Google LLC v. Metarail Inc

1. Case Identification

• Case #: IPR2024-01270
• Patent #: 9,633,378
• Filed: October 2, 2024
• Petitioner(s): Google LLC
• Patent Owner(s): Metarail, Inc.
• Challenged Claims: 1-26

2. Patent Overview

• Title: Generating Deep-Linked Ads
• Brief Description: The ’378 patent describes a system for automatically generating "deep-linked" advertisements. The system purports to solve the problem of manually linking form fields on a publisher's website to corresponding fields on an advertiser's website by creating and utilizing a "universal variable map" populated by crawling internet websites to identify and categorize form data.

3. Grounds for Unpatentability

Ground 1: Claims 1-6, 9-16, 19-24, and 26 are obvious over Belanger in view of Halevy.

• Prior Art Relied Upon: Belanger (Application # 2005/0144048) and Halevy (Application # 2006/0230033).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Belanger disclosed the foundational system of a reservation booking website (e.g., for car rentals) that uses "promotional deep-links" to connect users to partner sites (e.g., amusement parks). Belanger taught that these deep-links could pre-populate forms on the partner site with user data (e.g., travel dates, location) from the original reservation. However, Petitioner asserted that Belanger's system relied on a manual, time-intensive process to create these deep-link mappings. Halevy was argued to disclose the missing automation piece. Halevy taught a system that automatically creates a "form database" by crawling web forms, analyzing and classifying them (e.g., "car," "hotel"), and mapping their fields to a shared set of normalized identifiers ("SO-properties"). Petitioner contended this "form database" is analogous to the ’378 patent’s "universal variable map" and could be queried to automatically generate the exact type of pre-populated deep-links described in Belanger.
  • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine Belanger's manual deep-linking system with Halevy's automation technology to overcome the recognized bottleneck of manual coding. Automating the creation of deep-links would improve the scalability and efficiency of offering cross-promotional advertisements, a known valuable feature in the travel industry.
  • Expectation of Success: A POSITA would have a reasonable expectation of success because combining the references involved applying Halevy's known automation solution to Belanger's known business problem. The integration would use standard web technologies like SOAP and XML, which were discussed in the references, to predictably achieve an automated system for generating deep-linked ads.

Ground 2: Claims 7-8, 17-18, and 25 are obvious over Belanger and Halevy in view of Haveliwala.

• Prior Art Relied Upon: Belanger (Application # 2005/0144048), Halevy (Application # 2006/0230033), and Haveliwala (Patent 7,756,887).

• Core Argument for this Ground:

  • Prior Art Mapping: This ground built upon the combination of Belanger and Halevy to address claims requiring ad generation or refinement based on a user's "implicit action." Petitioner argued that Haveliwala disclosed a search engine that monitored user browsing behavior—specifically, pointer "hover" time over informational items—to determine the items' relevance. This relevance data was then used to refine or reorder subsequent search results. Petitioner asserted that this monitoring of a user's hover activity constituted the claimed "information provided by a user's implicit action" used to "generate or refine" the deep-linked ads from the primary combination.
  • Motivation to Combine: A POSITA, having already created an automated deep-linking system by combining Belanger and Halevy, would be motivated to further improve the relevance of the displayed ads to increase user engagement and conversion rates. Haveliwala taught a known method for achieving this by incorporating user feedback (hovering) to better rank results. Therefore, a POSITA would integrate Haveliwala’s user feedback mechanism to refine the selection and ranking of deep-linked ads generated by the Halevy system.
  • Expectation of Success: Success would be expected, as implementing user activity monitoring with well-known scripting languages like JavaScript (as taught by Haveliwala) was within the skill of a POSITA. Applying the resulting relevance scores to rank a list of deep-links was a straightforward and predictable enhancement to the base system.

• Additional Grounds: Petitioner asserted additional obviousness challenges (Grounds 2 and 4) that were substantively similar to the grounds above but added Applicant Admitted Prior Art (APA) from the ’378 patent specification to further evidence the general knowledge in the art regarding deep-linked advertising on travel websites.

4. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under §314(a) based on the Fintiv factors would be inappropriate. It was asserted that the parallel district court litigation was at a very early stage with no trial date set, no claim construction conducted, and no substantive rulings issued. Petitioner also noted that the IPR challenges all 26 claims of the ’378 patent, whereas the district court complaint identified only one asserted claim, weighing against denial due to limited issue overlap.
• Petitioner further argued that denial under §325(d) was unwarranted because the primary prior art references (Halevy and Haveliwala) presented new art and arguments that were not before the Examiner during the original prosecution. It was contended that the new art was not cumulative to the art previously considered, thus presenting substantial new questions of patentability.

5. Relief Requested

• Petitioner requested the institution of an inter partes review and the cancellation of claims 1-26 of the ’378 patent as unpatentable.