TNS Media Research, LLC v. Tivo Research & Analytics, Inc., 2015 U.S. App. LEXIS 16517 (Fed. Cir. Sept. 16, 2015):

Tivo Research and Analytics, Inc. dba TRA, Inc. ("TRA") appeals a judgment of the district court granting summary judgment in favor of TNS Media Research, LLC dba Kantar Media Audiences and Cavendish Square Holding B.V. (collectively, "Kantar"). Kantar initially filed suit in the district court, seeking a declaratory judgment that it did not infringe U.S. Patent No. 7,729,940 (the "'940 Patent"). Tivo Research and Analytics, Inc. dba TRA, Inc. ("TRA") counterclaimed, asserting infringement of the '940 Patent, and also U.S. Patent Nos. 8,000,993 (the "'993 Patent"); and 8,112,301 (the "'301 Patent"), misappropriation of trade secrets, and breach of contract and fiduciary duty claims against Kantar. The district court determined at summary judgment that Kantar's two categories [*2]  of accused products--the Auto Products and the Consumer Packaged Goods ("CPG") Products--did not infringe the patents-in-suit, that Kantar did not misappropriate TRA's trade secrets, that TRA could not rely upon its damages expert's testimony or report to support its claim for damages with respect to its non-patent claims, and that TRA could not seek punitive damages against Kantar. Ultimately, the district court determined that TRA only could pursue a request for nominal damages for its remaining breach of contract and fiduciary duty claims, but, was not entitled to do so before a jury. The parties agreed to settle that remaining claim, however, and the court entered final judgment.

We affirm-in-part, reverse-in-part, vacate-in part, and remand. Specifically, because the district court correctly determined that TRA's Auto Products do not meet the "double blind matching" limitation, we affirm this part of the judgment. But because the district court's non-infringement ruling as to Kantar's CPG Products was based on a disputed stipulation, we vacate this decision and remand for further proceedings. We also reverse the district court's decision to dismiss TRA's claims for misappropriation [*3]  of trade secrets as a discovery sanction, and vacate and remand its alternative ruling that TRA's client secrets and its TRAnalytics product are not protectable. We reverse the district court's ruling that TRA's financial projections and strategic plans are not protectable as a matter of law. We affirm, however, its determination that TRA's product positioning secrets are not protectable as a matter of law. We also affirm the district court's exclusion of the "frozen market" opinion expressed by TRA's damages expert. But, we reverse its determination that, without this report, TRA submitted insufficient evidence to support a claim for compensatory damages. We also reverse the district court's conclusion that TRA was only entitled to nominal damages on its non-patent claims after its summary judgment ruling, thereby mooting the district court's determination that TRA was not entitled to a jury trial as to those claims. And, we reverse the district court's conclusion that TRA is not entitled to injunctive relief on its fiduciary duty claims as a matter of law. Ultimately, although we agree with some of the district court's rulings, we conclude that TRA has a right to a jury trial on [*4]  at least a subset of its claims. Finally, we reject TRA's request that the case be reassigned to another judge on remand.

Background

Companies spend billions of dollars on advertising each year hoping to reach potential customers. But it can be difficult to determine how effective these advertisements are. TRA sought to address this problem in the television context by developing techniques for processing television viewing data and consumer purchasing data to create reports that can be used to determine what households watch and what they buy. TRA claims its solution, implemented in its Media TRAnalytics product, allows companies to target their advertisements more strategically and to assess the effectiveness of these ads. This solution is protected by the patents-in-suit, which all relate to systems and uses of consumer data in advertising.

A. Patents-in-Suit

Specifically, the '940 Patent is directed to a method for collecting, matching, and analyzing television viewing and consumer purchasing data to create reports determining the return on advertising investments and other metrics. Although there were other prior art methods for performing such a task, the '940 Patent differed in that it claimed a method [*5]  that did not require the installation of supplemental equipment in homes or retail locations or a consumer's "opt-in" permission, all while maintaining the privacy of the consumer.

***

The '993 Patent, titled "Using Consumer Purchase Behavior for Television Target," is similar to the '940 Patent, and is directed to a system for facilitating the analysis of consumer data associated with advertising exposure. Likewise, the '301 Patent is directed to a related invention, a computer-implemented method and system directed towards facilitating the analysis of consumer behavior associated with advertising exposure.

To help aid the growth of the company, TRA sought outside investment. It successfully went through three rounds of financing, the first resulting in a post-money valuation of roughly $10 million in August 2007, the second resulting in a valuation of roughly $27 million [*8]  in May 2009, and the third resulting in a valuation of roughly $54 million in May 2010. Kantar, a market research company, participated in these rounds via its investment arm, Cavendish Square Holding B.V., investing a substantial sum in each round. These investments granted Kantar access to TRA's board which, in turn, gave it access to TRA's trade secrets.

During this time, the companies also engaged in merger discussions, but these plans fell through shortly after Kantar spent almost two billion dollars acquiring TNS Media Research, a competing market analytics company. Soon after Kantar's acquisition of TNS Media Research, Kantar released its own analytics product, which directly competed with TRA's Media TRAnalytics product. It was during this time that TRA attempted to undergo a fourth round of financing, but was unable to raise the necessary funds. This led to a substantial drop in value and, subsequently, TiVO purchased TRA for approximately $20 million in July 2012.

B. District Court Proceedings

Believing that Kantar's competing product was in fact infringing its patent rights, TRA contacted Kantar in March 2011 about the product. In response, Kantar filed a declaratory judgment [*9]  action, seeking a judgment that it did not infringe the '940 Patent. TRA answered and counterclaimed, alleging: (1) patent infringement of the '940, '993, and '301 Patents; (2) misappropriation of trade secrets; (3) breach of contract; and (4) breach of fiduciary duty. The district court ultimately resolved every one of the many issues raised by these pleadings as a matter of law. A detailed description of the court's multiple and interrelated rulings is, thus, necessary to understand the issues we address on appeal.

1. Patent Claims

At the time of the district court's November 2013 summary judgment order, TRA asserted that Kantar's two types of products--the Auto Products (comprising Kantar's Rapidview Auto and Charter with Auto products) and the CPG Products (comprising its Rapidview Retail, Rapidview for Retail, and Charter with CPG products)--directly infringed claim 71 of the '940 Patent, claims 1, 2, 3, 7, 8, and 9 of the '993 Patent, and claims 1, 23, 42, 47, 49, 63, 108, and 109 of the '301 Patent. TNS Media Research, LLC v. TRA Global, Inc., 984 F. Supp. 2d 205, 209-14 (S.D.N.Y 2013) ("Summary Judgment Op."). Kantar's Auto Products match information regarding motor vehicle registrations maintained by J.D. Power & Associates ("J.D. Power") with television viewing data, while its CPG Products match data collected from supermarket [*10]  loyalty cards with television viewing data. Kantar countered that none of its accused products directly infringe the asserted claims, alleging that: (1) its Auto Products did not perform double blind matching of data as required by all asserted claims; (2) its CPG Products did not collect "purchase data" as all fifteen asserted claims require; (3) TRA waived any argument that its CPG Products infringed under the doctrine of equivalents; (4) Kantar's accused products did not meet the "cleansing and editing algorithm to the matched and stored data" limitation of claim 71 of the '940 Patent; and (5) that none of the accused products meet the claim limitation embodied in claim 42 of the '301 Patent. The district court found that summary judgment was warranted based on the first three arguments, and declined to address the other two.

With respect to the "purchase data" limitation, the parties stipulated that "purchase data" meant "data describing the purchase of a particular product at a given time, obtained from a purchase data source, such as a shopping loyalty card, point of sale collection means, or other record of a sale of a product or service." Id. at 218 (emphasis added). Although Kantar conceded that its CPG [*11]  Products did utilize information from "purchase data source[s]," it argued that its products did not meet the "purchase data" limitation, because they did not collect data about the time at which a purchase was made, but only categorized the users into types, such as heavy, medium, or light user. Id.

Because the dispute between the parties concerned how to interpret the term "purchase data," the district court found the issue was purely one of claim construction, and, thus, was amenable to summary judgment. Id. at 243. It determined that the asserted patents distinguish between "purchase data" and "user types." For example, the '301 Patent explains that "[t]he ROI [return on investment] report may use the following data as inputs, for example: . . . purchase data; user type (heavy, medium, or light category purchase rate). . . ." 301 Patent col. 29:48-51). Although TRA argued that this distinction was meaningless because a classification of users into types necessarily must consider the users' purchase of products over a given time, the district court disagreed, ruling that "user type" only conveys relative purchasing information, whereas "purchase data" requires data about when an actual purchase was made. Summary Judgment Op., 984 F. Supp. 2d at 243. Additionally, the [*12]  district court rejected TRA's allegation that Kantar should be estopped from arguing that it does not meet the "purchase data" limitation, because Kantar argued at the preliminary injunction stage that its current method of matching viewer data to purchase data was nearly identical to what it had been doing before TRA received a patent. Under applicable Second Circuit law, judicial estoppel only applies when the court relies on a party's prior inconsistent arguments. Here, because the district court did not rely upon Kantar's assertion in denying TRA's motion for a preliminary injunction, it declined to find that TRA was estopped from raising its present argument. Id. at 244.¹ Accordingly, the district court concluded that Kantar's CPG Products did not literally infringe the asserted claims.

1   This estoppel issue is not before us on appeal.

The district court also concluded that the CPG Products did not infringe the asserted claims under the doctrine of equivalents. It found that TRA's expert report only made conclusory statements about the doctrine of equivalents and did not mention the "purchase data" limitation in this discussion. Id. at 244. Because it felt that TRA failed to explain this theory [*13]  adequately in its report, the district court found that TRA could no longer assert it. Although TRA attempted to amend its expert report by submitting an expert declaration during summary judgment briefing, the district court found that these untimely disclosures were barred under Fed. R. Civ. P. 37(c). Thus, the district court also dismissed TRA's doctrine of equivalents infringement theory for the CPG Products.

Regarding the accused Auto Products, both sides agreed that all the asserted patent claims require double blind matching. Double blind matching is a process where no data supplier provides both behavioral data--in this case automobile registration data and television viewing data--and personally identifying information--such as names--to the same party at the same time. See '940 Patent, col. 9:26-34 ("[N]o single party has access to both household identity and household purchase or viewing behavior. A party that knows a household identity, for example, will not know the behavior of the household; likewise, a party that knows the behavior of a household will not know the identity of that household."). Instead, one data supplier will provide a trusted third party with personal information and an abstract identification [*14]  code for each household, while a second data supplier will provide the same third party with personal information and an abstract identification code. At the same time, the two data suppliers will each send the behavioral data linked to an abstract identification code to the party that ultimately will match the two sets of behavioral data, in this case the market researcher. By separating the identifying information from the behavioral data, a party, aside from the original data collectors, cannot know both the name of a person and behavioral information about that person. The third party will then correlate the two identification codes via matching personal information. It will then provide this correlation to the market researcher, who will use information to associate the two different behavioral data points, collected by the two data suppliers in order, to generate a market research report.

It was undisputed that in order to generate reports regarding the effectiveness of automobile advertisements, Kantar matches household automobile registration data and television viewing data. And, it was also uncontested that Kantar relies upon Experian to provide it with automobile registration [*15]  data. But the parties disputed whether Experian receives personally identifiable information along with the automobile registration data. To support its claims that Kantar's Auto Products utilize double blind matching, TRA cited an unsigned draft contract between J.D. Power and Experian, which stated that J.D. Power would not send personal identifying information to Experian, and that Experian would use anonymous, blind matching to pair data from set-top providers, like DirecTV, with auto registration data provided by J.D. Power, using a unique identifier, as opposed to relying upon the names of consumers. Summary Judgment Op., 984 F. Supp. 2d at 221. Kantar disputed TRA's interpretation of the contract, arguing that the contract did not address whether J.D. Power provides additional information to Experian and did not explain what "Unique Key ID" meant.

The district court agreed with Kantar. It first noted that TRA's Experian contract was unsigned, and, thus, was of limited value. Id. at 245. But, more importantly, the district court concluded that the unsigned contract did not contradict the declaration from Kantar Media's president, Shabbab, who stated that "Experian . . . appends J.D. Power[s] purchase attribute data to the DirecTV [identifier], [*16]  removes all [personally identifying information], and sends that data to [Kantar Media]. [Kantar Media] uses the same DirecTV [identifier] to correlate [set-top box] data with J.D. Power[s] purchase attribute data." Id. In the absence of any evidence contradicting Shabbabb's testimony, the district court granted Kantar's motion for summary judgment of non-infringement with respect to Kantar's Auto Products.

2. Trade Secret Claims

TRA also alleged that Kantar misused its confidential information--which Kantar received during the merger discussions and via its appointed board member--in order to assess whether to launch its competing product and to accelerate its own product development. But for this improper use, TRA contended that Kantar would have been unable to release a competing product as quickly as it did. Originally, TRA asserted that Kantar misappropriated twenty four categories of trade secrets but, in order to streamline this case for trial, the district court ordered TRA to reduce the number of asserted trade secrets in April 2013. Following this order, TRA agreed to reduce the number of trade secrets to the following five: "(1) Media TRAnalytics'--TRA's product--speed, reliability, scalability [*17]  and performance; (2) TRA's client lists and client interactions[;] (3) TRA's strategic plans[;] (4) TRA's product positioning[;] and, (5) TRA's capital structure, financials, financing proposals target investor list, and offers to acquire or merge the company." Summary Judgment Op., 984 F. Supp. 2d at 222.

After this reduction, Kantar filed a motion for summary judgment arguing, in part, that TRA could not show any misappropriation of a trade secret because it had failed to sufficiently identify any trade secret during discovery. The district court agreed, ruling that TRA's attempt to identify a small number of documents to support its five categories of trade secrets for the first time on the eve of summary judgment briefing was in violation of Fed. R. Civ. P. 26(e), which requires a party to supplement or correct its disclosure or response in a timely manner. Because allowing TRA to remedy its deficiency after the close of discovery would be prejudicial to Kantar and taxing on the district court, the district court dismissed TRA's trade secret claim as a sanction under Fed. R. Civ. P. 37. Id. at 239.

The district court, moreover, concluded that, even if it did consider the merits of TRA's misappropriation claims, TRA had failed to submit sufficient evidence that would demonstrate [*18]  that Kantar used TRA's alleged trade secrets or that TRA's secrets were protectable. Summary Judgment Op., 984 F. Supp. 2d at 239. Specifically, with respect to the Media TRAnalytics trade secret, the district court determined that TRA had disclosed most of the properties of the system to the public, such as its use of application program interfaces ("APIs") to handle legacy clients, and that TRA failed to provide evidence that Kantar's products used any of TRA's technical information. Id. Similarly, the district court found that TRA's client lists were also disclosed to various companies, that there was no evidence to support the claim that TRA took the necessary steps to protect this information, and that there was no proof that Kantar used these client lists. Id. at 240. Regarding TRA's strategic plans, the district court concluded that these plans were merely goals of the company, which are not protectable trade secrets under New York law, and, thus, failed as a matter of law. Id. Likewise, the district court determined that TRA's product placement strategies did not qualify as a trade secret, because they were mere marketing plans, which are unprotectable as a matter of law. Id. Lastly, the district court concluded that TRA's [*19]  financial information was not secret because it was publicly disclosed by TRA and that, even if the information was undisclosed, there was no evidence that Kantar used this information. Therefore, assuming arguendo that TRA was allowed to supplement its responses to add details regarding its trade secrets, the district court concluded that summary judgment would have been appropriate nevertheless. Id.

3. Non-Patent Damages

In addition to arguing that TRA could not satisfy its burden of proving that Kantar misappropriated TRA's trade secrets and Kantar infringed TRA's patents, Kantar also argued that TRA could not establish that it was entitled to damages for Kantar's alleged breach of fiduciary duty, breach of contract, and misappropriation of trade secrets. Summary Judgment Op., 984 F. Supp. 2d at 233-34. TRA's main damages theory turned on Kantar's decisions to release a competing product and to file a declaratory judgment against TRA which, according to TRA, froze the market. Specifically, TRA alleged that Kantar's actions spooked investors, because investors are hesitant to fund companies embroiled in litigation and are reluctant to invest in less-established companies--like TRA--when they are competing against a powerful and well-known [*20]  company such as Kantar. Id. at 235. TRA estimated Kantar's actions cost it $21-23 million. TRA calculated this figure by considering the value of TRA before Kantar's allegedly improper acts--$54 million--in May 2010 and after--$20 million--in July 2012 and reducing this figure by thirty percent because only seventy percent of this loss was attributable to Kantar. Id. at 234. To support its damages theory, TRA submitted: (1) an expert report by its damages expert, Melissa Bennis; (2) a portion of a declaration of its CEO, Mark Lieberman; (3) a portion of a declaration of Naveen Chopra, TiVo's CEO; and, (4) a portion of a declaration of Stephen B. Morris, another TRA expert that had been proffered, in support of its motion for an injunction. Id. at 234.

As for TRA's claim that Kantar's release of a competing product caused TRA harm, the district court determined that the opinion of TRA's damages expert, Ms. Melissa Bennis, did not satisfy the requirements of Federal Rule of Evidence 702 ("FRE") for several reasons. Bennis asserted that TRA's inability to acquire more financing in a fourth round of investments was caused, in large part, by Kantar's decision to release a competing product. The district court found that this opinion, based solely [*21]  on a temporal relationship, was insufficient to satisfy the rigors of FRE 702(c). Summary Judgment Op., 984 F. Supp. 2d at 241. The district court also found that Bennis's opinion was flawed because it failed to consider the alternative explanations given for TRA's failure--that investors were concerned about TRA's lack of revenue given the amount of capital already raised. Id. Lastly, the district court determined that Bennis's report relied upon mere speculation by Chopra that Kantar's product caused the market to freeze, making it difficult for TRA to compete. Id. For these reasons, the district court excluded Bennis's report.

Without Bennis's report, the only other evidence proffered by TRA to support its "frozen market" theory was Liberman's testimony. But the district court found that his testimony was based on conjecture and could not alone be relied upon to justify TRA's damages. Id. at 242. Thus, the district court excluded the entire theory from the case. Id. Without this theory, the court found no other basis for a request for compensatory damages and, thus, barred TRA from seeking any.

As for TRA's possible claim for nominal damages for TRA's remaining non-patent claims--breach of fiduciary and contract claims--its request for attorneys' [*22]  fees for its breach of fiduciary duty claim and its request for equitable relief for both claims in the form of an injunction, the district court concluded that Kantar's motion for summary judgment did not dispose of these claims, and the court delayed consideration of these claims until a later date. Summary Judgment Op., 984 F. Supp. 2d at 242, 246. The district court subsequently ordered the parties to meet and confer to decide what, if any, remedies were still available to TRA and, if the parties disagreed, to file an appropriate motion. TNS Media Research, LLC v. TRA Global, Inc., No 1:11-cv-4039, slip op. at 2 (S.D.N.Y. April 7, 2014), ECF No. 185 ("Damages Order"). Unable to reach an agreement, Kantar filed a motion to limit TRA's remedies and strike its jury demand. Upon consideration, the district court granted in part the motion, concluding that TRA was not entitled to any remedy, aside from nominal damages. Id. at 13.

With respect to TRA's request for injunctive relief for its breach of fiduciary duty claim, the district court concluded that such relief was now moot. TRA argued that Kantar breached its fiduciary duty to TRA by manipulating its representative on TRA's board to gain improper access to TRA's confidential information. In light of this access, TRA sought [*23]  to enjoin Kantar from continuing to breach its fiduciary duties to TRA. Since 2012, Kantar no longer had a representative on the board, however. Id. at 3. With no fiduciary to enjoin, the district court concluded that TRA had no basis for injunctive relief.

As for TRA's breach of contract claim, the district court determined that two out of the four asserted contracts--a 2008 non-disclosure agreement ("NDA") and a 2009 NDA--had expired, mooting a request for injunctive relief as to those contracts. Id. at 4. And, although the other two contracts--a 2007 NDA and an End-User License Agreement ("EULA")--had yet to expire, the district court concluded that its prior summary judgment ruling foreclosed any relief. Under the 2007 NDA, TRA granted Kantar access to its confidential information. TRA alleged Kantar used this information to copy TRA's patented method of creating target indices, but the district court had already determined that Kantar's accused method of creating market research reports was different than TNA's and granted summary judgment of non-infringement on this basis. Damages Order, at 5. Therefore, the court concluded that TRA could not now claim damages for Kantar's alleged misuse of this [*24]  information. Similarly, the EULA between TRA and Kantar granted Kantar access to TRA's Media TRAnalytics product. TRA asserted that Kantar breached the contract by incorporating this information into its own products, but the district court had already determined that Kantar did not use any of TRA's technical information related to the TRAnalytics product. Without evidence that TRA improperly used this confidential information, the district court determined that TRA could not pursue damages for Kantar's alleged breach of this agreement.

TRA also sought a jury trial on attorneys' fees, based on the disputed contracts.² The district court denied this request. It explained that attorneys' fees are only awarded after entry of judgment unless the law requires those fees be proven at trial as an element of damages. Id. at 6. Because none of the contracts asserted by TRA included a provision for attorneys' fees, the district court concluded that TRA could not rely on those contracts to support its claim to a jury trial. Accordingly, the district court found that TRA was not entitled to a trial on attorneys' fees, but the court recognized that TRA could request those fees after a judgment had been entered. [*25]

2   The district court rejected this claim, but TRA does not dispute it on appeal.

Next, the district court considered whether TRA could recover nominal damages for its breach of contract and fiduciary duty claims. Id. at 7. TNS alleged that TRA waived any claim to such damages by failing to specifically plead them in its complaint. But, the district court found that this failure did not prevent TRA from pursuing such relief, because a general request for damages can include nominal damages. Id. Thus, the district court found that TRA could seek nominal damages in this case. Id. It, however, concluded that TRA could not try this issue to a jury, noting that TRA's request was below twenty dollars--the threshold amount for a right to a jury trial under the Seventh Amendment to the U.S. Constitution. Damages Order, at 7 n.23.

In addition to its other damages theories, TRA also argued that the district court's prior decision eliminating its frozen market theory did not foreclose its ability to pursue its compensatory damages theories, including "loss of exclusive use" and unjust enrichment. The district court disagreed. It first noted that it understood TRA's only compensatory damages theory to be its "frozen market" theory, because TRA stated [*26]  in its summary judgment briefing that it was not seeking "lost profits," "lost sales," or "price erosion." Id. at 8. Because the district court had already addressed this theory in its resolution of Kantar's motion for summary judgment, the district court found that TRA could no longer pursue any other theory for compensatory damages.

The district court concluded, moreover, that even considering the merits of these alternative theories, it would conclude that TRA's alternative theories failed as a matter of law. Id. TRA alleged that part of its loss in value was attributable to its loss in the exclusive use of its trade secrets. The district court, however, found it already had ruled that TRA failed to put forth sufficient evidence that Kantar caused any of TRA's loss in value, thereby barring a "loss of exclusive use" claim. Id. at 9. As for TRA's unjust enrichment claim, the district court found that TRA's damages expert's brief discussion of this theory was legally insufficient. Thus, the district court found that TRA had failed to meet its burden that it was entitled to damages under this theory. Id. at 10.

Lastly, the district court considered TRA's assertion that it was entitled to punitive damages on [*27]  its breach of contract and fiduciary duty claims because of Kantar's egregious behavior. The district court stated that, in order to be awarded punitive damages, one must "satisfy the [] 'very high threshold of moral culpability.'" Damages Order, at 10-11 (quoting In re Methyl Tertiary Butyl Ether, 725 F.3d 65, 128 (2d Cir. 2013)). Under New York law, a party must demonstrate that a defendant acted with actual malice or such reckless disregard that borders on criminality to be awarded punitive damages. Although there was evidence that Kantar hoped to starve TRA for cash, there was no evidence that Kantar actually acted on those intentions. It determined that there was also evidence that TRA authorized the disclosure of confidential information to Kantar, which suggested that Kantar had a good faith belief that it was authorized to share this information. Without any proof of egregious conduct, the district court concluded that TRA was not entitled to punitive damages. Id. at 13.

Following this decision, only TRA's claim for nominal damages arising from Kantar's alleged breach of contract and fiduciary duty remained. TNS Media Research, LLC v. TRA Global, Inc., No. 1:11-cv-4039, slip op. at 1 (S.D.N.Y. July 2, 2014), ECF No. 188. Ultimately, the parties agreed that it would be wasteful to conduct [*28]  a bench trial on these claims when so little was at stake. Accordingly, Kantar agreed to pay TRA $1 in nominal damages, thereby mooting the final issue left in the case. The parties acknowledged that this agreement did not foreclose the parties from pursuing claims reinstated after an appeal. Id. In light of the parties' agreement, the district court entered a final judgment.

TRA timely appealed to this court. We have jurisdiction pursuant to 28 U.S.C. § 1295(a)(1) (2012).

Discussion

A. Trade Secret Claims

Trade secret misappropriation is a matter of state law. See Atl. Res. Mktg. Sys., Inc. v. Troy, 659 F.3d 1345, 1356 (Fed. Cir. 2011). The parties agree that New York law applies to TRA's trade secret claims against Kantar. We review the grant of summary judgment under the law of the regional circuit. Charles Mach. Works, Inc. v. Vermeer Mfg. Co., 723 F.3d 1376, 1378 (Fed. Cir. 2013). The Second Circuit reviews the grant or denial of summary judgment de novo. Petrosino v. Bell Atl., 385 F.3d 210, 219 (2d Cir. 2004). Summary judgment is appropriate when, drawing all justifiable inferences in the non-movant's favor, there exists no genuine issue of material fact and the movant is entitled to judgment as a matter of law. Fed. R. Civ. P. 56(c); Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255 (1986). Additionally, this court reviews a district court's decision to exclude evidence and impose discovery sanctions under the law of the relevant regional circuit. See Meyer Intellectual Props. Ltd. v. Bodum, Inc., 690 F.3d 1354, 1358 (Fed. Cir. 2012). In the Second Circuit, such [*29]  rulings are reviewed for abuse of discretion. See Phoenix Assocs. III v. Stone, 60 F.3d 95, 100 (2d Cir. 1995); Thomas E. Hoar, Inc. v. Sara Lee Corp., 900 F.2d 522, 525 (2d Cir. 1990).

1. Discovery Sanctions

TRA argues that the district court erroneously dismissed its trade secret claims, contending that the district court abused its discretion in dismissing the claims as a sanction for TRA's discovery violations. First, TRA disputes the district court's conclusion that it improperly waited until summary judgment to narrow its claims. See Summary Judgment Op., 984 F. Supp. 2d at 239. According to TRA, it voluntarily agreed to reduce the number of asserted trade secrets and did so, in part, at the district court's express invitation. To be punished for such a reduction was improper especially considering that it had not violated any court order previously, which is a prerequisite to dismissal of all or part of a cause of action as a sanction under the Federal Rules of Civil Procedure. Additionally, TRA claims that the district court's conclusion that Kantar was severely prejudiced by TRA's discovery failures is unsupported as there was no evidence suggesting that Kantar misunderstood TRA's claims or that it required additional depositions or subpoenas for additional documents. Because this was TRA's only discovery shortcoming, it argues that there was no reason [*30]  for the district court to issue such as a harsh sanction.

Kantar disputes TRA's allegations, explaining that TRA failed throughout the course of the litigation to sufficiently identify its trade secrets. For example, when Kantar requested more information about TRA's trade secrets in June 2012, TRA conceded that its responses were inadequate. Although TRA amended these responses in October 2012, Kantar alleges that these amendments were still insufficient because they merely listed several hundred documents, without explaining how these documents demonstrated the existence of trade secrets. And, even though the district court suggested that TRA reduce the number of its trade secrets in April 2013, Kantar contends that this was not an invitation for TRA to supplement its deficient disclosures because the discovery deadline had passed. Because TRA failed to identify its trade secrets adequately, Kantar contends that the district court correctly concluded that TRA had violated Federal Rule of Civil Procedure 26(e) ("FRCP"). And, because FRCP 37(c) permits a district court to exclude information that should have been disclosed under FRCP 26, Kantar contends that dismissing TRA's claims was well within the district court's discretion. [*31]  Federal Rule of Civil Procedure 26(e)(1)(A) provides that a party who has responded to an interrogatory "must supplement or correct its disclosure or response . . . in a timely manner if the party learns that in some material respect the disclosure or response is incomplete or incorrect, and if the additional or corrective information has not otherwise been made known to the other parties during the discovery process or in writing . . . ." "If a party fails to provide information or identify a witness as required by Rule 26(a) or (e), the party is not allowed to use that information or witness to supply evidence on a motion, at a hearing, or at a trial, unless the failure was substantially justified or is harmless." Fed. R. Civ. P. 37(c) (emphasis added).

Additionally or instead of this sanction, a district court:

(A) may order payment of the reasonable expenses, including attorney's fees, caused by the failure;

(B) may inform the jury of the party's failure; and

(C) may impose other appropriate sanctions, including any of the orders listed in Rule 37(b)(2)(A)(i)-(vi).

Id. Therefore, a failure to follow Rule 26(e) will warrant preclusion of omitted information, "unless the failure was substantially justified or is harmless." Id. "Failure to timely amend a contention interrogatory can bar use of [*32]  a theory of liability, especially when such failure results in prejudice to the adverse party." N.J. Dep't of Envtl. Prot. v. Atl. Richfield Co., No. 1:00-cv-1898, 2014 U.S. Dist. LEXIS 15966, at *9 (S.D.N.Y. Feb. 6, 2014) (citing Unigene Labs. v. Apotex, Inc., No. 06-cv-5571, 2010 U.S. Dist. LEXIS 67444, at *6 (S.D.N.Y. July 7, 2010), aff'd, 655 F.3d 1352 (Fed. Cir. 2011) ("[W]here there is substantial prejudice to the Plaintiffs--namely, not being advised of the contours of [a] claim until long after the termination of discovery and the filing of dispositive motions--the Defendants' failure to amend their contentions results in [a] claim being deemed waived.")).

Here, the district court concluded that TRA violated Rule 26 when it narrowed its trade secret claims on May 10, 2013. We note that on April 23, 2013, the court ordered TRA to reduce its trade secret claims. In response to that order the parties stipulated, with the court's approval, that TRA would reduce its trade secret claims by May 10. While this does not necessarily establish that TRA complied with Rule 26, it is also not clear that TRA's actions--to reduce its trade secret claims on the court-approved timeline--rise to the level of violating the rules of procedure. For the reasons below, we conclude the district court abused its discretion when it dismissed all of TRA's trade secret claims [*33]  as a discovery sanction. On remand, the district court should both address whether TRA violated Rule 26 and, if there was a violation, craft a more appropriate sanction.

We find that the district court abused its discretion when it dismissed all of TRA's trade secrets claims as a discovery sanction. Generally, "[a] district court 'abuses' or 'exceeds' the discretion accorded to it when (1) its decision rests on an error of law (such as application of the wrong legal principle) or a clearly erroneous factual finding, or (2) its decision--though not necessarily the product of a legal error or a clearly erroneous factual finding--cannot be located within the range of permissible decisions.'" Zervos v. Verizon N.Y., Inc., 277 F.3d 635, 650 (2d Cir. 2002) (quoting Zervos v. Verizon NY, Inc., 252 F.3d 163, 169 (2d Cir. 2001)). In order to determine the appropriate sanction for a discovery violation, the Second Circuit considers several factors including "(1) the willfulness of acts underlying noncompliance; (2) the efficacy of lesser sanctions; (3) the duration of noncompliance; and (4) whether the noncompliant party was on notice that it faced possible sanctions,'" but no one factor is dispositive. Agiwal v. Mid Island Mortg. Corp., 555 F.3d 298, 302-03 (2d Cir. 2009); see also SEC v. Razmilovic, 738 F.3d 14, 25 (2d Cir. 2013) (noting that "these factors are not exclusive, and they need not each be resolved against the [sanctioned] party"). [*34]

Considering these factors, it is clear that a dismissal was an inappropriate sanction in these circumstances. First, there is no indication that TRA purposefully shirked its discovery obligations. Cf. Robertson v. Dowbenko, 443 F. App'x 659, 661 (2d Cir. 2011) (noting that willful non-compliance exists when a party has "repeatedly failed to respond to interrogatories and produce documents . . . in violation of the district court's orders"). Instead, the record suggests that TRA actually tried to meet its obligations, as evidenced by its decision to amend its initial disclosures in response to Kantar's complaint that such disclosures were deficient and the fact that it was not until Kantar lobbied the court in April 2013 to order TRA to identify and limit its trade secrets in anticipation of trial that TRA became aware that Kantar still believed its disclosures were inadequate. When the court ordered the parties to confer in the hopes of streamlining the case, moreover, TRA responded by reducing the number of trade secret claims asserted. It did so within days of the court's suggestion and well before Kantar filed its motion for summary judgment. We see no discovery violation which would warrant such a harsh sanction, especially one imposed [*35]  without warning.

There is also nothing in the record that evinces the district court considered the efficacy of lesser sanctions. World Wide Polymers, Inc. v. Shinkong Synthetic Fibers Corp., 694 F.3d 155, 159 (2d Cir. 2012) (finding that a dismissal of a party's damages claim was inappropriate, in part, because "there [was] no indication in the record that the district court considered any lesser sanctions"). Most importantly, there was no indication prior to the district court's summary judgment order that TRA's discovery shortcomings could result in a dismissal of its trade secret claims. Id. at 160 ("Parties must be given notice and an opportunity to respond before a cause of action, or potential remedy, is dismissed as a sanction for failure to comply with court orders."); cf. Sit-Up v. IAC/InterActive Corp., No. 05-CIV.-9292, 2008 U.S. Dist. LEXIS 12017, at *16-19 (S.D.N.Y. Feb. 20, 2008) (before dismissing trade secret claims, the court afforded plaintiff two opportunities to amend its disclosures and warned them that failure to provide more detailed responses would result in their dismissal). Although Kantar disputes this conclusion, arguing that the district court did warn TRA at the April 2013 conference that its trade secret claims may be dismissed, a review of the record belies that contention. The district court did no more than state [*36]  that Kantar's motion for summary judgment of no trade secret misappropriation would be granted if the secrets identified did not meet the legal definition of a trade secret--it did not say that TRA's claims might be dismissed as a discovery sanction. And, although the district court justified its decision because it said TRA's actions were prejudicial to Kantar, the district court failed to explain why this was the case. See Shcherbakovskiy v. Da Capo Al Fine, Ltd., 490 F.3d 130, 140 (2d Cir. 2007) ("With no findings or explanation from the district court, we cannot conclude that the sanction of dismissal of the complaint and granting of the counterclaims was appropriate."). Thus, while the facts perhaps suggest that some sanction was appropriate, the record before us does not support dismissal. Accordingly, we reverse the district court's decision to dismiss TRA's trade secret claims.