of judgment involved in selecting which particular “life factors” to utilize in adjusting the standard load rating calculation, based upon the record before the Court such judgment is very minimal given that the relevant life factors are published in industry guidelines. The level of judgment necessary to calculate the load rating information is undoubtedly no more than that needed to determine the settlement prices at issue in New York Mercantile ....

Id. Just as in New York Mercantile, though, the court did not rest its decision solely on this ground, and it stated that it would reach the same decision even if “the load bearing ratings are expressions rather than facts,” because of the court’s application of the merger doctrine. Id. at 23. The court’s reasoning, therefore, like the reasoning in New York Mercantile, could be considered dicta in some sense.

All of the authorities discussed so far were available to the Court at the time of its earlier decision in this case. Since then, the Second Circuit has not decided any additional on-point copyright cases, but its decision in Barclays Capital Inc. v. Theflyonthewall.com, Inc., 650 F.3d 876 (2d Cir.2011), provides some additional illumination. In that case, the Second Circuit held that the defendant was not liable under the state-law doctrine of hot news misappropriation for reproducing plaintiffs’ “actionable” stock ratings — for example, an analyst’s downgrade of or recommendation to buy a particular stock— without authorization. Id. at 881. While the majority did not discuss whether the ratings were uncopyrightable facts for purposes of the Copyright Act in particular, the majority opinion described the defendant as “collecting, collating and disseminating factual information — the facts that [plaintiff brokerage firms] and others in the securities business have made recommendations with respect to the value of and the wisdom of purchasing or selling securities — and attributing the information to its source. The [plaintiffs] are making the news; [the defendant], despite the [plaintiffs’] understandable desire to protect their business model, is breaking it.” Id. at 902 (emphasis removed). As one leading copyright scholar noted, the case represents “[t]he most stunning example of judicial skepticism of ratings” to date. James Grimmelmann, Three Theories of Copyright in Ratings, 14 Vand. J. Ent. & Tech. L. 851, 865 (2012); see also id. (noting that the Barclay’s opinion, despite dealing with state-law misappropriation, “casts grave doubt on copyright protection of influential ratings”).

In 2010, the Court summarized the doctrine as follows. Where: (1) the raw data used to create the final value were unprotectable facts; (2) the method of converting raw data into the final value was an industry standard, or otherwise widely accepted as an objective methodology; and (3) the final value attempted to measure an empirical reality, then the final value produced from raw data ordinarily is not protected by copyright. BanxCorp, 723 F.Supp.2d at 604. That summary not only appears to remain good law today, but it has been reinforced by the idea in Bar- day’s that the law should not allow a select group of people to gain a legal monopoly on information widely influential as a benchmark in financial markets. See Barclays, 650 F.3d at 896 n. 29 (noting that allowing plaintiffs to block the reproduction of their recommendations “would ensure that the authorized recipients of the [rjecommendations would in significant part be profiting because of their knowledge of the fad of a market-moving [r]ecommendation before other traders learn of that fact” (emphasis in original)).

A Plaintiff’s Averages Are Uncopyrightable Fads

Plaintiffs legal argument that any given individual average is protectable as a sufficiently original work withers away in light of the evidence of what Plaintiffs data is and the factual findings explained above. Each average is a fact, plain and simple: It is the national average rate of interest offered by major U.S. banks on a given financial product at a given point in time based on publicly available data. See supra Section II.B.2. That is how Plaintiff held out its averages over the relevant time period, how the media interpreted and reported on them, and how any relevant consumer would have understood them. Thus, on the spectrum from fact to estimate suffused with judgment and opinion outlined above, Plaintiffs data is legally equivalent to the unprotectable load ratings in RBC Nice Bearings, the likely unprotectable settlement prices in New York Mercantile, and the likely unprotectable analyst recommendations in Bardays. By the same token, Plaintiffs list of averages are unlike the protectable list of estimated prices of hypothetical used cars at issue in Maclean Hunter.

Plaintiff has not created a genuine dispute of fact that the type of judgment that would infuse the data with “originality” goes into the calculation of each individual average. Rather, Plaintiff inputs the relevant rates and the software runs an average, which Plaintiff then publishes verbatim. And any “judgment” that went into the initial selection of banks was both extremely straightforward — one large bank in each state and the District of Columbia — and infrequent — Plaintiffs list of banks did not change at all for four years for one of the averages at issue. Applying the specific three-part test stated above, (1) the raw data used to create the final value consists entirely of unprotectable facts; (2) the method of converting raw data into the final value is an industry standard and widely accepted as an objective methodology, because the method involves merely tracking the interest rates offered by large banks and computing a “simple mathematical average” of the inputted rates; and (3) the final value clearly attempts to measure an empirical reality. See BanxCorp, 723 F.Supp.2d at 604. Each individual average is thus an uneopyrightable fact.

Instead of trying to parse the doctrine differently, Plaintiff instead claims that the averages are protectable based solely on its factually unsupported view of its averages. “The Banxquote Indices are Purely Estimates or Predictions,” states one subheading in Plaintiffs principal brief. (Pl.’s Mem. 13.) But, as explained, this statement is contradicted by the record. In fact, Plaintiffs averages are not predictions at all; they are reports of historical interest rates offered by financial institutions. And, mathematically, they are not estimates at all; they are computed by computing the exact arithmetic mean of all the input values.

The only possible sense in which these averages could be considered “estimates” is by taking into account the representation or purpose of the data as it is presented in various publications of the data series — including the presentation in Defendants’ advertisements as the “national average rate.” That is, while any given value calculated by Plaintiff is exactly the average of the rates charged by the particular input banks, using this value to represent the “National Average Bank Rate” is in some sense an estimation, or perhaps a kind of shorthand, because Plaintiffs calculation does not use as input the average rate of every single financial institution in the United States. But the fact that not every bank’s interest rate enters the calculation is not nearly enough to move a reported average from the column of fact to that of judgment or opinion. After all, very often, data fails to be perfectly representative or entirely complete relative to what it is supposed to measure, but the data nevertheless remains fundamentally factual.

For instance, no white pages directory lists every single person living in a particular area, or gets every address, phone number, and name exactly right — indeed, the white pages at issue in Feist even contained four fictitious listings, inserted to detect copying — but that does not make the white pages a work of opinion regarding who lives in a given area. See Feist, 499 U.S. at 344, 111 S.Ct. 1282. Likewise, in a case about the census that did not address copyright issues, the Supreme Court acknowledged that no population census can possibly capture everything about the population it surveys with complete accuracy. See Dep’t of Commerce v. U.S. House of Representatives, 525 U.S. 316, 322, 119 S.Ct. 765, 142 L.Ed.2d 797 (1999) (describing the Census Bureau’s methods for compensating for the “under-count,” which is the portion of the population not directly surveyed either in person or by mail). And yet the Supreme Court stated in Feist that “[cjensus data ... do not trigger copyright” because “[cjensus takers ... do not ‘create’ the population figures that emerge from their efforts; in a sense, they copy these figures from the world around them.” Feist, 499 U.S. at 347, 111 S.Ct. 1282. So too here. Each average at issue in this case is a fact about the world — an “empirical reality” — even though it is in some sense an imperfect representation of some platonic ideal of a “national average bank rate.”

This explains why the fact that there are several competing companies that measure national average rates, all of which regularly computed slightly different final values, does not mean that Plaintiffs output is not fundamentally factual in nature. The difference between two particular values, according to an email in the record, likely arises “due to the fact that the Informa national average is $10k [i.e., for accounts with a minimum balance of $10,000] and Bankrate’s has no min [i.e., no minimum balance].” (Lipkis Decl. Ex. 77, at COB0014996.) Though not entirely clear from the record, the best inference is that each provider of national averages is actually collecting and computing a slightly different average, perhaps because each provider thinks that its own input is more relevant to its consumers. For one company, the relevant metric is the interest rate large banks pay on CDs with a $10,000 minimum deposit, and, for another, it is the interest rate large banks pay on CDs with no minimum deposit. These d