UNITED STATES DISTRICT COURT SOUTHERN DISTRICT OF NEW YORK
September 22, 2011
TNS MEDIA RESEARCH, LLC (D/B/A KANTAR MEDIA AUDIENCES), AND CAVENDISH SQUARE HOLDING B.V., PLAINTIFFS/COUNTERCLAIM DEFENDANTS,
TRA GLOBAL, INC. (D/B/A TRA, INC.), DEFENDANT/COUNTERCLAIM PLAINTIFF.
The opinion of the court was delivered by: Shira A. Scheindlin, U.S.D.J.
OPINION AND ORDER
Plaintiffs TNS Media Research, LLC d/b/a Kantar Media Services ("Kantar") and Cavendish Square Holdings B.V. ("Cavendish") bring this action against TRA Global, Inc. ("TRA") seeking a declaratory judgment that Kantar's RapidView for Retail product ("RVR") did not infringe on U.S. Patent No. 7,729,940 ("'940 Patent"), which TRA owns. Cavendish also brings a claim against TRA for breach of contract.*fn1 TRA asserts various affirmative defenses and brings counterclaims for patent infringement, breach of fiduciary duty, misappropriation of trade secrets, and breach of contract.
Currently before the court is TRA's motion for a preliminary injunction requiring Kantar to cease making RVR, and to cease selling it or offering it for sale, on the basis that it infringes Claim 71 of the '940 Patent. For the reasons set forth below, the motion is denied.
TRA is a media research company that began operations in 2007.
Cavendish invested in TRA in August of that year, and, in return, received TRA stock and the right to a seat on TRA's board of directors.*fn3 Near the end of 2008, Cavendish filled that seat with a senior vice president from WPP PLC, the parent company to both Cavendish and Kantar.*fn4
In June 2010, the Patent and Trademark Office issued the '940 Patent, with TRA designated as the sole assignee.*fn5 Shortly thereafter, TRA created Media TRAnalytics ("MTRA"), a product based on the '940 Patent. MTRA collects and matches four kinds of data: data from cable and satellite set-top boxes to determine what channel a household is watching ("clickstream data"); programming data that allows it to determine what show was on that channel; advertising data that allows it to determine what ads ran during that show; and, finally, data about what the household purchases. These data are then anonymously analyzed and used to generate reports to help advertisers determine the return on their investment in previous ads ("ROI reports") and how to invest their future advertising dollars ("planning reports"), as well as second-by-second TV ratings reports ("ratings reports").*fn6 TRA claims that Kantar approached it about licensing the '940 Patent in August 2010;*fn7 Kantar denies this and states instead that TRA made repeated efforts to license Patent '940 to it for a substantial fee.*fn8 The parties never agreed to a license.
Kantar released RVR in March 2011.*fn9 Kantar claims that RVR is based on its own technology and does not infringe the '940 Patent, while also pointing out differences between RVR and MTRA.*fn10 Kantar also argues that the '940 Patent is invalid as obvious.*fn11 For its part, TRA claims that Kantar obtained its confidential information from Cavendish's representative on the TRA board of directors, that Kantar used this information to develop RVR, and that RVR in fact infringes on Claim 71 of the '940 Patent.*fn12
The parties have submitted expert declarations regarding the technical merits of TRA's motion for a preliminary injunction. The experts disagree about whether RVR infringes on Claim 71 of the '940 Patent, and whether Claim 71 is valid.
A. Expert Declarations on Infringement
TRA's expert splits Claim 71 into five steps -- collecting data, matching it on a computer, storing the matched data electronically, applying a cleansing and editing algorithm, and calculating a "true target index metric" -- and states that RVR "literally infringes" each step.*fn13 Kantar's expert asserts that RVR does not infringe on Claim 71 for two reasons. First, he claims that RVR collects "market-level data" on programming and advertising, whereas Claim 71 contemplates collecting "household level data associated with multiple households" for programming and advertising.*fn14 Second, he asserts that RVR applies a "cleansing and editing algorithm" to its input data sets before matching them, whereas Claim 71 applies the algorithm to the data post-matching.*fn15
TRA's expert makes two points in rebuttal. First, he states that "market-level data" is synonymous with "household level data associated with multiple consumer households" for purposes of Claim 71.*fn16 Second, he asserts that RVR "appl[ies] at least one cleansing and editing algorithm to the matched and stored data" by editing its raw clickstream data prior to matching, and by converting the matched data sets into user-friendly form.*fn17
B. Expert Declarations on Validity
Kantar's expert argues that Claim 71 is "invalid because it covers subject matter that would have been obvious at the time of the invention."*fn18 He states that with two exceptions, Claim 71 was taught by prior art as early as 1985.*fn19
As to the first exception -- that the 1985 prior art did not teach the use of a set-top box to collect clickstream data -- he claims that "artisans would substitute the [now] commonplace set-top boxes" for older external meters. and that a 2002 patent intimated such a substitution.*fn20 He claims that the second exception -- that the 1985 prior art did not teach the use of a trusted third party to group and match the collected data sets "without processing any personally identifiable information associated with the consumer household" to which an account identifier has been assigned -- was taught by prior art from 2000, and that privacy protection was required by Federal law.*fn21 As a result, he concludes that the method of matching and grouping data required by Claim 71 "would have been obvious to try. . . ."*fn22
TRA's expert responds in three ways. First, he asserts that the 1985
prior art made use of a supplemental data collection device to obtain
purchase data, the need for which was eliminated by Claim
71.*fn23 Second, he asserts that prior art did not
teach matching of purchase and media-exposure data.*fn24
Third, he notes that Kantar previously offered a product
without Claim 71's privacy protections.*fn25
III. LEGAL STANDARDS
A. Patent Validity
Patents are presumptively valid by statute.*fn26 Nonetheless, they may be deemed invalid if "the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject matter pertains."*fn27
1. Claim Construction
The court must construe claim terms before determining whether a patent is obvious in light of prior art.*fn28 First, the court should look at the actual claim language. Absent "an express intent to impart a novel meaning to claim terms," the claim language is given its "ordinary and customary meaning. . . ."*fn29
During this process, the court may look to the claim and the specification, the prosecution history, and dictionaries, amongst other things. While a novel or unique definition in the specification is "usually . . . dispositive,"*fn30 "[t]he claim is the measure of [the patentee's] right to relief, and while the specification may be referred to to limit the claim, it can never be made available to expand it."*fn31
2. Comparing Claim to Prior Art
Once the court has construed the claim terms, it proceeds to determine whether a patent claim is obvious by looking at "the scope and content of the prior art, the differences between the prior art and the claims at issue, and the level of ordinary skill in the pertinent art."*fn32 Additionally, a court may find a patent obvious "if it would have been obvious to a person having ordinary skill to try a course of conduct. . . ."*fn33 At trial, the alleged infringer must make a showing of obviousness by clear and convincing evidence.*fn34
3. Secondary Factors of Non-Obviousness
Once the alleged infringer has made out a prima facie case that a patent is obvious, a court must consider secondary factors of non-obviousness before reaching its final decision.*fn35 The patentee bears the burden of demonstrating a nexus between any secondary factor of non-obviousness and the patent claim.*fn36 Therefore, if a secondary factor arises from something beyond the claim or something that is taught by prior art, it is insufficient to overcome a prima facie case of invalidity.*fn37
B. Patent Infringement
If the court finds that a patent is valid, the court should proceed to analyze infringement claims.*fn38 Analysis of patent infringement involves two steps. First, as in the validity analysis, the court must construe the patent claims. Second, the court must compare "the properly construed claims to the device accused of infringing."*fn39 If the patentee can then "show the presence of every element or its equivalent in the accused device," the court should find that the accused device infringes the patent.*fn40
C. Preliminary Injunction
Injunctive relief in patent cases is authorized by statute.*fn41 A preliminary injunction, however, is "an extraordinary remedy never awarded as of right."*fn42 A party seeking a preliminary injunction must therefore establish " that [it] is likely to succeed on the merits,  that [it] is likely to suffer irreparable harm in the absence of preliminary relief,  that the balance of equities tips in [its] favor, and  that an injunction is in the public interest."*fn43 A movant seeking a preliminary injunction that would alter the status quo must establish the same factors by a "clear showing" that it is entitled to the relief it requests.*fn44
In order to show a likelihood of success on the merits in a patent case, the patentee "must show that it will likely prove infringement, and that it will likely withstand challenges, if any, to the validity of the patent."*fn45 This means that "[a] preliminary injunction should not issue if an alleged infringer raises a substantial question regarding either infringement or validity."*fn46 Raising such a question requires "more than a scintilla but less than a preponderance of the evidence."*fn47
Once the accused infringer has made its required showing, "the burden shifts to the patentee to show that the defense lacks substantial merit."*fn48 A patentee who does not meet this burden "is not entitled to a preliminary injunction," regardless of the weight of the other factors, because it has not shown a likelihood of success on the merits.*fn49
Validity of Claim 71 of the '940 Patent
A. Claim Construction
As a threshold matter, I must construe the disputed language in Claim 71. The parties dispute the meaning of the phrases "household level data associated with multiple consumer households" and "[applying] a cleansing and editing algorithm to the matched and stored data."
1. "Household Level Data Associated with Multiple Consumer Households"
TRA's expert asserts that the market-level programming and advertising data used by RVR falls within the meaning of "household level data associated with multiple consumer households" because "each set of this data. . . discloses. . . what show each of multiple consumer households in a given area is watching, and what advertisement they saw in that area."*fn50 There are several problems with this assertion.
As a preliminary matter, TRA's expert misstates the way RVR works, as explained by Kantar's expert. According to Kantar's expert, the programming and advertising data sets that RVR uses do not, and alone cannot, reveal what any individual consumer household has seen by way of programming or advertisement.
Instead, RVR uses these data sets to determine what the raw clickstream data it obtains from set-top boxes actually means.*fn51
Furthermore, "household level data associated with multiple consumer households," when given its ordinary meaning, would not cover a data set simply because it could be used to determine something about multiple households on a household-by-household basis. Instead, when given its ordinary meaning, that phrase would require that there be some household level data to begin with, and that the same be later aggregated into a data set including multiple consumer households.
Claim 71 attaches this phrase to each of the four sets of data its
proposed system collects. The specification and description, however,
do not contemplate that each data set would be of the same kind or
collected in the same way. The claim contemplates that clickstream
data will be collected from each individual household via set-top
boxes, and provided to the proposed system by a third-party data
vendor;*fn52 likewise, the claim contemplates that
purchase data will be collected from each individual household by means of a unique store
discount card, also provided by a third-party data vendor.*fn53
Programming data, however, would come directly from cable
operators or a third-party data vendor without ever being collected on
the household-level, and advertising data would come either directly
from the television networks or the advertisers seeking to make use of
the proposed system.*fn54
It appears, then, that "household level data associated with multiple consumer households" is to have different meanings in different contexts. In the context of clickstream and purchase data, its meaning is the ordinary meaning explained above. In the context of programming and advertising data, however, the inventors of the '940 Patent appear to have intended a specialized meaning of "usable to determine what programs and advertising each of multiple consumer households saw." I therefore construe the phrase according to its ordinary meaning when applied to the former data sets, and according to its specialized meaning when applied to the latter.
2. Applying a "Cleansing and Editing Algorithm to the Matched and Stored Data"
Kantar and Cavendish assert that "[applying] a cleansing and editing algorithm to the matched and stored data" requires the algorithm to be applied after the relevant data sets are matched and stored in a computer system.*fn55 Relying on the specification accompanying Claim 71, TRA's expert asserts that the algorithm may be applied before any matching of the data sets occurs, and that the term "matched and stored" "is a descriptive reference to the data to which the algorithm is applied" that "does not impose a sequential limitation. . . ."*fn56 This interpretation does violence to the ordinary grammatical understanding of the past tense in a manner that would expand the coverage of the claim rather than limit it. Because the specification cannot, as a matter of law, expand the claim, I find that "[applying] a cleansing and editing algorithm to the matched and stored data" means "applying a cleansing and editing algorithm to the data after it is matched and stored."
B. Comparison of the Prior Art and the '940 Patent
Having construed the disputed portions of Claim 71, I must now determine what differences, if any, exist between Claim 71 and the prior art. At the preliminary injunction stage, this effectively means that I must determine whether Kantar and Cavendish have raised substantial questions of obviousness and whether TRA has shown that those questions lack substantial merit.
I must first examine Claim 71, so that I may compare it to the prior art. Claim 71 consists of five basic steps. First, the proposed system collects clickstream, advertising, programming, and purchase data without the use of supplemental data collections devices. Second, the proposed system matches the data on a household-by-household basis, without using personally identifiable information. Third, the proposed system stores the matched data on a computer. Fourth, the proposed system applies a "cleansing and editing algorithm" to the data stored on the computer after it is matched. Fifth, the proposed system generates a planning report based on the matched data.*fn57
After a careful review of the record currently before me, I conclude
that Kantar and Cavendish have shown the prior art to teach six
things. First, I find that prior art taught the use of a set-top box
--which is not a "supplemental data collection device" -- to collect
clickstream data.*fn58 Second, I find that one of
Kantar's own products taught the collection and use of programming and
advertising data from third parties without the use of a supplemental
data collection device, in a manner consistent with the specialized
meaning of "household level
data associated with multiple consumer households" noted
above.*fn59 Third, I find that the same Kantar product
taught to match programming and advertising data with clickstream data
to determine what programs and advertisements individual households
actually saw.*fn60 Fourth, I find that prior art
taught to compare information about what programs and advertisements
households actually saw with purchasing data, said purchasing data
being "household level data associated with multiple consumer
households" consistent with the ordinary meaning of that phrase noted
above;*fn61 particularly, it taught the use of
household-specific cards to track buying behavior at the point of
purchase.*fn62 Fifth, I find that prior art taught the
use of a trusted third-party "matcher" to protect the personal
information of households whose behavior is analyzed.*fn63
Sixth, I find that prior art taught the generation of planning reports.*fn64
Comparing these findings to Claim 71, I find that the prior art substantially taught all of the limitations in Claim 71. I further find that it would have been obvious to a person reasonably skilled in the art to attempt to combine these teachings, even though no single device embodied them all. I therefore conclude that, on the record currently before me, Kantar and Cavendish have met their burden of raising a substantial question about the validity of Claim 71 by "more than a scintilla" of evidence.
C. Whether the Question of Claim 71's Validity Lacks Substantial Merit
I must now consider whether TRA has shown that the substantial question of validity that Kantar and Cavendish raise lacks substantial merit. While the Federal Circuit has not clearly stated what kinds of evidence a court may consider in this regard, I must, at the very least, consider the secondary factors of non-obviousness that TRA raises.*fn65 As noted above, however, if TRA cannot demonstrate a "nexus" between these factors and its patent, it is legally inappropriate for me to consider them.
Amongst the secondary factors TRA raises, only the allegation of copying is supported with the required nexus. TRA argues that the similarities between RVR and MTRA are most likely explained by copying, and that Claim 71 is therefore not obvious. TRA presents the following four facts as evidence of copying: first, that Kantar and Cavendish were investors in TRA; second, that Cavendish's duly appointed member of TRA's board of directors had access to inside information; third, that RVR and MTRA provide similar reports to their users; and fourth, that RVR and MTRA use many of the same data vendors.*fn66
Kantar and Cavendish dispute each of these arguments. First, Cavendish's representative on TRA's board denies sharing any confidential information with Kantar or Cavendish, or knowing of RVR before its release.*fn67
Second, Kantar and Cavendish flatly deny that RVR produces ROI reports.*fn68
Third, Kantar and Cavendish point out that Kantar used the data vendors common to RVE and MTRA since before TRA entered the market.*fn69 Because I find these arguments approximately equal in persuasive value to TRA's arguments, I find that TRA has not shown by even a preponderance of the evidence that the obviousness defense raised by Kantar and Cavendish lacks substantial merit.*fn70
Kantar and Cavendish have raised a substantial question about the validity of Claim 71 of the '940 Patent. Because TRA has not shown that said question lacks substantial merit, TRA's motion for a preliminary injunction is denied. The Clerk of Court is directed to close this motion (docket #11). A conference is scheduled for October 20 at 3:30 PM.