TNS Media Research LLC V. Tivo Research & Analytics, Inc.
Docket: 11-cv-4039 (KBF)
Court: District Court, S.D. New York; November 28, 2016; Federal District Court
On February 22, 2016, Judge Shira A. Scheindlin dismissed all patent claims by TiVo Research and Analytics, Inc. (TRA) in a ruling that attempted to interpret patent law following the Supreme Court's decision in Alice Corp. Pty. Ltd. v. CLS Bank International. Since then, the Federal Circuit has provided further clarity on the application of Alice, prompting the reassignment of the case to Judge Katherine B. Forrest on May 10, 2016, after Judge Scheindlin's retirement. The Court invited additional briefings to reassess the February 22 decision in light of the clarified legal framework. Upon reviewing Federal Circuit case law and the relevant patent claims, the Court found that vacatur was warranted. Consequently, the February 22, 2016 Opinion and Order has been vacated, and TRA's patent claims are reinstated.
This case began on June 15, 2011, when TNS Media Research, LLC and Cavendish Square Holding B.V. sought a declaratory judgment of non-infringement regarding U.S. Patent No. 7,729,940. TRA counterclaimed for infringement, and over the following year, acquired two additional patents, U.S. Patent Nos. 8,000,993 and 8,112,301. An amended answer filed by TRA in June 2012 included claims for these patents. The case faced various legal challenges, including a 2013 summary judgment in favor of the plaintiffs, which was partially vacated by the Federal Circuit in 2015. In the wake of the Supreme Court's Alice decision, the plaintiffs moved for summary judgment on patent ineligibility, leading to the February 22, 2016 ruling, which is now under reconsideration.
The February 22, 2016 Opinion focuses on claim 71 of the '940 Patent, which addresses an innovative method for measuring advertising effectiveness in a fragmented digital landscape. The invention, detailed in the patent issued in 2010, involves gathering, storing, and analyzing data related to media consumption (e.g., television programming) alongside individual product-purchasing data and demographic information. This comprehensive data analysis aims to determine correlations between advertising exposure and subsequent purchases, thus allowing advertisers to assess the return on investment for their campaigns.
To mitigate privacy concerns associated with tracking viewer behavior and purchasing habits, the invention incorporates a 'double-blind' matching process, although this process itself is not the primary claim of the patent. The Court's analysis includes a review of both the claim language and the specification, referencing relevant case law to contextualize the claims within the broader framework of patent law, particularly the Alice framework for determining patent eligibility.
The '940 Patent, titled "Analyzing Return on Investment of Advertising Campaigns by Matching Multiple Data Sources," outlines a system and methods for processing and analyzing advertising data within a software-as-service framework or through a downloadable program. It focuses on integrating various data types—household media exposure, product purchases, advertising, program information, and demographics—sourced from devices such as television set-top boxes. The invention aims to enhance media placement effectiveness and tailor advertising content to specific households.
Key components include methods for data storage distributed across multiple shards and measures to ensure data privacy. The invention addresses the critical commercial need for accountability in advertising expenditures, specifically the challenge of accurately assessing the return on investment (ROI) of different media channels. It highlights the complications arising from the fragmentation of television programming, which questions the reliability of current media accountability methods.
To effectively measure advertising impact, the invention requires a large sample size, potentially necessitating data from hundreds of thousands of households to achieve stability in audience ratings. Privacy considerations are integral to the data collection and analysis process, and the system allows for measuring sales effects of advertising using household-level data rather than broader market averages. It also enables implementation without additional data collection devices, simplifying the measurement process.
The invention provides a method for advertising measurement that enables cross-correlation of audience data obtained from television distribution systems, such as digital set-top boxes, with other media touchpoints (e.g., Internet, radio, print) and product purchase data from multiple households. This method allows for an analysis of consumer purchases in relation to prior advertising stimuli. Data collection includes a 'clickstream' that captures time-stamped media exposure events (like channel changes) from devices controlled by consumers. The invention also measures advertising effectiveness through Internet and DVR devices, facilitating real-time access to calculations and metrics that can optimize advertising resource allocation by shifting funds from less effective advertising to more effective campaigns, thereby enhancing return on investment (ROI). The inventors note that raw viewing data from digital set-top boxes can be insufficient, prompting the use of cleansing algorithms to improve data accuracy. Additionally, the invention addresses the challenges of measuring television exposure among large household samples and correlating it with purchasing behavior, while ensuring cost-effectiveness and precision in data collection across various media and marketing channels.
The specification outlines methods for conducting multivariate regression analyses and marketing-mix modeling for advertisers, enabling detailed evaluations of ad exposure and purchase behaviors at the household level. It emphasizes the importance of aggregating data to facilitate granular analysis of various characteristics pertinent to advertising effectiveness. The document includes approximately forty figures illustrating key processes, such as data collection, storage, and report generation. It introduces several algorithms for metric measurement, allowing for audience engagement analysis based on specific advertisements, calculated at the household level using time-based averages. Furthermore, it details the computation of cost per minute (CPM) metrics, incorporating variables like household counts, program view instances, and comparisons between expected and actual viewing impressions. Additional CPM algorithms are specified across columns 29 to 37.
Claims 1 and 71 of the '940 Patent are similar, differing mainly in their focus on the final calculation: Claim 1 centers on calculating a return on investment (ROI), while Claim 71 pertains to determining a 'one true target index metric.' Claim 71 outlines a computer-implemented method for analyzing consumer behavior related to advertising exposure or program delivery. The method involves:
1. Collecting clickstream data from a consumer's program delivery source, which does not require a supplemental data collection device and must include household-level data from multiple households.
2. Collecting advertising data related to the program delivery, also independent of a supplemental device, with household-level data.
3. Collecting program data associated with the delivered program, again without needing a supplemental device and including household-level data.
4. Collecting purchase data from a purchase data source, independent of a supplemental device, including household-level data.
The method requires matching the collected data (advertising, clickstream, purchase, and program data) at a household level using a centrally located electronic computer processor. This includes grouping data by account identifiers while ensuring no personally identifiable information is processed, and matching account identifiers for the same household. The matched data is stored in a central data storage medium and subjected to cleansing and editing algorithms. Finally, the method calculates at least one true target index metric based on the stored data.
Claim 71 outlines a process consisting of four key steps:
1. **Data Collection**: This involves gathering various types of data, including clickstream, advertising, program, and purchase data.
2. **Matching Step**: This requires matching portions of the collected data at the household level using a centralized electronic processor. The matching process involves grouping data with a household account identifier while ensuring no personally identifiable information is processed.
3. **Cleansing and Editing Step**: This utilizes an algorithm for data refinement.
4. **Calculation Step**: This involves calculating a single target index metric.
Each step has specific limitations, particularly regarding the nature of data collection, which must include household-level data from multiple consumer households without dependence on external data collection devices. The process necessitates computer use for matching, cleansing, and calculation, emphasizing that central storage and processing capability are essential.
In the February 22, 2016 Opinion, the Court classified claim 71 as an abstract idea under the Alice framework, determining it involved the digital matching of purchase and program delivery data to households. The Court noted that the data could be sourced from the Internet, and no tangible machine was necessary for data collection, framing the claim as an abstract concept. The Court also found that the claim did not demonstrate inventiveness, merely representing conventional data collection and processing methods without improving existing technologies or fields.
Federal Rule of Civil Procedure 60(b) allows courts to relieve parties from final judgments, orders, or proceedings due to 'mistake or any other reason that justifies relief,' emphasizing the need to balance justice and the finality of judgments. Such relief is considered 'extraordinary' and appropriate only in exceptional circumstances, yet must be broadly construed to achieve substantial justice.
Section 101 of the Patent Act defines the scope of patentable subject matter, allowing patents for new and useful processes, machines, manufactures, or compositions, subject to specific conditions. The Supreme Court has established a significant implicit exception to this rule: laws of nature, natural phenomena, and abstract ideas are not patentable to prevent the monopolization of essential scientific tools, which could hinder rather than promote innovation. Analyzing information through mental processes or mathematical algorithms is categorized as an abstract idea. However, the application of these exceptions should not extend beyond their intended purpose of preventing preemption of new discoveries, as doing so could undermine patent law. The inquiry into patent eligibility under § 101 is a legal question distinct from assessments of novelty or obviousness under §§ 102 and 103.
At the first step of the Alice framework, courts assess whether a claimed invention is 'directed to' ineligible subject matter, specifically 'abstract ideas.' This requires a comprehensive evaluation of the claims to determine their overall character. The Supreme Court has not defined a precise standard for identifying an 'abstract idea,' but courts may compare claims to those previously deemed abstract. Since the Alice decision, the Federal Circuit has frequently invalidated patents for involving fundamental economic practices, conventional business methods, or abstract mathematical formulas applied generically on computers without any technological improvement. Merely stating an abstract idea with the phrase 'apply it with a computer' does not suffice; a more substantive application is necessary. Claims that only require generic computer implementation without enhancing technology are typically found ineligible. However, the Federal Circuit clarified that not all improvements in computer technology are inherently abstract, emphasizing that the focus should be on whether the claims assert specific advancements in computer capabilities rather than merely invoking computers as tools for abstract processes. The inquiry must distinguish between patent-ineligible concepts and routine claims that incorporate natural laws or phenomena, recognizing that all inventions may, in some respect, rely on these concepts.
At step one of the patent eligibility analysis, patent claims must be evaluated in conjunction with the specification to determine if their overall character pertains to excluded subject matter. This assessment, informed by the Enfish decision and subsequent rulings following the Alice framework, focuses on whether the claim advances beyond a well-established concept or merely seeks to monopolize it by applying technology. The court's task involves defining the underlying idea and determining if that idea, across its various generic forms, constitutes the invention itself or if the invention pertains to executing the abstract idea in a specific manner.
A claim directed to an abstract concept is not automatically patent-ineligible under § 101. Courts must examine the claim's elements to determine if they contain an "inventive concept" that transforms the abstract idea into a patent-eligible application. This evaluation looks at the remaining elements, either in isolation or in combination, to find an inventive concept that offers significantly more than the abstract idea itself. The inquiry is limited to eligibility under § 101 and does not address issues of novelty or non-obviousness under §§ 102 and 103. The novelty of an element or process is irrelevant to determining if a claim fits within the eligible categories of patentable subject matter.
Eligibility analysis assesses whether steps combined with an abstract idea are insignificant, conventional, or routine, which would render the claim effectively covering the idea itself. Merely reciting routine steps or adding a computer to an abstract idea does not satisfy the inventive concept requirement. However, claims that improve computer functionality or involve a non-conventional arrangement of known elements may be patent-eligible. The focus remains on whether the claim presents an inventive concept sufficient to differentiate it from an abstract idea.
A list of non-exhaustive questions has been established by the Court for the step-two analysis of patent eligibility, focusing on aspects such as improvements, benefits, novelty, specific applications, and the procedural steps in applying the invention. These inquiries are essential to determine if the claimed invention represents an actual advancement in human knowledge rather than merely attempting to monopolize ineligible subject matter. TRA seeks vacatur, arguing that the February 22, 2016 Opinion oversimplified the claims and overlooked the invention's benefits, particularly emphasizing that 'double-blind matching' is merely one component of the claimed method. The Court agrees, noting that the invention in the '940 Patent, especially claim 71, addresses the challenge of evaluating advertising campaign effectiveness in a fragmented digital landscape. The patent highlights the difficulty of data acquisition due to diverse media consumption and identifies the potential value of capturing such data.
Claim 71 presents a concrete method to analyze detailed viewing data alongside purchasing behavior, countering any high-level abstract characterization. The Court asserts that the claim is not merely an abstract idea but rather a specific, actionable method for gathering and analyzing data. Even if the claim were abstract, its inventive nature allows it to satisfy the second step of the Alice analysis, demonstrating that it embodies more than mere abstract concepts through its detailed, multi-sourced data collection approach. The Court emphasizes the significance of claim limitations as understood within the context of the patent specification.
The specification outlines key issues addressed by the patent, notably the inability of advertisers to assess utility in real time and with adequate sample sizes. Claim 71 proposes a computer-implemented method that systematically collects, stores, cleanses, and analyzes data to improve advertising effectiveness. This method is innovative, as it captures specific viewer behaviors, such as a temporary volume reduction alongside competitor purchases, to provide actionable insights for advertisers. The method also facilitates large demographic studies to analyze correlations between education, income levels, and purchasing behaviors. Claim 71 is deemed patent-eligible under the Alice framework, and the court finds no failure at its second step. Consequently, the prior ruling from February 22, 2016, is vacated, and both parties are ordered to propose a plan for final resolution within two weeks. TRA currently asserts infringement of specific claims from three patents, with the previously broader claims now narrowed. The court acknowledges the interchangeable use of "viewer" and "household," noting that while viewing and purchasing may not occur simultaneously, data storage allows for retrospective analysis.
