PTAB

IPR2021-00034

Facebook Inc v. USC IP Partnership LP

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: System and Method for Intent Data Processing
  • Brief Description: The ’300 patent describes systems and methods for determining a website visitor's intent to predict and suggest other webpages. The system can prompt a visitor to indicate an intent, process that intent, and display links to recommended webpages based on ranking data provided by previous visitors.

3. Grounds for Unpatentability

Ground 1: Claims 5-6, 8, and 10 are obvious over Gregov, Dai, Hartman, Linden, and Holzner.

  • Prior Art Relied Upon: Gregov (Patent 8,032,506), Dai (Application # 2008/0235204), Hartman (Patent 6,615,226), Linden (Patent 6,912,505), and Holzner (Ajax for Dummies (2006)).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that the primary references, Gregov and Dai, teach all key limitations of the independent claims. Gregov, assigned to Amazon.com, discloses a system for generating "user-directed product recommendations." A user is prompted to indicate an intent by selecting a "More like this" button next to a displayed item, which serves as a "seed item" to generate a new ranked list of recommended items. Petitioner asserted this maps to prompting for, receiving, and processing an "indicated intent" to determine and display recommended webpages. Dai discloses a method for improving search results by using aggregated user feedback (e.g., 1-5 star ratings) to rank documents, including webpages. Petitioner mapped this to the limitation of using "intent ranking data" provided by previous visitors to rank the recommended webpages. Hartman and Linden were cited for teaching basic, well-known web technologies like using hyperlinks for recommended items that lead to item detail pages. Holzner was cited for teaching Ajax technology, which allows a webpage's content to be updated without a full page reload, thereby rendering a subsequent page of recommendations on the "same webpage" as the initial prompt.
    • Motivation to Combine: A POSITA would combine Gregov's recommendation engine with Dai's feedback-based ranking system to improve the relevance and quality of the recommendations. Using aggregated historical user ratings (Dai) to further rank the items identified by Gregov’s "seed" mechanism would predictably lead to better results. The motivation to incorporate Hartman and Linden's hyperlinking is to improve user convenience and increase sales by allowing easy navigation to product details. A POSITA would be motivated to use Holzner's Ajax technology to provide an improved, faster, and more "desktop-like" user experience by avoiding disruptive full-page reloads.
    • Expectation of Success: Petitioner contended that a POSITA would have had a high expectation of success. The references are all within the analogous art of web-based information retrieval and presentation. Combining them involved the application of known techniques (feedback ranking, hyperlinking, Ajax) to a known system (a recommendation engine) to achieve the predictable result of an improved user experience and more relevant recommendations.

Ground 2: Claims 7 and 9 are obvious over Gregov, Dai, Hartman, Linden, Holzner, and Nickerson.

  • Prior Art Relied Upon: Gregov (Patent 8,032,506), Dai (Application # 2008/0235204), Hartman (Patent 6,615,226), Linden (Patent 6,912,505), Holzner (Ajax for Dummies (2006)), and Nickerson (Patent 7,827,487).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground adds Nickerson to the combination from Ground 1 to teach the specific limitations of dependent claims 7 and 9, which recite prompting for a rank and storing a "datapoint" comprising the webpage's identity, the intent, and the ranking. Nickerson discloses a "feedback measurement tool" that can be incorporated into a webpage to prompt a user to provide a rating on a multi-level scale (e.g., "++" to "--"). Nickerson explicitly teaches storing this feedback in a database, including the webpage's identity (URL) and the subjective rating provided by the user. Petitioner argued this directly teaches the limitations of prompting for and storing a rank and webpage identity. The "indicated intent" (the "seed item" from Gregov) would obviously be stored with the rank to provide context, a practice taught by Dai, which associates feedback data with the query that produced the results. Claim 7's requirement of processing a "plurality of datapoints" is met by using the collected feedback data from multiple pages to generate the ranked list of recommendations.
    • Motivation to Combine: A POSITA would be motivated to incorporate Nickerson's feedback tool into the Gregov/Dai system to implement a standardized and user-friendly interface for collecting the user ratings that Dai's system uses. This would improve the quality of data for the ranking engine and provide valuable feedback to the website owner about page performance, leading to an improved user experience.
    • Expectation of Success: Nickerson is designed for easy incorporation into existing websites. A POSITA would have readily understood how to add Nickerson's feedback tool to the webpages in Gregov's system to collect user rankings and store them in a database, with a high expectation that the combined system would function as intended.

4. Arguments Regarding Discretionary Denial

  • Petitioner argued against discretionary denial under §314(a), asserting that the parallel district court litigation should not preclude institution. It was noted that the district court's trial date of December 13, 2021 was tentative and could be delayed due to the court's docket and the effects of COVID-19. Petitioner also argued it acted with diligence, filing the IPR well before the one-year statutory bar and shortly after receiving the patent owner's infringement contentions.

5. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claims 5-10 of the ’300 patent as unpatentable under 35 U.S.C. §103.