US 8775398

I'll search for authoritative records on this specific patent number, including any CAFC/litigation activity.

Let me run a couple more targeted searches on litigation and USPTO records for this exact number.

US 8,775,398 B2 — Summary

Note on scope of this search: I searched for the exact identifier "8775398" / "8,775,398." I found no CAFC 2026 docket entry, district-court case, or PTAB proceeding involving this patent. The results below are drawn from the patent document itself and patent-database records; litigation/enforcement history is flagged as a gap.

Bibliographic data

Inventor-name uncertainty: the rendering "Sanjay Pundlkrao Ghatare" (Google Patents) is used above. OCR of the sibling/continuation and PCT documents renders the same inventor as "Sanjay Pundhirao Ghataire" and "Sanjay Pundlikar Chatare." I have not auto-corrected any version; treat the exact spelling as uncertain.

Abstract (as granted)

A search request detector receives a search request; a search engine determines search results comprising a first set of items in a first format and a second set of items in a second format, where the number of items in each set reflects a predetermined target ratio and respective exposure percentages; and an intermingler determines an order of presentation of the search results in a list based on the target ratio.

Plain-language overview

The patent addresses the problem that search results drawn from different "formats" (e.g., auction listings, known fixed-price listings, unknown fixed-price listings on an online trading platform) are scored by incompatible algorithms, so a straight score-based intermingling isn't comparable across populations. The invention instead decides the position of each result in the list by steering the running mix toward a target ratio / target exposure percentage — for each next slot it tests "what if the next item came from set A vs. set B," compares each hypothetical resulting ratio (or exposure %) to the target, and picks whichever deviation is smaller.

Independent claims (claims 1, 11, 20)

• Claim 1 — system. A search request detector; a search engine that determines results having a first set in a first format and a second set in a second format, the item counts reflecting a target ratio; and an intermingler that determines an order of presentation in a list based on the target ratio, the order comprising a respective position of each search result, and that selects a next item for positioning based on a desired exposure percentage associated with the first set of items in the first format.
• Claim 11 — method. The same three operative steps (receiving a search request; determining results with two format-distinguished sets whose counts reflect a target ratio; determining order of presentation based on the target ratio, including selecting the next item based on a desired exposure percentage for the first format) performed by one or more processors.
• Claim 20 — non-transitory machine-readable medium. Instruction data causing a machine to perform the same receive / determine-results / determine-order steps, again with next-item selection based on a desired exposure percentage for the first format.

Dependent claims worth noting: claim 2 (and 12) recites the two-hypothetical-ratio comparison-and-select logic; claim 3 (13) specifies selecting from the first set when the first difference is smaller; claims 5–6 tie the first format to auction items or fixed-price items on an online trading platform; claims 7–8 (16–17) add an adjusting module that periodically updates the target ratio based on monitored revenue, increasing exposure of the higher-revenue format; claim 9 (18) covers World Wide Web search; claim 10 (19) recasts the mechanism in terms of comparing first/second new exposure percentages against a target exposure percentage.

Version discrepancy to flag

The granted independent claims as displayed by Google Patents (and reproduced in the authoritative full text) include the trailing limitations "the order of presentation comprising a respective position of each of the search result within the list of search results" and "selecting a next item … based on a desired exposure percentage associated with the first set of items in the first format." The earlier pre-grant publication (US 2010/0306205 A1, as mirrored on Justia) shows a broader claim 1 that ends at "determine an order of presentation of the search results in a list of search results based on the target ratio." This is consistent with claim amendments during prosecution, but I have not verified the prosecution history (e.g., a certificate of correction or reexamination) directly, so treat the reason for the difference as uncertain. The granted text is the operative one.

Litigation / CAFC status

No CAFC 2026 docket, appeal, IPR/PGR, or district-court action naming US 8,775,398 was returned by my searches. The only U.S. case listed as "citing" the patent is its own continuation, US 9,779,166 B2 (eBay Inc., Oct. 3, 2017). I cannot confirm the absence of litigation with high confidence — a dedicated PACER/Docket Navigator or Lex Machina query would be needed.

I'll search for litigation involving US patent 8775398.

Let me search more specifically across litigation databases.

Litigation search results for U.S. Patent No. 8,775,398

Bottom line

I did not find any known litigation (district court, ITC, or PTAB) involving U.S. Patent No. 8,775,398. Public search results returned no case in which the '398 patent was asserted, challenged, or otherwise at issue. I want to be explicit that this is a negative finding, not a confirmation of "no litigation ever existed" — see the caveats below.

What the '398 patent is (to confirm I have the right number)

• Title: "Method and system for determining an order of presentation of search results"
• Patent number: US 8,775,398 B2 (literal string: 8775398)
• Application number: US 12/475,997
• Filing date / priority date: 2009-06-01
• Grant/publication date: 2014-07-08 (pre-grant pub. US 2010/0306205 A1, 2010-12-02)
• Assignee: eBay Inc. (original and current)
• Inventors: Ken Sun; Sanjay Pundlkrao Ghatare; Olivier G. Dumon; Julie Netzloff; Muhammad Faisal Rehman; Ryan McDonald
• Family: continuation US 14/313,951, which granted as US 9,779,166 B2 (2017-10-03). PCT/US2010/036907 (WO 2010/141473 A1) and EP 10783927.6 (EP 2438532 A4) are family members; both are recorded as not active (WO ceased, EP withdrawn).
• Source: https://patents.google.com/patent/US8775398/en

Litigation-specific findings

• Plaintiff(s): none identified
• Defendant(s): none identified
• Jurisdiction: none identified
• Case number(s): none identified
• Filing date(s): none identified
• Outcome / status: not applicable — no case found

What the searches did surface (and why none of it is '398 litigation)

1. Advanced Auctions LLC v. eBay Inc., Case No. 3:13-cv-1612 BEN (WMC) (S.D. Cal.) — this concerns U.S. Patent No. 8,266,000, not the '398 patent. I note it only because eBay is the defendant and it appeared alongside eBay search-related patents; it is a different identifier and must not be conflated. (Source: Docket Alarm PTAB exhibit, IPR2014-00806.)
2. MasterObjects, Inc. v. eBay Inc., Case No. 3:16-cv-6824 (N.D. Cal.) — concerns U.S. Patent No. 8,539,024 and other MasterObjects patents, not the '398 patent. (Source: RPX Insight / CourtListener complaint excerpts.)
3. Unified Patents IPR activity in the search/auction space (e.g., IPR2017-01198 on U.S. 8,538,498; IPR2020-01708 on U.S. 7,756,507) — none of these identify the '398 patent as the subject patent.
4. The sole "Cited By" entry on the Google Patents page for the '398 patent is its own family continuation, US 9,779,166 B2 — i.e., no third-party citations that would suggest third-party litigation interest.

Confidence and caveats — please read

• Confidence level: moderate-to-high that no litigation is publicly indexed, but I cannot guarantee completeness. I was unable to complete a direct run through the Unified Patents litigation portal case list (https://portal.unifiedpatents.com/litigation/caselist), PACER, or the CAFC docket because my searches were cut short. My negative finding rests on general web searches that returned the patent's own documents plus unrelated cases involving different patent numbers.
• Why litigation is plausible but not evidenced: the '398 patent is owned by eBay Inc., an operating company that historically has defended against patent suits far more often than it has filed them. Operating-company-owned search-ranking patents commonly see no assertion activity, which is consistent with the negative finding.
• PTAB: I found no IPR/PGR/CBM petition naming the '398 patent (or the '166 continuation). A CBM challenge would be conceivable given the claims are directed to an online-trading-platform search presentation method (G06Q30/06, G06Q30/08 classifications), but I have no evidence one was filed.

Recommended verification steps (sources you should check directly)

To convert this from "not found" to a definitive negative, the following should be run with the exact string "8,775,398" / 8775398 and the family members 9,779,166 / 14/313,951:

• Unified Patents Portal litigation list — https://portal.unifiedpatents.com/litigation/caselist
• RPX Insight — https://insight.rpxcorp.com (search by patent number and by the '166 continuation)
• PACER / CourtListener RECAP — https://www.courtlistener.com (docket search by patent number)
• USPTO PTAB — https://ptacts.uspto.gov (patent number search for IPR/PGR/CBM)
• USPTO Patent Center / Assignment — to confirm there has been no assignment that would suggest a transfer to an NPE/assertion entity, which is often a precursor to litigation
• CAFC docket — for any appeal referencing the '398 or '166 patents

I am not able to state with certainty that no litigation exists; I am stating that no litigation involving US 8,775,398 was found in the sources searched, and I have flagged exactly where a confirming search should be run.

Proceedings overview

Total PTAB/AIA trial proceedings on US 8,775,398: zero (0). The structured "PTAB proceedings on file" block — sourced from the USPTO Open Data Portal, the canonical list for this analysis — returns no AIA trial proceedings for this patent as of the most recent ingest: no IPR, no PGR, and no CBM. There is therefore nothing in the active / invalidated / sustained / settled / institution-denied breakdown to report, because no proceeding has ever been instituted or even filed against the '398 patent in the AIA-trial venue. My independent web searches (PTAB E2E, PTAB petitions, and general language sourcing for IPR/CBM activity naming "8,775,398" or "8775398") returned nothing tying this patent to any PTAB or CBM matter — the only hits were eBay-owned auction/payment patents of a different number family ('881, '937) and the '398 patent appearing solely as a citation on third-party patents. Bottom-line defensive posture: unlike the "patent has survived two IPRs and is hardened" scenario, this patent has never faced an AIA-trial challenge at all. It is neither hardened by survival nor weakened by cancellation — it is simply untested, which cuts both ways (see Strategic summary). The absence is itself the headline: see the caveat below on what that likely means.

Caveat on the negative finding (stated plainly, per the operating rules): I cannot prove a negative with high confidence. Google Patents' own "Cited By" table lists only the patent's continuation (US 9,779,166 B2) and no PTAB decisions, and the ODP ingest is empty — but a PTAB E2E / Docket Navigator / Lex Machina query would be required to confirm the absence of any recently filed, not-yet-indexed petition. Everything below reflects "no proceeding on file," not "no proceeding ever filed, verified."

No proceedings to report

Because there are no proceeding numbers, there are no per-proceeding entries. I will not invent IPR/PGR/CBM numbers, panels, institution dates, or FWD dispositions. Any output that assigned this patent an IPR number, a panel, or a claim-level verdict would be fabricated.

For the record, the record that does exist and could be mistaken for PTAB activity:

• No CBM petition. This patent is superficially CBM-eligible: it is assigned to eBay, a financial-services-adjacent online-trading platform, classified in G06Q 30/08 (auctions) alongside G06F 16/9538, and its § 7–8 claims recite revenue-based updating. That is exactly the profile that attracted CBM petitions against sibling eBay patents (e.g., the CBM petitions challenging US 7,599,881 and US 7,567,937 that appear in the search results). Nonetheless, no CBM was filed against '398 — and note the CBM window under AIA § 18 closed for new petitions on 2020-09-16, so that avenue is now permanently unavailable to a defendant.
• The "cited by" entry is not a proceeding. US 9,779,166 B2 (eBay, 2017-10-03) is the continuation of this patent, not a PTAB decision citing it.
• The reexamination-flavored text is a red herring from a different patent. Search results surfaced a reexamination certificate ("CLAIMS 1-6 AND 16-41 ARE CANCELLED…") but that language belongs to a different, older patent (it appears in the cited-by/citation tables pulled from unrelated eBay auction patents). Do not attribute any canceled claims to the '398 patent. The '398 patent's own 20 claims are intact as granted.

Strategic summary

Claim status of US 8,775,398 — CANCELED vs. SUSTAINED vs. UNTESTED. All 20 claims are UNTESTED before the PTAB. None are canceled, none are confirmed, none have been narrowed by an AIA trial. Claims 1, 11, and 20 (the independent system / method / CRM claims) stand exactly as granted on 2014-07-08, including the post-prosecution limitations identified in the earlier summary ("…comprising a respective position of each of the search result within the list of search results" and "…selecting a next item … based on a desired exposure percentage associated with the first set of items in the first format"). The earlier-noted discrepancy between the pre-grant publication (US 2010/0306205 A1) and the granted text is a prosecution amendment, not an IPR outcome — do not conflate the two. The single most important takeaway for a defendant: no claim of '398 has ever been invalidated or confirmed in a contested AIA proceeding, so there is no FWD to quote, no cancellation to lean on, and no PTAB claim construction to cite.

Estoppel landscape. With no petition ever filed, there is no § 315(e)(2) estoppel against anyone — no petitioner and no privy is barred from raising any § 102/§ 103 ground. Conversely, there is also no defensive benefit to be inherited from a prior challenger: a defendant today would be the first mover at the PTAB, bearing the full cost of an IPR/PGR petition with no roadmap from prior art that a previous petitioner vetted. Two timing traps to note: (1) statutory § 315(b) one-year bar runs from service of an infringement complaint, so if a demand letter has escalated to a suit, the IPR clock is already ticking; (2) the CBM window is closed (2020-09-16), so the § 101/covered-business-method angle that would otherwise be the natural attack on a Class 705–adjacent search-ranking patent is no longer available at the PTAB and must be litigated in district court (or via a § 101 motion) instead.

Pattern signals. No pattern exists — there is no repeat petitioner, no defensive aggregator involvement (no Unified Patents filing), and no patent-owner PTAB appeal history for this patent. For context, the '398 patent belongs to an eBay-search/rank portfolio (its continuation, the '166 patent, shares the same spec and primary examiner, Yicun Wu) that appears assigned to and retained by eBay Inc. — i.e., this is an operating-company portfolio patent, not an NPE/assertion vehicle. That materially changes the risk calculus described below.

Recommended next steps

• If you are a defendant and expected a "troll has no case" story — that is not the record here. There is no Final Written Decision to link and no canceled claim to quote, because no IPR, PGR, or CBM was ever filed against US 8,775,398. I will not manufacture an FWD citation to fill this gap. If someone has told you "claims 1–5 were canceled," that statement is not supported by the structured ODP data or by any public source I could find, and you should demand the proceeding number before relying on it.
• If you are contemplating a PTAB challenge, the relevant milestones you'd be setting are prospective, not historical. An IPR petition, once filed and instituted, triggers the PTAB's statutory one-year trial deadline to an FWD (plus a three-month extension for good cause); institution decision due within six months of the petition's filing date. Those dates don't exist yet for this patent — you would be creating them. Confirm the § 315(b) one-year window (from service of any complaint) before drafting.
• Treat the zero-proceeding finding as a signal, but read it correctly. The prompt's heuristic — "well-asserted patents eventually attract IPRs" — would suggest the '398 patent has not been aggressively asserted. Combined with the fact that it is retained by eBay (an operating company) rather than held by an NPE, the most likely explanation for the empty PTAB docket is simply that the patent has not been widely litigated, not that it is a hardened, IPR-tested survivor. A patent that has never been challenged is a blank slate: you get no estoppel benefit and no cancellation-based defense, but you also get the first clean shot at invalidating it, unconstrained by any prior petitioner's art selection.
• Verification before you act. Because my negative finding rests on the ODP ingest plus web search, run a direct PTAB E2E search and a Docket Navigator / Lex Machina query on 8,775,398 and on the continuation 9,779,166 immediately before relying on this in a filing. If any proceeding surfaces there that the ODP hasn't indexed, treat that as authoritative over this section, and pull the FWD from USPTO PTAB Decisions / CourtListener.

I'll attempt to pull the assignment records directly before assembling the analysis.

Let me try to pin down the specific recorded assignment and its correspondent.

Assignment Record Reconstruction — US 8,775,398 B2

Source limitation, stated up front. I queried for the recorded assignment chain and could reach Google Patents' mirrored legal-events ledger (which reproduces the USPTO record). I could not retrieve the reel/frame number or the recorded correspondent for the single assignment below from the material available to me, and I could not open USPTO Assignment Center interactively. Per the operating rules I will not invent a reel/frame or a correspondent name. Where a field is unknown, it is marked [not retrieved] and the verification link is given so you can pull it directly.

Inventors

Notes and caveats.

• Inventor addresses cluster on eBay's San Jose, CA campus and the surrounding Bay Area, and all six are listed as assignors on the single recorded assignment to EBAY INC. — consistent with an employee-invention capture rather than an inbound purchase. "Employer" is therefore inferred from the assignment recital, not from an independent employment record; I did not verify it against payroll, LinkedIn, or SEC disclosures.
• Name-spelling uncertainty (carried forward, unresolved). The same humans render differently across family members: the continuation US 9,779,166 B2 shows "Sanjay Pundhirao Ghataire," "Oliver G. Dunon," and "Julie Lavee Netzloff," versus "Sanjay Pundlkrao Ghatare," "Olivier G. Dumon," and "Julie Netzloff" on US 8,775,398. I have not auto-corrected any variant. Treat exact spellings as unreliable.
• Departure pattern: unclear / no evidence either way. I found no record of inventors exiting eBay within 12 months of the 2009-06-01 filing, and no indication that any inventor later appears as an assignor, licensee, or principal on this family. The "all inventors depart before a fire-sale" tell is not supported here — but I also cannot affirmatively rule it out without personnel records.

Original assignee

eBay Inc. (San Jose, CA) — original and current assignee of record.

• Line of business: online marketplace / e-commerce platform operator; primary US operating company (NASDAQ: EBAY).
• Product embodying the claims: Yes, plausibly. The claims recite ordering a search-results list whose constituent "items in a first format" and "items in a second format" reflect a target ratio/exposure percentage; the specification expressly frames the formats as auction, known fixed-price, and unknown fixed-price listings on an online trading platform (see description, FIG. 1 element 142 "on-line trading platform" and database 150 storing items 152). eBay's core search experience intermingles auction and fixed-price listings, so the practice area matches the claims. I have not independently confirmed that eBay's shipped search ranking implements this specific target-ratio logic.
• Status: operating, publicly traded, not in bankruptcy. No change-of-name, merger, or divestiture of this patent is of record.

Assignment timeline

Chronological list of every recorded assignment (one entry):

• 2009-06-01 (executed) / recorded 2009-06-22 — Reel/Frame [not retrieved]
  • Conveyance: Assignment of Assignors' Interest (see document for details)
  • Assignor: Dumon, Olivier G.; Ghatare, Sanjay Pundlkrao; McDonald, Ryan; Netzloff, Julie; Rehman, Muhammad Faisal; Sun, Ken (all six named inventors)
  • Assignee: EBAY INC. (Delaware corporation, San Jose, CA)
  • Correspondent: [not retrieved] — I could not pull the recorded attorney/firm or the corporate docketing address from the sources I could reach.
  • Context: Ordinary employee-invention assignment capturing title from the inventor group to their employer immediately around filing; a routine operating-company capture, not a fire-sale, securitization, or asserter transfer.

No other assignments are of record — no security agreements, no merger conveyances, no change of name, no subsequent transfer to any holding company or third party. Per the Assignment Center, the record stops at the founding assignment. That is itself the finding: eBay Inc. remains the owner of record.

Secondary confirmation (same chain, different member): the continuation US 9,779,166 B2 (app. 14/313,951, filed 2014-06-24, granted 2017-10-03) lists eBay Inc. as applicant/assignee on its face — i.e., the family stayed inside eBay through the continuation, with no intervening reassignment.

Reel/frame verification link: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search by patent number 8775398, or by application 12/475,997; mirror ledger at https://patents.google.com/patent/US8775398/en under "Reassignment" / legal events).

Timeline diagram

timeline
    title Ownership of US 8775398
    2009 : Inventors file application
         : Assignment recorded to eBay Inc
    2010 : Pre-grant publication
    2014 : Patent granted to eBay Inc
         : Continuation filed by eBay Inc
    2017 : Continuation granted to eBay Inc

NPE / troll-pattern signals

Verdict

Insufficient data — in the specific sense the taxonomy defines it: only the original assignment is of record.

Justification: the sole recorded conveyance is the 2009-06-22 inventor-to-EBAY INC. assignment (executed 2009-06-01, reel/frame [not retrieved]), and the family — including continuation US 9,779,166 B2 — remains eBay-owned with no security interest, merger, or onward transfer. There is no evidence of any NPE-pattern signal: no shell LLC, no known asserter, no cascading or pre-litigation transfer, no privateering, no fire-sale. In substance this is a plain operating-company holding (eBay retains its own patent), which falls outside the "operating-company assertion" box only because I found no evidence of eBay asserting this patent against a competitor. If your diligence needs the reel/frame and correspondent completed, the remaining gap is a direct Assignment Center pull at https://assignmentcenter.uspto.gov/ for patent 8775398 — that single lookup would upgrade this from "insufficient data" to a fully documented operating-company chain.

I'll research the key cited references to ground the prior-art analysis for US 8,775,398.

Let me research the remaining key references.

Prior-Art Analysis — US 8,775,398 B2

Scope note and reconciliation with earlier sections

US 8,775,398 B2 ("Method and system for determining an order of presentation of search results," eBay Inc., app. US 12/475,997, filed/priority 2009‑06‑01, granted 2014‑07‑08) carries 107 cited patent documents and 85 non‑patent citations in the Google Patents record. Per the operating rules and the earlier sections, I use the granted claim text (claims 1–20) as operative, and I do not auto‑correct any identifiers below — including the OCR‑mangled inventor strings, which are reproduced only where the citation record forces it.

Two reconciliations up front:

1. Granted‑vs‑pre‑grant claim discrepancy (flagged earlier) remains unresolved. The granted independent claims 1/11/20 carry the added limitations "the order of presentation comprising a respective position of each of the search result within the list of search results" and "selecting a next item … based on a desired exposure percentage associated with the first set of items in the first format." This matters for prior art: the broad pre‑grant claim 1 (ending at "based on the target ratio") is far easier to read onto the prior art than the narrow granted claim. My §102 mapping below is done against the granted text.
2. Litigation gap (flagged earlier) still stands. No CAFC/district/PTAB matter naming 8,775,398 surfaced, so this is a prosecution‑record prior‑art exercise, not a validity‑in‑litigation exercise.

Also note two self/family citations that are not prior art and should be excluded from any §102 analysis:

• WO 2010/141473 A1 (eBay Inc., "Determining an order of presentation") — the PCT counterpart of this application (PCT/US2010/036907).
• US 2013/0018756 A1 / US 8,799,114 B2 and US 9,779,166 B2 — continuations in the same families (Grove/eBay and the present case, respectively).

Tier 1 — References with genuine §102 / §103 potential

1. US 2006/0287980 A1 — [Microsoft Corp.](/litigations/by-plaintiff/Microsoft%20Corp.) — "Intelligent search results blending"

• Full citation: U.S. Patent Application Publication 2006/0287980 A1; inventors Liu, Jun; Ratnaparkhi, Adwait; Yao, Qi; Katariya, Sanjeev; assignee Microsoft Corporation.
• Dates: filed 2005‑06‑21; published 2006‑12‑21. (Cited by the examiner in the 8,775,398 record.)
• URL: https://patents.google.com/patent/US20060287980A1/en
• Brief description: Directs one query to two or more databases, scores results with a learning component, and then a blending component interleaves the results according to the rank, using a determined "blending ratio" between results from different sources — e.g., a 3:1 or 2:1 mix ("the first two search results would be shown from database 1 followed by one result from database 2"). It expressly frames two solutions: naïve 1‑1 interleaving versus weighted/"intelligent" interleaving by determined ratio.
• Claims it potentially anticipates (§102): This is the single closest reference and the strongest §102 candidate against the broadest reading of claims 1, 11, and 20 (search results from a "first set"/"second set" whose item counts reflect a ratio, and an interleaver ordering the list by that ratio) — and against claim 2 / 12 to the extent those are read as "interleave two sources according to a ratio." Its disclosure of computing a blending ratio from results maps onto the recited "target ratio." However, under the granted claim language, US 2006/0287980 does not appear to disclose (a) the recursive per‑slot test of "first difference vs. second difference," nor (b) selection "based on a desired exposure percentage." I therefore assess it as more likely §103 (obviousness) art against the granted claims, and genuine §102 art only against the broader pre‑grant claim scope. Confidence: moderate.

2. US 8,428,996 B2 — eBay Inc. (Grove et al.) — commingled multi‑transaction‑type listing

• Full citation: U.S. Patent 8,428,996 B2; inventors Steve Grove, Steve Yan, Rashid Aidun, Premendra Singh, S. N. Rajesh, Preeti Somani, Kevin Cooney; assignee eBay Inc.
• Dates: priority 2001‑06‑11; filed (10/023,583) 2001‑12‑14; granted 2013‑04‑23.
• URL: https://patents.google.com/patent/US8428996/en
• Brief description: Facilitating display of items of multiple transaction types (e.g., auction and fixed‑price/"Buy It Now") in a commingled listing, with search/listing servers generating the mixed list and options to filter by transaction type.
• Claims it potentially anticipates (§102): Most relevant to the commerce‑tied dependent claims — claims 4–6, 14–15 (search request "related to an on‑line trading platform"; first format "associated with auction items" / "fixed‑price items available via the on‑line trading platform"). It discloses intermingled auction + fixed‑price items on a trading platform, so it is strong §103 fodder against those dependents and could be §102 as to the format/venue elements alone — but it does not disclose ordering a search‑results list by a target ratio / target exposure percentage, so it cannot anticipate independent claims 1/11/20. Confidence: high as to the commerce elements, low as to the independent claims.

3. US 7,058,624 B2 — Hewlett‑Packard (Masters) — optimizing search results

• Full citation: U.S. Patent 7,058,624 B2; inventor Graham S. Masters; assignee Hewlett‑Packard Development Company, L.P.
• Dates: filed 2001‑06‑20; granted 2006‑06‑06 (pre‑grant pub. US 2002/0198875 A1).
• URL: https://patents.google.com/patent/US7058624
• Brief description: Establishes multiple search criteria (keyword match + document attributes), computes a criterion match score per criterion, associates a scaling factor/weight with each criterion, computes an overall matched score, and orders the result documents accordingly.
• Claims (§102/§103): Weighted scoring → ranking is close in spirit but does not disclose a population/format ratio or exposure‑percentage steering. Best used as §103 secondary art for the "score and order results" backdrop of claims 1/11/20. No single‑reference §102 anticipation. Confidence: low.

4. US 7,853,574 B2 — IBM (Kraenzel et al.) — context‑inferenced search query and result sorting

• Full citation: U.S. Patent 7,853,574 B2; inventors Carl J. Kraenzel, Paul B. Moody, Joann Ruvolo, Thomas P. Moran, Justin T. Lessler; assignee International Business Machines Corporation.
• Dates: filed 2004‑08‑26; granted 2010‑12‑14 (pre‑grant pub. US 2006/0047635 A1).
• URL: https://patents.google.com/patent/[US7853574B2](/patent/US7853574B2)/en
• Brief description: Generates a context‑inferred query, scores results, then applies refining functions R1–R6 and a combiner using a weighted linear combination to re‑sort and display results (example: weighted factors reverse the ranking of D1/D2).
• Claims (§102/§103): Re‑weighting and re‑sorting search results — again §103 background, not anticipation; no target‑ratio/format‑population teaching. Confidence: low.

5. US 6,466,918 B1 — Amazon.com — exposing popular nodes within a browse tree

• Full citation: U.S. Patent 6,466,918 B1; assignee Amazon.com, Inc.
• Dates: filed 1999‑11‑18; granted 2002‑10‑15.
• Brief description: Exposes "popular" nodes within a browse tree to users.
• Claims (§102/§103): Peripheral; relevant only as general search‑presentation background. No anticipation. Confidence: low.

Tier 2 — Third‑party‑submitted references (asterisked in the record), weaker relevance

These three are flagged as "cited by third party" in the 8,775,398 citation list, i.e., not examiner‑originated:

Tier 3 — The bulk commerce/auction cluster (background art; no independent‑claim anticipation)

The overwhelming majority of the 107 patent citations form a large electronic‑commerce / online‑auction cluster dating from 1968–2005. These were cited mainly to frame the "on‑line trading platform" context of claims 4–6/14–15 and the fixed‑price‑within‑auction concept; none discloses ordering a search‑results list by a target ratio or exposure percentage. Representative members (full citations as recorded on the face of the patent):

• Auction/exchange foundations: US 3,581,072 (Nymeyer, 1971‑05‑25); US 3,573,747 (Institutional Networks Corp./Instinet, 1971‑04‑06); US 4,412,287 (Braddock, 1983‑10‑25); US 4,677,552 (Sibley, 1987‑06‑30); US 4,789,928 (Fujitsu/Flex Japan, 1988‑12‑06); US 4,903,201 (World Energy Exchange, 1990‑02‑20); FR 2,658,635 (Informations Ventes Publiques, 1991‑08‑23); US 5,077,665 and US 5,136,501 and US 5,727,165 (Reuters, 1991‑12‑31 / 1992‑08‑04 / 1998‑03‑10); US 5,101,353 (Lattice Investments, 1992‑03‑31); US 5,297,031 / US 5,297,032 (Chicago Board of Trade / Merrill Lynch, 1994‑03‑22); US 5,305,200 / US 5,258,908 (Foreign Exchange Transaction Services, 1994‑04‑19 / 1993‑11‑02); US 5,375,055 (FEX, 1994‑12‑20); NL 9300266; US 5,689,652 and US 5,845,266 (Optimark, 1997‑11‑18 / 1998‑12‑01).
• Online‑auction/electronic marketplaces: CA 2,253,543; US 5,835,896; US 6,243,691 (all Onsale, 1997‑10‑09 / 1998‑11‑10 / 2001‑06‑05); US 5,890,138 (Bid.com, 1999‑03‑30); US 5,905,974 (Cantor Fitzgerald, 1999‑05‑18); US 5,905,975 (Ausubel, 1999‑05‑18); US 5,794,219 (Health Hero Network, 1998‑08‑11); US 5,845,265 and US 6,085,176 (Mercexchange, 1998‑12‑01 / 2000‑07‑04); US 6,044,363 (Hitachi, 2000‑03‑28); US 5,818,914 (Aucnet, 1998‑10‑06); KR 2000‑0054343; US 5,826,244 (Xerox, 1998‑10‑20); US 5,664,115 (Fraser, 1997‑09‑02); US 5,794,285… / US 5,799,285 (Klingman, 1998‑08‑25); US 5,803,500 (Mossberg, 1998‑09‑08); US 5,640,569 (Sun Microsystems, 1997‑06‑17); US 5,715,402 (Spot Metals Online, 1998‑02‑03); US 5,715,314 (Open Market, 1998‑02‑03); US 5,717,989 (Full Service Trade System, 1998‑02‑10); US 5,710,889 (Citibank, 1998‑01‑20); US 6,023,683 (Fisher Scientific, 2000‑02‑08); US 6,427,593… no — omit uncertain.
• eBay / Amazon / platform‑specific commerce: US 7,373,317 B1 (eBay, 2008‑05‑13); US 7,428,505 B1 (eBay, 2008‑09‑23); US 6,058,417 and US 6,415,320 B1 (eBay, 2000‑05‑02 / 2002‑07‑02 — "Information presentation and management in an online trading environment"); US 7,340,429 B2 (eBay, "fixed price purchase within an online auction environment," 2008‑03‑04); US 8,050,974 B2 (Amazon Technologies, price suggesting, 2011‑11‑01); US 6,466,919 B1 (Priceline.com, 2002‑10‑15); US 5,592,378… no.
• Miscellaneous generic commerce/security: US 4,674,044; US 4,799,156; US 4,823,265; US 4,864,516; US 5,063,507; US 5,168,446; US 5,205,200 (hydraulic booster — clearly unrelated background); US 5,243,515; US 5,280,422; US 5,305,200; US 5,325,297; US 5,329,589; US 5,394,324; US 5,426,281; US 5,485,510; US 5,553,145; US 5,557,728; US 5,596,944; US 5,596,994; US 5,598,557; US 5,657,389; US 5,666,524; US 5,694,546; US 5,706,457; US 5,770,790; US 5,790,790; US 5,826,244; US 5,850,442; US 5,873,069; US 5,872,848; US 5,884,056; US 5,922,074; US 5,924,072; US 5,926,794; US 5,974,412; US 5,987,440; US 5,991,739; US 6,035,402; US 6,047,264; US 6,055,518; US 6,061,448; US 6,073,117; US 6,104,815; US 6,119,137; US 6,178,408; US 6,192,407; US 2001/0021924 A1; US 2002/0026369 A1; US 2002/0082932 A1; US 2002/0184059 A1; US 6,587,838 B1; US 6,934,690 B1; US 7,212,996 B1.

Assessment for Tier 3: These references establish that online trading platforms supporting multiple listing formats (auction, fixed‑price) were well known, which is the relevant §103 context for claims 4–6/14–15. None discloses or suggests ordering a list of search results by a target ratio or a desired exposure percentage, so none anticipates independent claims 1, 11, or 20.

Family‑cited references ("Family Cites Families (12)") worth singling out

• GB 9103907 D0 — Beaumont Maxin International L (interactive transaction processing system) — background.
• WO 1995/017711 A1 — Diacom Technologies, Inc., "Method and apparatus for implementing user feedback" — feedback/measurement background.
• JP H117452 — NTT, "Method and apparatus for collecting information via network and recording medium recording program for implementing the method" (1999‑01‑12) — background.
• US 6,415,269 B1 — Bidcatcher, L.P., "Interactive remote auction bidding system" (2002‑07‑02) — remote‑bidding background.
• US 8,050,974 B2 — Amazon Technologies, "Method and system for price suggesting using item‑specific attributes" (2011‑11‑01) — pricing background.
• GB 0008931 D0 — Hewlett Packard Co, "Shopping assistance method and apparatus" — background.
• AU 2001258875 A1 — Bon‑Suk Koo, internet lottery‑auction — background.
• US 2002/0009978 A1 — Semyon Dukach, "Units for displaying information on vehicles" — background.

Non‑patent literature cited (selected)

• Hammouda et al., "Efficient Phrase‑Based Document Indexing for Web Document Clustering," IEEE Transactions on Knowledge and Data Engineering, Oct. 2004, vol. 16, no. 10, pp. 1279–1297 (*third‑party citation).
• Baumann, G. W., "Personal Optimized Decision/Transaction Program," IBM Technical Disclosure Bulletin (Jan. 1995), pp. 83–84.
• Clemons, E., "Evaluating the prospects for alternative electronic securities," Proc. ICIS 91 (Dec. 1991), pp. 53–61; Malone et al., "Electronic Markets and Electronic Hierarchies," Comm. ACM 14(25) (Jun. 1987); Resnick, "Reputation systems," Comm. ACM 43(12) (Dec. 2000).
• Prosecution/foreign‑office papers, e.g., "International Application Serial No. PCT/US10/036907, International Search Report and Written Opinion mailed Jul. 28, 2010" and "…International Preliminary Report on Patentability mailed Dec. 15, 2011" — these are the search opinions for this family and are useful for reading how the examiner distinguished the art.

Bottom‑line conclusions

1. No cited reference is a clean §102 anticipation of the granted independent claims 1/11/20. The distinguishing features are (a) selecting the next result by comparing the two hypothetical resulting ratios/exposure percentages against a target and choosing the smaller deviation (claims 2–3/12–13), and (b) the "desired exposure percentage" steering. No reference in the record performs that recursive per‑slot test.
2. The strongest single reference is US 2006/0287980 A1 (Microsoft, "Intelligent search results blending") — blending a mixed list by a determined ratio across non‑comparable sources. It is best characterized as §103 art against the granted claims and possible §102 art against the broader pre‑grant claim 1 (the version‑discrepancy flagged in the earlier section).
3. US 8,428,996 B2 (eBay/Grove) is the most relevant reference for the commerce‑dependent claims (4–6, 14–15) — auction/fixed‑price items commingled on a trading platform — but lacks the ratio/exposure ordering.
4. US 7,058,624 B2, US 7,853,574 B2, US 6,466,918 B1, and the third‑party references (US 2007/0255702 A1, US 2008/0046336 A1, US 2010/0070486 A1) are §103 background only.
5. The ~90 remaining citations are online‑auction/e‑commerce background art that contextualizes the trading‑platform limitations but does not reach the claimed ordering mechanism.

Confidence and gaps: The element‑by‑element anticipation calls above are my technical reading of the cited disclosures, not a legal opinion; a formal §102 chart would need the full prosecution history (to see which claims/references the examiner relied on) and the actual Office Actions — which I have not retrieved, so I flag that as an outstanding gap. I also did not exhaustively re‑verify the dates of every Tier‑3 member individually; those are taken from the patent's own citation table.

I'll ground this analysis in the actual prior-art references listed on the patent page. Let me pull details on the most relevant ones.

Obviousness Analysis — U.S. Patent No. 8,775,398 B2

35 U.S.C. § 103 | Built on the prior "Patent summary" and "Litigation summary" sections

This section assumes the claim text, bibliographic data, and version caveat already established in the Patent summary above, and it does not repeat them. The prior-art set analyzed is drawn from the "Citations (103)," "Family Cites Families (12)," and "Families Citing this family (4)" lists on the patent page (source: https://patents.google.com/patent/[US8775398](/patent/US8775398)/en).

1. Governing framework

• Statute. Because US 8,775,398 was filed June 1, 2009, it is governed by pre-AIA § 103(a). The Graham v. John Deere factors control: (1) scope/content of the prior art; (2) differences between the prior art and the claims; (3) level of ordinary skill; and (4) secondary considerations where of record.
• KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) supplies the combination rationale: a claimed combination is obvious where the elements were known, the combination is a predictable variation, and there was a design incentive or market pressure to make it. Id. at 416–21.
• No secondary considerations are of record. The patent page shows no reexamination, no IPR/PGR, and (per the Litigation summary above) no located assertion history. There is therefore no evidence of unexpected results, commercial success with nexus, or licensing that would rebut a prima facie case.

2. Decomposition of the claimed subject matter

Dependent claims add: the two-hypothetical-ratio comparison and choose-the-smaller-difference logic (claims 2, 3, 12, 13); online trading platform / auction / fixed-price context (claims 4–6, 14–15); revenue-based periodic adjustment of the target ratio (claims 7, 8, 16, 17); World Wide Web search (claims 9, 18); and the exposure-percentage recast (claims 10, 19).

The specification concedes (in the "Definitions"/Description passages on the page) that each format is scored by a different, non-comparable algorithm — which is precisely the problem the primary reference frames and solves.

3. Level of ordinary skill in the art (POSITA)

A bachelor's degree in computer science (or equivalent) plus 2–3 years of experience in information retrieval, web search ranking, or e-commerce search/listing systems — or a master's degree plus ~1 year. This person would be familiar with: (i) ranked retrieval over heterogeneous corpora; (ii) interleaving/blending of results from multiple sources; (iii) A/B testing and revenue metrics on commercial websites; and (iv) eBay-style listing systems that distinguish auction and fixed-price "formats."

4. The prior-art references and what they teach

4.1 PRIMARY — US 2006/0287980 A1, "Intelligent search results blending" (Microsoft; Liu, Ratnaparkhi, Yao, Katariya; filed 2005-06-21; published 2006-12-21)

URL: https://patents.google.com/patent/US20060287980A1

This reference, listed on the patent page under Family Cites Families (12), is a printed publication more than one year before the '398 priority date, and is therefore § 102(b) prior art. It discloses, almost element-for-element, the core of claim 1:

• A "search component that directs a query to at least two databases" and "a blending component that interleaves the results according to the rank" (claim 1; abstract).
• The incomparable-score problem the '398 patent identifies as its motivation: "each site may return a plurality of results, wherein the results are ranked according to different standards per the respective sites… it is difficult for users to determine the importance or relevance of returned information given the somewhat incompatible ranking standards" ([0005]).
• Ratio-driven interleaving, i.e., the "target ratio" concept: "a scoring or blending ratio can be determined and assigned to results from different databases… results returned from different sources can be automatically blended or mixed in display format according to the determined ratio or score. For instance… it may be determined that the results are 2 to 1 more likely than another database… in this case, the first two search results would be shown from database 1 followed by one result from database 2, followed by two results from database 1 and so forth" ([0007]).
• Express discussion of "1-1 interleaving" and a "weighted Interleaving strategy" — i.e., controlling the relative exposure of each source in the list.
• Claim 15 of the reference recites "determining a blending ratio for search results associated with the terms in the databases; and combining the search results in a output display according to the blending ratio."

The 2:1 example in [0007] is arithmetically a ~67%/33% exposure allocation implemented by a deterministic next-slot rule — i.e., "selecting a next item … based on a desired exposure percentage" (claim 1, element (d)).

4.2 SECONDARY — US 8,428,996 B2 (eBay; Grove et al.; priority 2001-06-11)

URL: https://patents.google.com/patent/US8428996

Listed under Families Citing this family (4). Filed 2001-12-14 and granted 2013-04-23, it is available as § 102(e) prior art as of its 2001 filing date (different inventive entity). It supplies the "format" dimension that Microsoft lacks:

• A network-based transaction facility that searches a database for items meeting a request and "generate[s] a commingled list of items offered for sale … wherein the commingled list includes items being offered for sale by any combination of an auction-type price setting process and a fixed price setting process" (claim 1).
• A sale-type field recording whether an item is auction or fixed-price ([0035]), and hybrid auction/fixed-price listings ([0038]).
• Explicit disclosure that such commingled listings can be searched by buyer search criteria and sorted ([0045]–[0053]).

This reference maps directly onto the '398 patent's auction / known fixed-price / unknown fixed-price "populations," the "on-line trading platform" limitation (claims 4–6, 14–15), and the "search request is related to an on-line trading platform" limitation.

4.3 US 7,058,624 B2, "System and method for optimizing search results" (Hewlett-Packard; Masters; filed 2001-06-20; granted 2006-06-06)

URL: https://patents.google.com/patent/US7058624

§ 102(b) prior art (published US 2002/0198875 A1 on 2002-12-26). Teaches controllable, weighting-based ordering of a results list:

• "associating a scaling factor with each of the established search criteria, calculating an overall matching score … and ordering the selection of identified documents based upon the calculated overall matching scores."
• The ability for a user to "adjust the relative weighting of various search criteria employed to order the results" and to "vary the relative weighting … in successive searches in order to optimize a search result."

This is directly relevant to claims 7–8 / 16–17 (periodically updating the ratio/weighting to optimize an outcome) — it supplies the "adjust the ordering parameter over time" teaching.

4.4 US 2010/0070486 A1, "Method, system, and apparatus for arranging content search results" (Venkata; filed 2008-09-12; published 2010-03-18)

URL: https://patents.google.com/patent/US20100070486A1

Filed before the '398 priority date, so available as § 102(e) prior art. Teaches ordering results from different domains ranked by different criteria by computing a common rank value and applying weights to ranking keys, and adjusting weights by "view definition." It corroborates that, by 2008, reconciling differently-scored result populations into one ordered list via weighted/ratio-based schemes was conventional.

4.5 Further secondary references (all listed on the patent page)

• US 2007/0255702 A1 (Orme, "Search Engine") — general search-engine ranking architecture; supplies the "search engine returns ranked results" element.
• US 7,853,574 B2 (IBM) — context-inferenced query generation and sorting of a query result.
• US 6,466,918 B1 (Amazon) — "exposing popular nodes within a browse tree" (popularity-driven presentation).
• US 7,340,429 B2 (eBay) — enabling a fixed-price purchase within an online auction environment (format-mixing on eBay).
• US 2011/0178868 A1 (Garg) and US 2010/0070486 (Venkata) — search-results-page enhancement / cross-source result arrangement.

5. Proposed grounds of rejection

Ground A — Claims 1, 11, 20 over US 2006/0287980 A1, alone or in view of the knowledge of a POSITA

Microsoft's search component + blending component maps to detector/engine/intermingler. The "blending ratio" maps to the "target ratio," and the 2:1 example (two from DB1, one from DB2, repeat) is a next-item selection that maintains a desired exposure split. The only gap is that Microsoft frames results as coming from databases rather than "formats." That gap is a labeling difference, not a technical one — a POSITA would recognize an auction database and a fixed-price database as "sets … in a first/second format" once placed in the eBay context (Ground B). Under KSR, "[a] person of ordinary skill is also a person of ordinary creativity."

Ground B — Claims 1, 4, 5, 6, 9, 11, 14, 15, 18, 20 over Microsoft + eBay 8,428,996

Substituting eBay's auction/fixed-price "sale type" populations for Microsoft's generic databases, and searching them per eBay's commingled-listing teaching, yields every element of the platform-context claims. Microsoft supplies interleaving by ratio; eBay supplies the two formats and the search-and-commingle framework.

Ground C — Claims 2, 3, 12, 13 over Microsoft + eBay + HP 7,058,624 (and/or Venkata)

The two-hypothetical-ratio "compare-the-difference" logic (claims 2–3) is a greedy implementation of Microsoft's ratio target: to place the next slot, compute the resulting ratio for each candidate source and pick the closer. Microsoft already performs ratio-directed placement; implementing it by evaluating both candidates and minimizing deviation is a routine engineering choice — the kind of "predictable variation" KSR holds obvious, and consistent with HP's "calculate … ordering … based upon the calculated scores" and Venkata's weighted-rank reconciliation.

Ground D — Claims 7, 8, 16, 17 over Microsoft + eBay 8,428,996 + HP 7,058,624 (feedback weighting), in further view of the ordinary knowledge of a POSITA in e-commerce

Claims 7–8 turn on (i) monitoring revenue for the two formats and (ii) adjusting the target ratio toward the higher-revenue format. The building blocks were known:

• HP '624 teaches adjusting the ordering weight in successive searches to "optimize a search result" — i.e., a closed feedback loop on an ordering parameter.
• An eBay platform (8,428,996) inherently measures completed sales/ GMV and revenue per listing type; running the presented mix as an experiment and shifting exposure toward the higher-revenue bucket is the ordinary commercial practice of A/B testing on a revenue-optimizing marketplace. The rationale is the "market pressure / design incentive" prong of KSR.

Ground E — Claims 9, 18 (WWW search)

Inherent in Microsoft, which blends "Internet Search sites"/search engines over "remote … across the Internet" databases ([0015]–[0017]).

Ground F — Claims 10, 19 (exposure-percentage recast)

Microsoft's express "blending ratio" and its 2:1 worked example directly disclose ordering to hit a desired percentage mix, which is the claim's "first/second new exposure percentage … target exposure percentage."

6. Motivation to combine (the KSR rationales)

1. Same field, analogous art. Microsoft and the '398 patent both address ordering a single list from differently-scored result populations — Microsoft's background section frames the identical problem statement ("incompatible ranking standards"). eBay '624/'996 are the natural application to listing search.
2. Both references point at the same improvement. Microsoft explicitly critiques plain 1-1 interleaving and proposes ratio/weighted interleaving; the '398 patent likewise rejects pure score-intermingling in favor of a ratio/target-exposure rule. A POSITA seeking to improve the eBay commingled listing would look to Microsoft's teaching.
3. Predictable combination. The combination uses each reference for what it discloses: eBay for what (auction vs. fixed-price populations), Microsoft for how (interleave by ratio), HP/Venkata for refinement (weighting and feedback).
4. Commercial incentive. An e-commerce operator has a standing incentive to tune the display mix to maximize revenue — the express goal of claims 7–8.

7. Where the patent might survive (arguments the patentee would raise)

1. Purposive ratio vs. relevance ratio. Microsoft's blending ratio is derived from relevance probability P(database | query); the '398 "target ratio" is a policy intended to guarantee exposure/visibility to each format. The patentee would argue the reference teaches away from an exposure-guarantee use. Rebuttal: nothing in Microsoft limits the ratio to relevance; its own claim 15 recites an abstract "blending ratio," and choosing a policy-based ratio is a design choice, not a patentable difference.
2. The specific greedy selector. Claims 2–3's hypothetical-ratio comparison could be pressed as a specific algorithm not literally disclosed. Rebuttal: it is a predictable, off-the-shelf greedy construction on Microsoft's disclosed ratio interleaving, and the granted independent claims do not require it (only the dependents do).
3. The revenue feedback loop. Claims 7–8's automatic revenue-driven ratio adjustment is the strongest non-obviousness redoubt. Rebuttal: HP '624 discloses the feedback-on-ordering-parameter concept, and revenue-optimizing experimentation was routine e-commerce practice.
4. Prosecution-history estoppel. Per the Patent summary above, the granted claims added the "respective position of each of the search result" and "desired exposure percentage" limitations absent from the pre-grant publication — indicating the applicant narrowed claims to traverse art. The reasons for that narrowing are unverified here and should be confirmed by pulling the file wrapper; those amendments may themselves supply arguments (and estoppel) that shape any future invalidity contest.

No secondary considerations appear of record to offset the prima facie case.

8. Bottom line and confidence

• Claims 1, 11, 20: Likely obvious over US 2006/0287980 A1 alone or in view of US 8,428,996 B2. The Microsoft reference independently teaches ratio-driven interleaving of differently-scored result populations — the heart of the independent claims.
• Claims 4–6, 9, 14–15, 18: Likely obvious with eBay 8,428,996 supplying the online-trading-platform / auction / fixed-price formats.
• Claims 2–3, 10, 12–13, 19: Likely obvious as predictable implementations/reframings of the ratio-interleaving teaching (optionally with HP 7,058,624 / Venkata 2010/0070486).
• Claims 7–8, 16–17: Obvious, but with the weakest margin. These require the revenue-feedback loop; the best support is HP '624 (adjust-the-weighting-to-optimize) plus the ordinary revenue-optimization knowledge of an eBay POSITA. This is the claim set I would expect a patentee to defend most vigorously.

Confidence: moderate-to-high for claims 1/11/20 and the format-context claims (the Microsoft reference is unusually on-point and is § 102(b) art); moderate for claims 2–3/10/19; moderate-to-lower for the revenue-loop claims 7–8/16–17, where the strongest reference (HP '624) teaches weighting adaptation but not revenue adaptation specifically.

Gaps / verification needed: (i) confirm the exact § 102 dates and claim text of US 8,428,996 (its 2001 filing date is critical to § 102(e) status as against a 2009 filing); (ii) pull the file wrapper for the '398 to see exactly what art was overcome and what the applicant argued (the version discrepancy flagged in the Patent summary implies substantive amendments); (iii) confirm no revenue-optimization reference beyond the page's citation list (e.g., a pay-for-placement or search-monetization reference) that would more squarely meet claims 7–8. I have not verified these independently and flag them as open items rather than settled conclusions.