Home Depot USA Inc v. Security Technology LLC

1. Case Identification

• Case #: IPR2024-01420
• Patent #: 11,562,402
• Filed: September 17, 2024
• Petitioner(s): Home Depot U.S.A., Inc.
• Patent Owner(s): Security Technology, LLC
• Challenged Claims: 1-19

2. Patent Overview

• Title: Targeted Advertising Using Quality Levels
• Brief Description: The ’402 patent discloses a method for targeted advertising where "quality levels" are calculated for users. The system determines first and second quality levels based on user profiles, searches, and purchases to identify user interest in sequential product categories and displays advertisements based on these calculated levels.

3. Grounds for Unpatentability

Ground 1: Obviousness over Grannan, Chung, and Yang - Claims 1-19 are obvious over Grannan in view of Chung and Yang.

• Prior Art Relied Upon: Grannan (Application # 2007/0244750), Chung (Patent 7,809,740), and Yang (a 2003 paper on pattern-based clustering).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that the combination of Grannan, Chung, and Yang taught all limitations of the challenged claims. Grannan disclosed a targeted advertising system that calculates a "product interest correlation" (PIC) score, analogous to the ’402 patent's "first quality level," based on user profiles, browsing history ("surfing"), and purchases. To the extent Grannan did not explicitly disclose certain features, Petitioner asserted Chung supplied them by teaching a behavioral targeting system that generates "mapped scores" based on a user's "conversion propensity," which is calculated from user events like search queries and historical clicks. For the limitation requiring that the quality level is based on a "unique identifier and clustering," which was added during prosecution to overcome prior art, Petitioner argued Yang taught this element. Yang described using pattern-based clustering of user behavior to differentiate individual users on a shared device, a known problem in the art at the time.
  • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine Grannan and Chung to improve the accuracy of Grannan's PIC score. Both references are in the targeted advertising field and address generating user scores to predict purchase likelihood. A POSITA would have been motivated to incorporate Chung’s more sophisticated "mapped score" analysis, which explicitly uses conversion propensity and historical clicks, to make Grannan’s system more effective. Furthermore, a POSITA would combine the teachings of Yang with the Grannan/Chung system to solve the well-known problem of identifying unique individuals on shared household computers, thereby improving the personalization and accuracy of the targeted advertising.
  • Expectation of Success: A POSITA would have had a reasonable expectation of success because combining these references involved applying known techniques to improve a known system for their intended purposes. Integrating Chung's conversion assessment into Grannan's scoring system and using Yang's established clustering method to refine user identification are predictable improvements, not inventive leaps.

Ground 2: Obviousness over Grannan, Chung, and Ramaiyer - Claims 1-19 are obvious over Grannan in view of Chung and Ramaiyer.

• Prior Art Relied Upon: Grannan (Application # 2007/0244750), Chung (Patent 7,809,740), and Ramaiyer (Application # 2012/0109956).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground is identical to Ground 1, except it substitutes Ramaiyer for Yang to teach the "clustering" limitation. Petitioner argued Ramaiyer, which was cited during prosecution but not applied in this combination, disclosed a profile prediction system that uses a clustering algorithm to identify one of multiple user profiles for a single device. Ramaiyer taught that user identification can be based on a clustering model that considers browsing characteristics to identify different clusters corresponding to a particular user or user profile.
  • Motivation to Combine: The motivation is identical to that for combining Yang in Ground 1. A POSITA would have been motivated to use Ramaiyer's clustering technique with the Grannan/Chung system to solve the known problem of distinguishing between multiple users sharing a single device, thereby improving the targeting accuracy of the advertising system.
  • Expectation of Success: As with Ground 1, a POSITA would have expected success in combining these elements because it represented the application of a known solution (clustering for user differentiation) to a known problem in the field of targeted advertising.

4. Key Claim Construction Positions

• Petitioner stated that for the purposes of the petition, it adopted the Patent Owner's purported licensee's position from parallel district court litigation that the claim terms should be accorded their plain and ordinary meaning. Petitioner reserved the right to argue for different constructions if needed in litigation.

5. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under §314(a) based on Fintiv factors would be inappropriate. The parallel district court litigation was in its earliest stages, with no scheduling order entered, no discovery commenced, and no Markman hearing date set.
• Petitioner contended that the Final Written Decision (FWD) in the IPR would issue approximately three months before the anticipated trial date in the parallel litigation, favoring institution.
• Petitioner also argued against denial under §325(d), stating the challenge was not substantially the same as those considered during prosecution. The primary references, Grannan and Chung, were not considered by the examiner, and Ramaiyer (cited in Ground 2) was not considered in the asserted combination.

6. Relief Requested

• Petitioner requests institution of an inter partes review and cancellation of claims 1-19 of the ’402 patent as unpatentable.