DCT
2:09-cv-00681
Amazon.com Inc v. Discovery Communications Inc
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Amazon.com, Inc. (Delaware)
- Defendant: Discovery Communications, Inc. (Delaware)
- Plaintiff’s Counsel: Corr Cronin Michelson Baumgardner & Preece LLP; Gibson, Dunn & Crutcher LLP
- Case Identification: 2:09-cv-00681, W.D. Wash., 05/15/2009
- Venue Allegations: Plaintiff alleges venue is proper in the Western District of Washington because Defendant has transacted business, supplied goods or services, and committed acts of patent infringement in the district.
- Core Dispute: Plaintiff alleges that Defendant’s e-commerce website infringes four patents related to refining search queries and generating personalized product recommendations.
- Technical Context: The patents address foundational e-commerce technologies for improving user search experiences and providing recommendations, which are critical for driving engagement and sales on large online retail platforms.
- Key Procedural History: The complaint does not mention any prior litigation, Inter Partes Review (IPR) proceedings, or licensing history related to the patents-in-suit.
Case Timeline
| Date | Event |
|---|---|
| 1998-06-15 | Earliest Priority Date for ’225 and ’986 Patents |
| 1998-09-18 | Earliest Priority Date for ’649 and ’722 Patents |
| 1999-12-21 | U.S. Patent No. 6,006,225 Issues |
| 2001-01-02 | U.S. Patent No. 6,169,986 Issues |
| 2001-07-24 | U.S. Patent No. 6,266,649 Issues |
| 2001-11-13 | U.S. Patent No. 6,317,722 Issues |
| 2009-05-15 | Complaint Filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 6,006,225 - "Refining Search Queries by the Suggestion of Correlated Terms from Prior Searches," Issued Dec. 21, 1999
The Invention Explained
- Problem Addressed: The patent describes a problem with existing search engines where suggested terms for refining a query are often unhelpful, processor-intensive to generate, or lead to "NULL" (no-result) searches (’225 Patent, col. 2:13-24). Existing methods often parsed the content of the current search results to find related terms, which was slow and inefficient (Id., col. 1:60-64).
- The Patented Solution: The invention proposes generating suggested search terms not from the immediate search results, but from a pre-compiled database of "query term correlation data" based on a history of prior successful searches by all users (’225 Patent, Abstract). This off-line process analyzes query logs to find which terms are frequently searched together, stores these correlations in an efficient look-up table, and uses that table to suggest refinements to a new user's query, ensuring suggestions are relevant and fast (Id., col. 4:1-20; Fig. 1).
- Technical Importance: This approach shifted query refinement from a real-time, content-analysis problem to a faster, data-driven historical analysis, reflecting the collective search behavior of the user community rather than just the content of a specific set of documents (’225 Patent, col. 2:31-36).
Key Claims at a Glance
- The complaint asserts infringement of "one or more claims" (Compl. ¶10). Independent claim 1 is representative:
- Processing search queries submitted by a plurality of users over time to generate "query term correlation data" reflecting how frequently terms appear together.
- Receiving a search query from a current user.
- Using the pre-generated correlation data to identify additional query terms related to the user's query.
- Presenting these additional query terms to the user for selection to refine the search.
U.S. Patent No. 6,169,986 - "System and Method for Refining Search Queries," Issued Jan. 2, 2001
The Invention Explained
- Problem Addressed: As a continuation of the ’225 Patent, the ’986 Patent addresses the same problem of providing users with valuable and efficient suggestions to refine their search queries and avoid frustrating "no result" outcomes (’986 Patent, col. 2:14-26).
- The Patented Solution: The ’986 Patent claims a system and method that uses a "history of search queries" submitted by a community of users over a period of time to identify and suggest refinements (’986 Patent, Claim 1). Like its parent, the solution relies on an offline analysis of historical query data to build a correlation table (Id., Figs. 4, 6), which is then used to generate suggestions for a live user search. The method emphasizes using data from successful prior queries to increase the likelihood that the suggested refinements will also be successful (Id., col. 3:1-5).
- Technical Importance: This patent further solidifies the concept of leveraging collective user search history as a data asset to improve the individual user experience, a foundational concept in modern e-commerce and web search functionality.
Key Claims at a Glance
- The complaint asserts infringement of "one or more claims" (Compl. ¶13). Independent claim 1 is representative:
- In a computer system implementing a search engine, receiving a search query from a user.
- Using a "history of search queries submitted to the search engine over a selected period of time by the community of users" to identify at least one refinement to that search query.
- Suggesting the identified refinement to the user.
U.S. Patent No. 6,266,649 - "Collaborative Recommendations Using Item-to-Item Similarity Mappings," Issued Jul. 24, 2001
- Technology Synopsis: This patent addresses product recommendations. It discloses a method that generates recommendations by analyzing item-to-item correlations from user purchase histories, rather than the more computationally expensive user-to-user similarity calculations (’649 Patent, Abstract). The system determines that two items are similar if a high percentage of customers who bought one also bought the other, with this data being pre-calculated offline and stored in a table for fast retrieval (Id., col. 10:49-67).
- Asserted Claims: The complaint asserts "one or more claims" (Compl. ¶16). Independent claims are 1, 16, 28, 33, and 43.
- Accused Features: The complaint alleges that the operation of the Discovery Store website infringes, which suggests the product recommendation features on the website are at issue (Compl. ¶16).
U.S. Patent No. 6,317,722 - "Use of Electronic Shopping Carts to Generate Personal Recommendations," Issued Nov. 13, 2001
- Technology Synopsis: This patent describes a method for generating personalized recommendations based on the current or recent contents of a user's electronic shopping cart (’722 Patent, Abstract). This approach aims to provide recommendations that are highly relevant to a user's immediate, short-term shopping interests, which may differ from their long-term purchase history (Id., col. 3:47-54). The system identifies items in the cart and uses an item-to-item similarity table to find and recommend related products (Id., Fig. 7).
- Asserted Claims: The complaint asserts "one or more claims" (Compl. ¶19). Independent claims are 1, 11, 16, 20, 28, and 40.
- Accused Features: The complaint's allegations against the Discovery Store website suggest that recommendation features tied to the shopping cart functionality are accused of infringement (Compl. ¶19).
III. The Accused Instrumentality
- Product Identification: The accused instrumentality is the "Discovery Store website," identified in the complaint as being accessible at http://store.discovery.com (Compl. ¶9).
- Functionality and Market Context: The complaint describes the website as an "online store" that sells products related to Discovery's media channels, such as the Discovery Channel, The Learning Channel, and Animal Planet (Compl. ¶9). The infringement allegations are directed at the "operation of the Discovery Store website," implying that its e-commerce functionalities for search and product recommendations are the features at issue (Compl. ¶¶10, 13, 16, 19). The complaint does not provide specific technical details about how the website's search or recommendation features operate. No probative visual evidence provided in complaint.
IV. Analysis of Infringement Allegations
The complaint provides only general allegations of infringement without mapping specific features of the Discovery Store website to the claim elements. The analysis below is based on the infringement theory implied by the assertion of the patents against an e-commerce website.
’225 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| (a) processing search queries submitted to the search engine by a plurality of users over a period of time to generate query term correlation data... | The backend systems of the Discovery Store website are alleged to analyze historical search logs from its user base to build a database of correlated search terms. | ¶10 | col. 7:44-67 |
| (b) receiving a search query from a user, the search query including at least one query term; | A user enters a search term into the search bar on the Discovery Store website. | ¶10 | col. 7:14-16 |
| (c) using at least the query term correlation data to identify a plurality of additional query terms that are deemed to be related to the at least one query term; and | The website's system allegedly uses the pre-compiled correlation data to look up terms historically associated with the user's submitted term. | ¶10 | col. 7:22-31 |
| (d) presenting the plurality of additional query terms to the user for selection to allow the user to refine the search query. | The Discovery Store website allegedly displays suggested or related search terms to the user on the search results page. | ¶10 | col. 7:32-38 |
- Identified Points of Contention:
- Technical Questions: A central question will be how the Discovery Store website actually generates its related search suggestions. Does the complaint provide any evidence that the system uses "query term correlation data" derived from historical user searches, as claimed, or does it use an alternative method such as analyzing the content of the current search results or using a manually curated thesaurus? The complaint's lack of detail makes this a primary point for discovery.
’986 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| receiving a search query submitted by a user, the search query comprising at least one term; | A user submits a search query on the Discovery Store website. | ¶13 | col. 7:14-16 |
| using a history of search queries submitted to the search engine over a selected period of time by the community of users to identify at least one refinement to the search query; and | The website's system is alleged to access a database of historical community search data to identify potential refinements for the current user's query. | ¶13 | col. 7:44-54 |
| suggesting the at least one refinement to the user. | The website allegedly presents the identified refinement (e.g., a related term or modified query) to the user for selection. | ¶13 | col. 7:32-38 |
- Identified Points of Contention:
- Scope Questions: The case may raise the question of what constitutes a "history of search queries." Does this require logging and processing raw user inputs over time, as described in the patent, or could a broader interpretation cover other forms of historical data? The determination will depend on evidence of how Discovery's system is architected.
V. Key Claim Terms for Construction
- The Term: "query term correlation data" / "history of search queries"
- Context and Importance: These related terms from the ’225 and ’986 patents are central to the core novelty asserted: using historical, collective user behavior to refine searches. The infringement analysis will depend entirely on whether the accused system generates suggestions using this specific type of historical data, as opposed to other methods. Practitioners may focus on this term because it distinguishes the claimed invention from prior art methods that relied on content analysis of search results.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The claims use the general phrase "correlation data," which could arguably encompass various statistical relationships beyond the specific frequency-counting embodiment. Claim 1 of the ’225 patent requires data that "reflect[s] frequencies with which query terms appear together," which could be one of many ways to implement a broader "correlation" concept.
- Evidence for a Narrower Interpretation: The specification provides a detailed embodiment where correlation data is generated by an offline process that parses daily log files, counts co-occurrences of terms in successful queries, and stores the results in a "correlation table" (’225 Patent, col. 8:43-67; Fig. 4). A defendant may argue that the claims should be limited to this disclosed process, particularly the requirement that the data be generated from past user queries, not other data like purchase history.
VI. Other Allegations
- Indirect Infringement: The complaint alleges that Discovery has been "indirectly infring[ing], solely or jointly with others, or induc[ing] others to infringe one or more claims" of the asserted patents (Compl. ¶¶10, 13, 16, 19). The complaint does not plead specific facts to support the knowledge and intent required for a claim of induced infringement, such as alleging that Discovery instructs its users to operate the website in an infringing manner.
- Willful Infringement: The complaint does not contain an allegation of willful infringement.
VII. Analyst’s Conclusion: Key Questions for the Case
The disposition of this case will likely depend on the answers to two central questions that are not addressed by the complaint's high-level allegations:
- A question of technical operation: How does the Discovery Store's search and recommendation functionality actually work? The core of the dispute will be a factual comparison between the specific methods claimed in the Amazon patents—which rely on analyzing historical user queries and purchase data to create item-to-item or term-to-term correlation tables—and the architecture of the accused website.
- An evidentiary question of proof: Given the notice-pleading style of the complaint, the case will turn on what is revealed in discovery. A key issue will be whether Amazon can produce evidence demonstrating that the accused website performs the specific, multi-step data processing and suggestion-generation methods required by the asserted claims, or if Discovery's systems operate on a fundamentally different, non-infringing principle.