US 7747745

I'll search for authoritative information on this patent number and any 2026 CAFC docket activity.

Let me run more targeted searches on the specific patent number and its litigation history.

One more targeted search for any 2026 CAFC activity specifically naming this patent.

US Patent 7,747,745 — Analyst Summary

Search scope note: I ran searches against public patent/litigation aggregators (Google Patents, Docket Alarm, CourtListener, PTACTS, CAFC opinions/orders pages) and the CAFC 2026 dockets. I found no CAFC docket in 2026 that names U.S. Patent No. 7,747,745 itself as the patent on appeal. Two 2026 AlmondNet appeals do exist, but they involve other patents in the same family (details below). I flag this explicitly rather than implying otherwise.

---

Bibliographic Data (authoritative, from the patent text)

Classification: G06Q 30/02 and subclasses (advertising/targeted ads), H04L 67/306 (user profiles), etc.

---

Abstract (verbatim)

"An automatic system facilitates selection of media properties on which to display an advertisement, responsive to a profile collected on a first media property, where a behavioral-targeting company calculates expected profit for an ad correlated with the profile and arranges for the visitor to be tagged with a tag readable by the selected media property. The profit can be calculated by deducting, from the revenues that are expected to be generated from an ad delivered based on the collected profile, at least the price of ad space at a media property where the BT company might like to deliver ads to the profiled visitor. When the calculated profit is positive (i.e., not a loss), the BT company arranges for the visitor to be tagged with a tag readable by the selected media property through which the BT company expects to profit."

---

Plain-Language Overview of the Independent Claims

Claim 1 — Method. Two steps performed with one or more computers:

• (a) Automatically determine a profile-attribute-dependent revenue available for showing a profile-attribute-related ad (i.e., how much money an ad tied to a given profile attribute is expected to bring in).
• (b) For each of many visitors to a first media property, upon receiving information about one or more profile attributes that may apply to that visitor, automatically electronically authorize a third-party second media property to allow display of a correlated ad to that visitor when the visitor later visits the second media property — but only at a price that does not exceed a price cap, where the price cap is a selected amount less than the revenue available for the ad.

In short: compute what an ad is worth for a profile, cap what you'll pay for the off-site ad space below that figure, and authorize/wire up the downstream media property accordingly.

Claim 24 — Tangible computer server device programmed to carry out the same (a) and (b) steps. This is the apparatus counterpart to claim 1.

Claim 37 — System comprising one or more computers programmed to implement the same (a) and (b) steps.

Notable dependent-claim coverage

• Claim 2/27: step (a) performed before step (b).
• Claim 3/25: the price cap is met by ensuring revenue ≥ (media price + an amount depending on cost or desired profit).
• Claim 4/26: (a) before (b); build a look-up data structure mapping profile attributes → second media properties; authorize responsive to an electronic redirect carrying profile-attribute data, only within a specified time period, via placing a readable tag.
• Claims 6–7: authorization triggered by a search request or by a visitor characteristic.
• Claim 8: second media property controls ad space only temporarily.
• Claim 9/36: translating revenue between two different payment methods (e.g., CPC vs. CPM).
• Claims 10, 15, 21–22: selecting media properties; delivering the price cap to an entity controlling ad space across multiple properties; Internet-site ad networks, including sites not owned by the second media property.
• Claims 11, 12, 18, 19: look-up structures; redirect-based authorization; redirecting the visitor to the second media property to place a tag, or placing the tag directly.
• Claims 13–14: recognizing a visitor action on the first property; auto-categorizing the profile attribute.
• Claims 20, 45: time-limited authorization (expiring tags).
• Claim 23/35/46: second media property in a different medium than the first (e.g., web → TV/mobile).

---

Litigation and PTAB / CAFC Context

District court (as reported on the patent page and in court records):

• AlmondNet, Inc. v. [Amazon.com, Inc.](/litigations/by-plaintiff/Amazon.com%2C%20Inc.), No. 6:22-cv-01204 (W.D. Tex.) — the '745 patent was one of the asserted patents (with 7,979,307; 8,204,783; 8,775,249; 8,959,146; 10,715,878). This action was dismissed with prejudice by joint stipulation dated April 26, 2023 — a stipulation that expressly preserved AlmondNet's assertions in the separate Amazon I case (6:21-cv-00898).
• AlmondNet, Inc. v. Meta Platforms, Inc., No. 6:22-cv-01205 (W.D. Tex.) — the other 2022 Texas case flagged on the patent page (later associated with N.D. Cal. 4:22-cv-08911).
• Samsung asserted-patent complaint referencing the '745 patent appears in IPR2024-00701 exhibit materials (Samsung v. AlmondNet).

PTAB: The '745 patent's prosecution history appears as an exhibit in Microsoft v. AlmondNet petitions (IPR2022-01319, IPR2022-01324), and in Docket Alarm records for IPR2022-01324. Uncertainty note: seeing the '745 prosecution history listed as an exhibit does not by itself establish that the '745 patent was the challenged patent in those IPRs — I could not confirm the challenged patent number in those two proceedings from the retrieved material.

Federal Circuit — 2026 activity (none names the '745 patent):

1. AlmondNet, Inc. v. Meta Platforms, Inc., No. 2024-1834 (appeal from PTAB IPR2022-01436). Decided April 13, 2026 — AFFIRMED under Fed. Cir. R. 36 (nonprecedential; Lourie, Dyk, Taranto). This IPR concerned the '582 patent (US 8,775,249), not the '745. (cafc.uscourts.gov judgment PDF; Justia)
2. AlmondNet, Inc. v. Amazon.com, Inc., Nos. 2025-1967 / 26-1326 (appeal from W.D. Tex. 6:21-cv-00898, Judge Albright). Dismissed by joint motion under Fed. R. App. P. 42(b) on April 6, 2026; no merits ruling. The judgment on appeal arose from the ~$122M/$136M verdict involving US 7,822,639 and US 8,671,139 — the latter being a continuation in the '745 family, but not the '745 patent itself. (CAFC order; GAE docket)

Bottom line on the CAFC question: As of the 2026 dockets reviewed, there is no Federal Circuit appeal docketed in 2026 specifically identifying US 7,747,745 as the patent on appeal. The 2026 AlmondNet appellate activity involves sibling/descendant patents ('249, '639, '139). I cannot rule out a docket I did not retrieve, but nothing in the material found supports a 2026 CAFC case on the '745 patent.

---

Family Context (relevant, from the patent record)

The '745 patent is the root of a large continuation family, all claiming the June 16, 2006 priority date, including: US 8,204,783; 8,200,822; 8,671,139; 8,959,146; 9,208,514; 9,508,089; 9,830,615; 10,134,054; 10,475,073; 10,839,423; 11,301,898; 11,610,226; 11,836,759; and pending/abandoned applications (e.g., US 2024/0281847 A1, abandoned). Because these are continuations of 11/763,286, their independent claims share the '745's price-cap/profit-based authorization architecture with varying claim language.

---

Uncertainty Flags

• No confirmed CAFC 2026 docket naming 7,747,745 — stated affirmatively only as to the dockets I could retrieve.
• PTAB proceeding numbers for the '745 patent itself are unconfirmed; I verified only that its prosecution history appears as an exhibit in Microsoft/AlmondNet IPRs (IPR2022-01319, IPR2022-01324). Some sources associate IPR2022-01324 with the '615 patent, which would make the '745 history a background exhibit rather than the challenged patent.
• Issue date, filing date, inventor, assignee, claim count, title, and abstract are taken from the authoritative patent text and are high confidence. Legal-status/expiration fields (e.g., "expires 2028-05-13") are Google Patents assumptions and should not be treated as legal conclusions.

I'll research litigation involving US 7,747,745 specifically. Let me run several targeted searches.

Let me dig into the specific cases and the docket numbers the patent record flags.

Answer — Litigation involving U.S. Patent No. 7,747,745

I searched litigation-focused sources (Stanford NPE Litigation Database, Docket Alarm, Justia/UniCourt PACER-derived dockets, Patexia, Google Patents' litigation metadata for US7747745, and PTAB petitions that reproduce district-court complaints). Note: I interpreted the ID literally as 7,747,745 / 7747745 and excluded same-family siblings (e.g., 8,204,783; 8,671,139; 8,959,146; 10,715,878), which are commonly confused with it.

Bottom line: I confirmed two district-court actions in which the '745 patent was expressly asserted, plus one closely parallel 2022 complaint I could not positively confirm. I found no CAFC appeal or PTAB/IPR instituted on the '745 claims themselves.

---

1. AlmondNet, Inc. et al. v. [Amazon.com, Inc.](/litigations/by-plaintiff/Amazon.com%2C%20Inc.) et al. — "Amazon II"

Sources: https://www.docketalarm.com/cases/Texas_Western_District_Court/6--22-cv-01204/AlmondNet_Inc._et_al_v._Amazon.com_Inc._et_al/docs/1/2.pdf ; https://docs.justia.com/cases/federal/district-courts/texas/txwdce/6:2022cv01204/1197266/36 ; https://www.docketalarm.com/cases/Texas_Western_District_Court/6--22-cv-01204/AlmondNet_Inc._et_al_v._Amazon.com_Inc._et_al/docs/34.pdf

⚠️ Caution on outcomes: AlmondNet's widely reported $122 million June 2024 W.D. Tex. jury verdict against Amazon concerned U.S. 7,822,639 and 8,671,139 — not the '745 patent. Do not attribute that verdict to 7,747,745. The '745-asserting case above was closed a year earlier.

2. AlmondNet, Inc. et al. v. Meta Platforms, Inc. — "Meta II"

Sources: https://npe.law.stanford.edu/case/269881 ; https://www.docketalarm.com/cases/PTAB/IPR2024-00701/Samsung_Electronics_Co._Ltd/docs/03-18-2024-Petitioner/Exhibit-1015-Exhibit_1015.pdf ; https://unicourt.com/case/pc-db5-almondnet-inc-et-al-v-meta-platforms-inc-1342872

3. AlmondNet, Inc. et al. v. Microsoft Corporation — 6:22-cv-01206 (W.D. Tex., filed Nov. 18, 2022)

Filed the same day, same venue, and same plaintiff group as the Amazon and Meta complaints above, and PTAB petitions in the Microsoft/AlmondNet IPRs (IPR2022-01319, IPR2022-01324) submit the prosecution history of U.S. Patent No. 7,747,745 as exhibits and cite the parallel Microsoft district-court case (6:21-cv-00897 / 6:22-cv-01206). I could not positively confirm from the sources retrieved that the '745 patent was named as an asserted patent in the Microsoft complaint. Treat this as unverified.

---

Related proceedings that mention the '745 patent (not '745 validity challenges)

• IPR2022-01319 (Microsoft Corp. v. AlmondNet/Intent IQ) and IPR2022-01324 — Exhibit 1005/1006/1010 is the "Prosecution History of U.S. Patent No. 7,747,745." These petitions appear directed at family members (e.g., the '783 patent), with the '745 prosecution history used as supporting evidence. Docket: https://www.docketalarm.com/cases/PTAB/IPR2022-01319/
• I found no IPR, PGR, CBM, or CAFC appeal instituted against the claims of 7,747,745 itself. Say so explicitly if you need a negative assertion: this is based on the sources searched, not a certified PTAB/CAFC docket check.

Explicit non-inclusions (to avoid confusion with the '745)

These AlmondNet campaigns assert other patents in the family and are frequently mis-attributed to the '745: Roku (6:21-cv-00731 / 6:21-cv-00876), Samsung (6:21-cv-00891; 5:22-cv-07515 / 3:22-cv-07515), Microsoft 2021 (6:21-cv-00897), Meta 2021 (6:21-cv-00896), Amazon 2021 (6:21-cv-00898), Yahoo/Oath (1:16-cv-01557; 1:19-cv-00247), Viant (1:23-cv-00174; IPR2025-01163), FreeWheel (1:23-cv-00220), LinkedIn (1:23-cv-01373), LiveIntent (1:24-cv-00831, D. Del.), Snap, and Adobe (7:25-cv-00487).

Confidence and caveats

• High confidence: the two confirmed cases (#1 and #2), their case numbers, filing dates, parties, venue, and the '745 assertions.
• Medium confidence: the N.D. Cal. procedural posture for the Meta case (transfer + stay); exact operative order dates after transfer should be verified on PACER for 4:23-cv-00438.
• Unverified: whether the Microsoft 6:22-cv-01206 complaint named the '745.
• Google Patents' own family metadata for US7747745 links only two WDTX cases as family litigation — 6:22-cv-01205 and 6:22-cv-01204 — consistent with findings #1 and #2. Source: https://patents.google.com/patent/US7747745/en
• I did not retrieve results specific to the '745 from Unified Patents' caselist or a PACER/CAFC docket directly in this session; the above rests on the secondary litigation databases cited. If you need certified docket currency, pull PACER for 6:22-cv-01204, 4:23-cv-00438 (N.D. Cal.), and the PTAB Patent Trial & Appeal Board docket for any '745-specific petition.

I'll check the structured block's default against live sources before writing anything up.

Initial results mention many AlmondNet PTAB matters — but on sibling patents, not necessarily the '745. Let me pin down whether IPRs were ever filed against the '745 itself.

One more check — CBMs from the Yahoo era, and a direct PTAB docket lookup for the '745 patent number.

PTAB / AIA Trial Landscape — US 7,747,745 (AlmondNet, Inc.)

Source of record: structured "PTAB proceedings on file" block (USPTO Open Data Portal ingest), which returns no AIA trial proceedings for this patent. I ran independent web searches (PTAB decisions, Docket Alarm PTAB dockets, Patexia PTAB dockets, Federal Circuit/CourtListener, and third-party petition exhibits) to test that default. Nothing surfaced that contradicts it.

Proceedings overview

Total AIA trial proceedings on US 7,747,745: 0 — 0 active, 0 with claims invalidated, 0 with claims sustained, 0 settled, 0 with institution denied, 0 with institution granted and 0 with any final written decision. The bottom-line defensive posture is therefore the opposite of "hardened": all 46 claims of the '745 are PTAB-untested, no claim has been canceled, and no § 315(e)(2) estoppel attaches to any party — but the patent has been through the PTAB meat-grinder only indirectly, as family lineage and as an exhibit in IPRs on its siblings, which means a defendant today gets no free invalidity finding and no free prior-art roadmap from the Board.

---

No AIA proceedings on file for US 7,747,745

Because the requirement is "don't invent proceeding numbers," the honest entry here is the null result, with the fields the task asks for answered as follows:

• Type: N/A — no IPR, PGR, or CBM was ever filed naming the '745 as the challenged patent.
• Filed: N/A.
• Status: No AIA trial proceeding on file (USPTO ODP), corroborated by independent search.
• Judge panel: N/A.
• Petition grounds: N/A.
• Institution decision: N/A.
• Final Written Decision: N/A. No claim-level verdict exists for any of claims 1–46.
• Settlement / termination: N/A.
• Appeal: N/A — there is no Board decision to appeal. (Note: 35 U.S.C. § 141/§ 144 CAFC jurisdiction runs from a Board decision; none exists here.)
• Defensive value: Neither you nor your adversary can point to a canceled claim. Conversely, there is no IPR record narrowing claim 1 (the "price cap… a selected amount less than the revenue available" claim), so claim scope and § 101 exposure for the '745 remain live district-court questions on a clean slate.

Evidence trail for the null result (rather than fabrication):

1. Google Patents' own prosecution/PTAB metadata for the patent shows litigation entries only — two W.D. Tex. cases (6:22-cv-01204, 6:22-cv-01205) — and no trial-proceeding entries: https://patents.google.com/patent/US7747745/en
2. Where the '745 number appears in PTAB records, it appears as family lineage or as an exhibit, not as the challenged patent. Examples:
   • IPR2022-01319 ([Microsoft Corp.](/litigations/by-plaintiff/Microsoft%20Corp.) v. AlmondNet, Inc.) is captioned U.S. Patent No. 8,671,139 — yet Petitioner filed "Prosecution History of U.S. Patent No. 7,747,745" as Exhibit 1006 (2022-07-27), and the Board denied institution (Paper 9, 2023-01-30): https://services.patexia.com/lawsuits/Microsoft-Corporation-v-AlmondNet-Inc-id-[179779](/patent/179779)/documents/sort_by/created/sort_dir/desc
   • IPR2022-01324 (Microsoft Corp. v. AlmondNet, Inc.) included the '745's prosecution history as Exhibit 1010; it was denied institution (Paper 9, 2023-03-13) and the fee was refunded (Paper 11, 2023-10-27).
   • In the '615 chain, the '745 appears only in a terminal disclaimer (Serial No. 15/360,762 disclaiming term beyond '783, '822 and '745) — i.e., as a family member, not a target.
3. The 2017 Yahoo! CBM wave targeted the '639, '586, '307, '249, '904, '582, '146, '139 and '574 — not the '745.

---

Family-context PTAB activity (NOT proceedings on the '745)

Flagging this only as pattern evidence and for the estoppel analysis. None of these involves US 7,747,745, and none of their outcomes changes a single claim of the '745.

Related Federal Circuit activity in the family: AlmondNet, Inc. v. Meta Platforms, Inc., No. 24-1834 (Fed. Cir. 2026-04-13, nonprecedential, per curiam (Lourie, Dyk, Taranto)) — AFFIRMED; an appeal from a PTAB IPR on a family patent (sibling of the '745, not the '745 itself). https://www.courtlistener.com/opinion/[10841919](/patent/10841919)/almondnet-inc-v-meta-platforms-inc/

Also worth knowing for context, though again not on the '745: in June 2024 a W.D. Tex. jury awarded AlmondNet $122 million against Amazon on the '639 and '139 patents (IAM, 2024-08-15). The '745 was pleaded in AlmondNet, Inc. v. Amazon.com, Inc., No. 6:22-cv-01204 (W.D. Tex.), which subsequent AlmondNet filings list as terminated; the companion Meta case, No. 6:22-cv-01205, was transferred/consolidated into N.D. Cal. 4:22-cv-08911.

---

Strategic summary

Claim status: everything is untested. The '745 issued with 46 claims — claims 1–23 (method; claim 1 independent, claims 2–23 dependent), claim 24 (tangible computer server device, claims 25–36 dependent), and claim 37 (system, claims 38–46 dependent). Because no AIA petition ever named the '745, there is no CANCELED / SUSTAINED split to report: all 46 claims are UNTESTED. Anyone telling you "the '745 claims are invalidated" is confusing the '745 with the '615 (claims 9–12 canceled, FWD 2023-11-28, affirmed at the Fed. Cir. 2026-04-13), the '878, the '398 or the '249. Note the practical corollary: where the '745 is part of the same patent family as the canceled '615 claims, a district court obviousness theory imported from the '615 FWD may still be available — but the '745's claim 1 "price cap" limitations are not the '615 claim 9–12 limitations, and the '615 FWD is not an invalidity judgment about the '745.

Estoppel landscape. There is no § 315(e)(2) estoppel against anyone on the '745, because § 315(e)(2) is triggered only by a final written decision in an IPR the party filed — and no IPR was ever filed on this patent. That cuts both ways:

• For you as a defendant: no petitioner has burned any prior-art ground on the '745, so the entire § 102/§ 103 field is open, with no IPR-estoppel shadow over your district-court invalidity case. Because the '745 has a 2006-06-16 priority date and a 2007 filing date, it is pre-AIA: PGR is categorically unavailable (PGR applies only to patents with effective filing dates on or after 2013-03-16), and the CBM program sunset on 2020-09-16, so IPR is the only AIA vehicle left. Ex parte reexamination under § 302 remains available to anyone at any time (no estoppel, lower cost, but the "substantial new question of patentability" threshold and no claim-construction-for-invalidity leverage).
• Against you: if you file an IPR, you alone bear § 315(e)(2) estoppel from that FWD. And check § 315(b) first — the one-year bar runs from service of a complaint alleging infringement of the patent. AlmondNet served Amazon and Meta with '745 complaints in the 2022 wave (No. 6:22-cv-01204 / 6:22-cv-01205), so those parties' clocks expired years ago; they are now time-barred from filing an IPR on the '745. If you were served on the '745 more than one year ago (or if your predecessor-in-interest was), you are barred too; if you are a new target, the clock starts when you are served. Also watch § 315(a)(1): a party that filed a declaratory judgment of invalidity on the '745 before petitioning cannot file an IPR (Roku's D. Del. DJ case, 1:21-cv-01035, is the family's precedent for that posture).

Pattern signals. AlmondNet is one of the most heavily IPR-challenged NPE-side portfolios of the last decade, and the pattern is instructive: (1) the 2017 Yahoo! CBM wave largely failed at the institution stage on § 18 CBM-eligibility, not on the merits; (2) the 2022–2023 IPR wave from operating companies (Microsoft, Meta, Amazon, Samsung, Roku) produced more denials than institutions, with the notable loss being the '615 (claims 9–12); (3) AlmondNet does appeal adverse decisions (Fed. Cir. 24-1834, affirmed 2026-04-13), so it is a litigant that goes the distance; (4) no defensive aggregator (Unified Patents) appears anywhere in this chain — the challenges have come from operating companies with their own commercial exposure, not from a subscriber-funded invalidity shop; and (5) across roughly two dozen IPR/CBM petitions against this family, none named the '745 — despite the '745 having been in suit against Amazon and Meta since 2022. Hold that fact loosely (my sweep was search-based, not an exhaustive PTAB docket query), but if it holds, the reasonable inference is that defendants evaluated the '745 and chose other targets, likely for cost/claim-coverage reasons rather than because the '745 is invulnerable.

Term pressure. The '745 carries an adjusted expiration of 2028-05-13 — under two years of life as of 2026-09-28. Any IPR would consume 12–18 months of that runway; if the patent is being asserted against you on a short demand letter or a narrow damages window, the economics of an IPR (typically mid-six figures all-in) versus a reexam or straight district-court invalidity case deserve an explicit analysis.

---

Recommended next steps

1. Do not concede that the '745 has PTAB history. Confirm the null result yourself in PTAB E2E / PTAB Center (https://ptacts.uspto.gov/) and the USPTO ODP API catalog (https://developer.uspto.gov/api-catalog/odp), searching by patent number rather than by party. If any '745 proceeding surfaces, it will be new since this ingest and I could not see it.
2. Get the service date on the '745 for your client. It drives § 315(b). If you are inside the one-year window, the petition must be filed before that date; if you are outside it, say so early, because it kills the option entirely.
3. If you are a defendant with an IPR window and a budget: the strongest candidates are pre-AIA § 103 combinations mapping (a) profile-based ad selection with (b) a price/cost comparator governing where the ad is placed. Because no petitioner has ever been estopped on this patent, there is no "reasonably could have raised" baggage — you get to pick the art freely. Serve the § 315(b) analysis first, institution-decision economics second.
4. If you are relying on a demand letter: quote the letter's claims back at the patent owner. If the letter cites claims 1, 24 or 37 (the independents) or any dependent claim, understand that those claims have never been invalidated or even reviewed by the Board. There is no canceled-claim short-circuit here; your invalidity case starts from scratch in district court or in a fresh IPR.
5. Pull the district-court record rather than the PTAB record. The '745's actual history is in W.D. Tex. No. 6:22-cv-01204 (listed as terminated) and the Meta case (6:22-cv-01205 → N.D. Cal. 4:22-cv-08911). Claim-construction positions on "price cap," "selected amount," and the pre-AIA "authorizing a third-party second media property" language will be your real leverage, and the FWDs on sibling patents (notably the '615, 2023-11-28) may supply useful admitted-art admissions even without estoppel effect.

Confidence flags, per the no-fabrication rule: (i) the count of zero is high-confidence on the ODP structured data plus search corroboration; (ii) the current status of the older CBM2017-00051 (instituted vs. denied) is genuinely conflicting between the sources I saw, so I declined to resolve it; (iii) the exact 2024–2025 status of IPR2025-00126 and IPR2025-01317 is unverified; and (iv) I did not locate the full docket of the Meta N.D. Cal. case, so its '745 disposition is unknown to me.

I'll research the assignment record for US 7,747,745 using multiple sources.

Let me search for the specific assignment records and any AlmondNet corporate structure changes.

Let me try to locate the specific reel/frame records and check for any security interests or liens.

Ownership Chain Analysis — US 7,747,745 B2

Access note / honesty flag up front: I could not retrieve USPTO Assignment Center records directly (the Assignment Center UI at https://assignmentcenter.uspto.gov/ and the legacy https://assignment.uspto.gov/patent/index.html index are JS-driven and my fetches did not return assignment rows). What follows is reconstructed from (a) the Google Patents legal-events block on the patent itself, (b) the assignment document text as reproduced verbatim inside the family's own file histories, and (c) court/agency filings and the assignee's own public statements. Where I cannot verify a reel/frame, I say so rather than invent one.

---

Inventors

Unusual-pattern check — the classic "inventors flee within 12 months" tell is NOT present. The opposite pattern obtains: the sole inventor is the founder/controlling principal of the assignee, has never left, and is the public face of the asserting entity 19 years after filing. Shkedi is the sole named inventor across essentially the entire 15-member family, which is characteristic of a founder-driven portfolio rather than an acquired/orphaned one.

Sources: Google Patents US7747745B2; GlobeNewswire, Apr 7 2026 — AlmondNet/Intent IQ/Datonics Amazon resolution; IAM interview with Shkedi.

---

Original assignee

AlmondNet, Inc. — a Delaware corporation, principal address Long Island City, NY (group address used throughout the family: 37-18 Northern Boulevard, Suite 404, Long Island City, NY 11101).

• Named on the issued patent: Yes — the front page carries "(73) Assignee: AlmondNet, Inc." and litigation filings state "AlmondNet owns all right, title, and interest in U.S. Patent No. 7,747,745."
• Did they ship a product embodying the claims? Partially / historically. AlmondNet was a 1998-vintage online advertising and data company (first online data marketplace; behavioral targeting; the company credits itself with the "Enhanced Notice"/AdChoices convention). Its current self-description is explicit that it is a licensing-and-R&D house: "the parent company AlmondNet Inc. … is focused on R&D and the licensing of its extensive portfolio of enabling technology and Intellectual Property." The group's operating subsidiaries are Intent IQ (identity resolution) and Datonics (audience data). There is no evidence in the record that AlmondNet, Inc. currently ships a product embodying the '745 claims.
• Current status: Operating — not dissolved, not in bankruptcy, not acquired. It is the parent of the AlmondNet Group and remains an active litigant/licensor (Amazon dispute resolved by license April 7, 2026).

Sources: US7747745B2; Joint Stipulation of Dismissal, W.D. Tex. 6:22-cv-01204 (Apr 26, 2023); GlobeNewswire Apr 7 2026.

---

Assignment timeline

Finding on record count: the USPTO record for this patent contains essentially ONE post-filing assignment — the original inventor-to-company assignment. There is no downstream chain.

I was unable to retrieve the reel/frame number for the recorded assignment; I flag that gap explicitly rather than guess. What IS verifiable is the assignment instrument text itself, reproduced in the file history of family continuation 14-EP-CON3 (filed June 7, 2012), which is an "Assignment For Published Patent Application" covering this exact title and the June 14, 2007 application.

---

2007-06-14 (executed/filed) / recorded 2007-06-14 — Reel/Frame not retrieved (see flag above). Google Patents legal events show a single event: "2007-06-14 — Assigned to ALMONDNET, INC. — ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: SHKEDI, ROY, MR."

• Conveyance: Assignment (Assignment For Published Patent Application — 37 CFR 3.11 recordation)
• Assignor: Roy Shkedi, the sole named inventor
• Assignee: AlmondNet, Inc., a Delaware corporation, "and any successor, assign, legal representative, or designated nominee"
• Correspondent / filer of record: Louis J. Hoffman of Louis J. Hoffman, P.C., 14301 North 87th Street, Suite 312, Scottsdale, AZ 85260 (USPTO Customer No. 26362); the assignment transmittal names "Louis J. Hoffman/Donald Hertz" as filer, with e-mail addresses louis@valuablepatents.com and donald@valuablepatents.com. (Flag: this is the address of record that appears on the family's filings; I am inferring it is also the Assignment Branch correspondent of record, which I could not separately confirm.) Recurrence flag: this firm is the sole correspondent across every single family record I retrieved — the 2007 original assignment, the 2010/2012 continuation filings (docket numbers 14-EP-CON, 14-EP-CON3, 14-EP-CON7, 14-TV1-CON2), and the later terminal-disclaimer and change-of-address filings.
• Context: Founder-to-harbor-company transfer at filing — ordinary start-up practice of taking assignment of the founder's invention into the operating corporation on the filing date. Not a fire-sale, not a transfer to an asserter.

Basis — the assignment instrument recites (verbatim excerpts): "FOR VALUE RECEIVED, I, the named inventor of the below-referenced patent application, sell, assign, and transfer to AlmondNet, Inc., a Delaware corporation … the entire right, title, and interest, for all countries, in and to certain inventions relating to: 'Media Properties Selection Method and System Based On Expected Profit from Profile-based Ad Delivery,' described in a U.S. utility patent application filed June 14, 2007, and a PCT application of the same date, and all the rights and privileges under all Letters Patent that may be granted for those inventions, including any regular applications and continuations, continuations-in-part, and divisions thereof…" Sources: assignment text reproduced in IPR2022-01319 Ex. 1004 (prosecution history of US 8,671,139); assignment text in 14-EP-CON3 filing materials.

No security agreement, license, merger, change-of-name, or release recorded against the '745 that I could find. Two non-assignment events worth separating out so they are not mistaken for ownership transfers:

• Terminal disclaimers (prosecution events, NOT assignments): the '745 is listed as a "prior patent" over which terminal disclaimers were filed by later family members — e.g., application 15/360,762 (filed Nov 23, 2016), owner listed as "ALMONDNET, INC. 100%." This confirms ownership but conveys nothing.
• Change of Correspondence Address, filed 2020-11-09, listing firm "INTENT IQ, LLC, 37-18 Northern Boulevard Suite 404, Long Island City, NY 11101" — but that filing is against sibling 8,677,398, not the '745. It does, however, evidence the shared group address/common principals across the family.

Notable NON-event: unlike several family siblings (8,677,398; 8,566,164; 8,595,069; 10,321,198; 10,715,878) which are held by Intent IQ, LLC, the '745 has never been moved to Intent IQ. Court AO-120 forms and complaints list ALMONDNET, INC. as the holder of the '745 in the Roku, Amazon, Meta, and Samsung matters. So within the group there has been some intra-group reallocation of patents — but not of this one.

---

Timeline diagram

timeline
    title Ownership of US 7747745
    2006 : Priority provisionals filed
    2007 : Shkedi assigns to AlmondNet Inc
         : Application 11-763286 filed
    2010 : Patent issued
    2016 : Family first asserted against Yahoo
    2022 : Asserted against Amazon and Meta
    2023 : Amazon 1204 suit dismissed
    2026 : Amazon dispute resolved by license

---

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. No assignment of the '745 to any "IP / Patents / Licensing / Holdings / Ventures" entity. The assignee since 2007-06-14 is AlmondNet, Inc., a Delaware operating corporation founded 1998, with a real New York office and two subsidiaries (Intent IQ, Datonics) — not a single-member registered-agent shell. The only recorded conveyance is the ordinary founder→corporation assignment.

2. Known asserter in the chain — PRESENT (qualified). AlmondNet, Inc. is not on the enumerated classic NPE lists (Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Round Rock, etc.). But it is surfaced as a high-frequency plaintiff by Unified Patents' litigation dataset, and it has maintained a sustained multi-front assertion campaign: AlmondNet v. Yahoo!/Oath, v. Roku, v. Facebook/Meta, v. Microsoft, v. Samsung, v. Amazon, v. LinkedIn, v. Viant, v. Oracle, v. Adobe. Its own counsel describes the entity as centered on "R&D and the licensing of its extensive portfolio." Cite: Unified Patents litigation entries on the patent page; the litigation roster in PTAB petition materials.

3. Repeat correspondent across the chain — PRESENT. Louis J. Hoffman, P.C. (Scottsdale, AZ; USPTO Customer No. 26362; filers "Louis J. Hoffman / Donald Hertz," louis@valuablepatents.com, donald@valuablepatents.com) is the single firm of record on every family link I could inspect: the 2007 original assignment, the 2010 and 2012 continuations (attorney dockets 14-EP-CON, 14-EP-CON3, 14-EP-CON7), and the TV-side family (14-TV1-CON2). Per the task's own caution, a single appearance is not a finding — but this is recurrence across the entire 15-member family over 10+ years, which is a genuine concentration signal (one lawyer/firm curating the whole portfolio chain-of-title and prosecution). No separate "assignment-only" correspondent appears, so I cannot point to a distinct NPE-recording attorney laundering reels — which itself weakens the classic troll tell.

4. Cascading transfers — NOT PRESENT. There are no consecutive assignments through chained LLCs. There is exactly one conveyance in 19 years, from the inventor to his own corporation. Nothing in <24 months, nothing sharing a correspondent address across unrelated assignee names.

5. Pre-litigation transfer — NOT PRESENT. The sole assignment is dated 2007-06-14. The earliest family assertion activity is ~2016 (AlmondNet v. Yahoo!, E.D.N.Y. 1:16-cv-01557) and the '745 itself was asserted in the 2022 W.D. Tex. filings (6:22-cv-01204/01205/01206). That is ~9–15 years between assignment and suit — the opposite of a venue/standing-setup transfer.

6. Bankruptcy fire-sale — NOT PRESENT. AlmondNet never filed Chapter 7/11. Shkedi has publicly described near-financial-ruin around 2013 resolved by a Google licensing deal ("take-it-or-leave-it"), not a bankruptcy sale. No bankruptcy-court docket, no trustee conveyance.

7. Privateering — NOT PRESENT. There is no operating-company-to-NPE transfer for AlmondNet to hide behind; AlmondNet is the original operating company and asserts in its own name. (The inverse arrangement — third parties licensing from AlmondNet — is documented, e.g., Google, Microsoft, and Amazon licenses, but that is licensing-to-alleged-infringers, not privateering.)

8. Defensive aggregator — NOT PRESENT. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN. It terminates at the original, still-asserting assignee.

---

Verdict

NPE — moderate confidence.

Justification: The ownership chain itself is conspicuously clean — a single 2007-06-14 founder-to-corporation assignment to AlmondNet, Inc. (Delaware) that has never been followed by any shell-LLC transfer, cascading conveyance, pre-litigation transfer, or bankruptcy sale, so signal 1, 4, 5, and 6 are all absent. The NPE character therefore does not come from chain maneuvering; it comes from the assignee's own posture — AlmondNet, Inc. is a high-frequency plaintiff surfaced in Unified Patents' litigation data and self-described as focused on "R&D and the licensing" rather than on shipping a product embodying the '745 claims (signal 2, present), with a single law firm (Louis J. Hoffman, P.C., Customer No. 26362) as recurring correspondent across every family record (signal 3, present, but weak in a one-assignment chain). One strong signal plus one weak signal lands at moderate, not high — and I explicitly decline to call the low-confidence "high" tier because the hallmark shell-entity and cascading-transfer indicia are affirmatively absent and the entity is the inventor's own organically-formed company, not an assembled assertion vehicle.

Alternative reading I considered and rejected: "Operating-company assertion" — AlmondNet does sue genuine competitors and its group operates real subsidiaries. Rejected because the record does not establish that AlmondNet, Inc. ships a product embodying the '745 claims; the group is licensing-centric, which is the classic non-practicing posture.

Verify at: USPTO Assignment Center / legacy Assignment Search — search patent number 7747745 and application 11/763,286. Please note: I could not extract a reel/frame from any accessible source; if the Assignment Center returns an original-assignment reel/frame for the 2007-06-14 Shkedi→AlmondNet, Inc. conveyance, that will complete this chain, and I expect it to be the only row.

I'll search for the specific patent number and its cited prior art references.

I have the patent identity confirmed. Now let me pull the actual "References Cited" (front-page) prior-art list for the '745 patent.

Let me pull the Justia page and try targeted queries for the front-page cited references of the '745.

Preliminary flags (read first)

1. Date contradiction. The task header states "Current Date: April 26, 2026," but the system date in this session is 2026-09-28. I am not silently reconciling these; flagging per the rule about contradictions. All statements below are as of 2026-09-28 unless quoted from a dated source.

2. Database-access caveat. I do not have direct query access to USPTO Patent Center / Patent Public Search (ppubs) in this session — my tools are general web search. So I cannot truthfully say "the USPTO database shows X." What follows distinguishes clearly between (a) what I retrieved and (b) what I could not retrieve.

3. The critical gap — I could not retrieve the '745's own front-page "(56) References Cited" list. The authoritative patent text supplied for this analysis (Google Patents' US7747745B2 page, fetched 2026-09-28) rendered without the (56) References Cited / PTO‑892 section. My targeted follow-up searches for the '745's examiner citations did not surface that list either. I am therefore not going to invent a per-reference citation list and present it as the '745's. Below I give what I did verify and label it precisely.

---

Patent identity (literal ID check — passed)

The identifier 7747745 / US 7,747,745 B2 resolves to exactly one patent, and I am not substituting any similar number:

This matches the previously generated sections; no contradiction with them.

---

What the citation search actually returned (and why it matters)

Three distinct citation streams exist around this patent number, and conflating them is the main error risk:

(A) Backward citations (prior art cited in the '745) — the thing you asked for. Not retrieved. See §"What I could not confirm" below.

(B) Forward citations (later patents citing the '745) — heavily retrieved. The '745 appears in the "References Cited / U.S. Patent Documents" lists of many later AlmondNet/Shkedi patents, e.g. US 10,009,636; US 10,405,058; US 11,003,970; US 11,503,348; and the Unified Patents art pages for US 9,508,089 and US 9,830,615 list US‑7747745‑B2 (2010‑06‑29) among their cited references. Sources: https://patents.justia.com/patent/10405058#7, https://patents.justia.com/patent/10009636#3, https://patentimages.storage.googleapis.com/3a/d1/81/e82d3946060c7d/US11093970.pdf#7#2, https://portal.unifiedpatents.com/patents/patent/9508089#1. These are not prior art against the '745 (they post-date it) and cannot anticipate it.

(C) Family-level IDS/Form PTO‑1449 citation universe — retrieved, and the closest proxy available. Because the '745 and its continuations share one specification, the family's Information Disclosure Statements list largely the same pre‑2006 references. I retrieved these lists from the '745's own prosecution-history exhibits and from sibling filings:

• IPR2022‑01319, Microsoft v. AlmondNet — Exhibit 1006 / 1005: "Prosecution History of U.S. Patent No. 7,747,745" (Docket Alarm: https://www.docketalarm.com/cases/PTAB/IPR2022-01319/Microsoft_Corporation/docs/07-27-2022-Petitioner/Exhibit-1005-5-Prosecution_History_of_US_Patent_No_8,204,783.pdf; Patexia confirms "Prosecution History of U.S. Patent No. 7,747,745" as Ex. 1006).
• IPR2022‑01324, Microsoft — Exhibit 1010: "Prosecution History of US Patent No. 7,747,745."
• A Form PTO‑1449 (Mod. 04‑03) listing for Shkedi app. 15/893,354 (Docket Alarm .../EXHIBIT-1013...) and the '898 Amazon complaint front page (.../6--21-cv-00898/.../docs/1/19.pdf) reproduce the same reference set.

These are the documents that contain the definitive answer. I did not open them in full, so the list below is a family IDS citation set, not a confirmed '745 PTO‑892 examiner-citation list. Treat every item as "likely relevant to the '745; confirm against the '745 file wrapper."

---

Candidate prior art (family IDS citation set) with §102 assessment

The short answer on §102

None of these references anticipates any claim of the '745 under 35 U.S.C. § 102. This is not a judgment call I'm guessing at: the '745 issued with 46 claims across three independent claims (1, 24, 37), and every claim in the set requires the same core limitation chain as claim 1 — (a) determining a profile-attribute-dependent revenue for an ad, and (b) authorizing a third-party second media property to display a correlated ad "for a price that does not exceed a price cap that is a selected amount less than the revenue available for display of the advertisement." The 1990s art cited in the family discloses profile/ad-targeting, cookie/tag persistence, and ad serving — but it predates and does not disclose the revenue-derived price-cap authorization to a third-party property. With one element missing from each reference, these are at most § 103 (obviousness) background, not § 102 anticipatory art. Any of these can only anticipate if a reference independently discloses all of: profile-attribute-dependent revenue determination and the capped-price authorization to a separate media property.

The references (as they appear in the family IDS lists)

Provenance caution: dates and short subjects above come from the family IDS/1449 listings and companion art-index pages I retrieved; where a listing gave only a number/inventor I have described the item's general subject rather than asserting a verbatim official title. Do not treat these titles as verified official titles without checking the printed front pages. Items I could not individually re-verify are labeled by their IDS grouping.

Which cited references are "most relevant" (technical, not §102)

If the question is which cited art is closest, rank by overlap with the '745's two inventive elements:

1. US 5,948,061 (Merriman) — closest on the ad-serving/selection mechanism.
2. US 5,933,811 (Angles) and US 5,848,396 (Gerace) — closest on profile-based ad targeting.
3. US 5,794,210 / 5,855,008 (Goldhaber) — closest on the advertising-economics framing, which is the '745's twist.

Even in combination, these supply targeting + serving + ad economics, but not the claimed linkage of "revenue available for the ad" → "price cap that is a selected amount less than that revenue" → "authorizing a third-party second media property." That gap is why the claims survived prosecution, and it is the element to attack (or defend) in any validity challenge.

---

What I could not confirm (explicit negative assertions)

• I did not obtain the '745 patent's own PTO‑892 "Notice of References Cited," so I cannot state which references the examiner cited (as opposed to those submitted by the applicant via IDS). This distinction matters: applicant-submitted IDS references carry no examiner-adopted §102/§103 characterization.
• I found no PTO‑892 or examiner's citation list for the '745 itself in the material retrieved, and no § 102 rejection of record for the '745 claims.
• No IPR/PGR/CBM instituted on the '745 claims and no Federal Circuit 2026 docket naming 7,747,745 — consistent with the litigation section already generated. No contradiction there.

To close the gap definitively, pull these three documents

1. USPTO Patent Center, app. 11/763,286 — the "References Cited" / PTO‑892 and any examiner's search notes (definitive backward-citation list).
2. IPR2022‑01319, Exhibit 1006 — "Prosecution History of U.S. Patent No. 7,747,745."
3. IPR2022‑01324, Exhibit 1010 — "Prosecution History of US Patent No. 7,747,745."

Any of these three should yield the authoritative per-reference list; items 2–3 are downloadable from the Docket Alarm PTAB case pages cited above.

Bottom line: The identifier 7747745 = US 7,747,745 B2 (Shkedi / AlmondNet) is confirmed. I could not retrieve the '745's own front-page prior-art list, so I have reported the family-level IDS citation set with that limitation stated plainly rather than presenting it as the '745's examiner citations. On the record available, no cited reference anticipates any claim under § 102 — each lacks the revenue-derived price-cap authorization to a third-party second media property that defines independent claims 1, 24, and 37.

I'll ground this in the actual cited prior art on the '745 patent's record rather than working from memory. Let me pull the "References Cited" data and check what the Prior Art/Similar Documents sections list.

The patent page's prior-art citation list didn't come through in the fetch. Let me pin down the specific references actually cited/relevant to this family, since those are the ones I need to build §103 combinations around.

Obviousness Analysis — U.S. Patent No. 7,747,745 B2

35 U.S.C. § 103 (pre-AIA), priority date June 16, 2006

---

0. Two threshold flags before the analysis

(a) Date contradiction. This task header states the current date is April 26, 2026; the session system date and the fetched patent record are dated 2026-09-28. I have not silently reconciled these. The analysis below is date-independent (it turns on art predating June 2006), but the discrepancy is noted per the "flag contradictions" rule.

(b) Evidence base for the "Prior Art section." The Google Patents page as I retrieved it renders the "Prior art keywords" field (visitor; profile; authorizing; media; media property) and the Background of the Invention, but its structured "References Cited" list did not render in the fetch. I therefore build the combinations on two independently verifiable sources, and I label which is which:

1. Applicant-admitted prior art inside the '745 specification itself (the Background's discussion of DoubleClick "Boomerang," BT-company cookie placement, and cookie matching). These are applicant admissions usable as prior art under § 102/MPEP 2129 and are, if anything, stronger than an examiner citation because they are the patentee's own characterization.
2. References actually applied against this specification by USPTO examiners and PTAB petitioners in the '745's continuation family (IPR2022‑01319 record and the 13/491,454 / '139 prosecution history): Merriman (US 5,948,061); Haveliwala (US 8,321,278 / US 2005/0222989 A1); Boubek; Bandura (US 2002/0099603 A1); plus Roth, Jaye, Minor, Coleman, Herz, Kublickis, and Jaschke.

(c) Caveat on reference numbers. I can verify US 5,948,061 (Merriman) and US 8,321,278 / US 2005/0222989 A1 (Haveliwala) with high confidence. I could not verify the patent numbers for Roth, Jaye, Minor, Coleman, Boubek, Herz, Kublickis, or Jaschke in this session. Where I name them, I name them as the family record names them and do not assert a number I have not confirmed. Full confidence tags are in § 7.

---

1. Governing law and level of ordinary skill

The application was filed June 14, 2007 with a June 16, 2006 priority date. Pre-AIA § 103(a) governs. The controlling framework is Graham v. John Deere as refined by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) — in particular KSR's holdings that (i) a combination of known elements yielding predictable results is obvious, (ii) a "design incentive" or "market pressure" can supply the motivation, and (iii) a known technique applied to a known device is obvious even without an express teaching in the references. MPEP 2144.04 (design incentives/market forces are "predictable variations") and 2144.05 (obvious to try in a finite set of identified, predictable solutions) are directly on point here.

POSITA: a person with a bachelor's degree in computer science or electrical engineering (or equivalent), plus roughly 2–4 years' experience with client-server systems and online ad serving — familiar with HTTP cookies and domain-scoping, server-to-server redirects, ad networks, CPM/CPC pricing, and bid/auction ad models. Notably, the record in this family itself treats Boomerang-style redirect/cookie architecture and profile-based bidding as the ordinary baseline, which supports a relatively low skill-gap for the asserted combination.

---

2. The operative legal question

The prior art (as detailed below) comfortably supplies: profiling a visitor at a first media property; tagging the visitor so a second property can recognize them later; correlating an ad to the profile; and determining a profile-attribute-dependent price that an advertiser will pay. The only limitations that carry any real § 103 weight are the two in claim 1(b):

1. the authorization is capped by a price cap that is a selected amount less than the revenue available (i.e., a media-buy ceiling, not just an advertiser bid), and
2. the actor "authorizes" a third-party second media property at that price, rather than merely serving the winning ad.

Both are addressed head-on by the combinations in § 3.

---

3. Combinations rendering the claims obvious

Ground 1 — Merriman + Haveliwala (primary combination)

Targets: independent claims 1, 24, 37 — and, with the sub-references below, essentially the whole claim set.

The gap and why it closes. Neither reference labels a media-cost ceiling a "price cap." That is a distinction without a difference under KSR: the '745 specification itself defines the test as Pr = Rev(profile) − P(mp) − C − Mar > 0, i.e., P(mp) < Rev(profile) − C − Mar (see the '745's own equations (4)–(9)). A POSITA implementing an ad-buying system on Haveliwala's profile-value machinery would necessarily (i) know the media property's ask, P(mp), and (ii) decline to authorize placement where it exceeds expected revenue less cost/margin. That is a predictable variation of a known commercial parameter, not an inventive step. The patents of the same inventor applied in the family's own IPRs likewise frame the problem as "the cost of ad space at a second media property might not be covered by revenues" — a business-economics problem, and the Federal Circuit treats such profit-constraint coding as routine when the data and comparison are known. KSR, 550 U.S. at 417–18; MPEP 2144.04.

Ground 2 — Haveliwala + Boubek

Targets: claims 4, 17, 20, 26, 33, 34, 45 (specified time period) and the price-cap/spending-limit aspect of claims 1/3/24/25/37.

This is not my invention — this is precisely the combination the examiner used in the 13/491,454 ('139) prosecution, and it is documented on the record: "Each information tag may be associated with a spending limit and based upon predetermined time period" (Boubek, Abstract, FIG.). The examiner's stated motivation is quotable and squarely on point: combining them would "result[] in a system that allowed … targeting of specific markets … during set times as deemed most effective for receiving a positive margin at a faster return and maximize the efficiency of advertising campaigns." (Examiner's Office Action, 13/491,454.)

Because the '745 and the '139 share the identical specification, Boubek's spending-limit/predetermined-period teaching is equally applicable to the '745's dependent time-period and cap limitations. A "spending limit" is the functional equivalent of a price cap applied to an authorized purchase — close enough that a POSITA would see the substitution as obvious, and the motivation (positive margin, timely return) is expressly economic and expressly stated.

Ground 3 — Haveliwala + Bandura

Targets: the tag-placement / redirect-dependent claims (12, 18, 19, 29, 40, 43, 44).
Bandura (US 2002/0099603 A1) was applied by the examiner in the same family prosecution against exactly the tag-placement and redirect-based authorizing sub-claims. Treat this ground as medium confidence — I did not retrieve Bandura's disclosure text in this session and cannot independently characterize it beyond the record's use of it.

Ground 4 — Haveliwala + Herz

Targets: the same dependent tier.
Grounds designated "Merriman–Minor, Roth–Jaye, Roth–Minor" and Grounds I–L (Herz) appear in IPR2022‑01319, and the PTAB record notes Patent Owner did not dispute the Herz grounds. I flag this as a procedural data point only: non-opposition is not an adjudication of obviousness, and Herz's own disclosure is unverified by me.

Ground 5 — The specification's own admission + ordinary ad economics ("clean hands" ground)

Targets: independent claims 1, 24, 37, plus claims 6, 7, 12, 13, 18, 19, 21, 22.

The '745 Background admits, as prior art:

• Boomerang: cookies placed at an advertiser site so that the same visitor is found "on other sites where DoubleClick is the ad server," with delivery "following a redirect."
• BT-company practice: cookies placed on a publisher's visitors "for the purpose of delivering targeted ads to the publisher's visitors on other sites."
• Cookie matching: sites/ad networks placing their own cookies on visitors redirected from a BT company, so they can later "call upon the BT company to serve an ad."
• The stated problem: "the cost of ad space at a second media property might not be covered by revenues generated for a BT company buying the space."

That last admission is close to dispositive of the claim's remaining novelty: the patentee has admitted that (i) the tagging/redirect/retargeting architecture was known, (ii) profile value varies by attribute, (iii) media prices vary, and (iv) the resulting risk is a loss on the media buy. Once the loss condition is admitted as the problem, the claimed solution — compare expected revenue against media price and only proceed when positive — is a result-oriented, predictable business rule implemented on known architecture. Where an applicant frames the invention as a profitability rule over admitted mechanics, § 103 is satisfied by the admitted mechanics plus a reference (Haveliwala) supplying profile-dependent pricing.

---

4. Motivation to combine (the KSR / Graham factors)

1. Same field of endeavor / analogous art (MPEP 2141.01(a)). All references are networked advertising, profile-based targeting, or ad inventory pricing. There is no field-of-invention gap to argue.
2. Common problem, common solution. Merriman, Haveliwala, and the admitted Boomerang art all address "find the profiled user again elsewhere and serve a relevant ad." The combination is additive, not transformative.
3. Market pressure and recognized inefficiency. The '745's own Background enumerates the inefficiency the claim is meant to solve (variable media prices; variable profile values; unsatisfactory margins after ad-serving and sales costs). KSR holds that such "market demand" and "design incentives" supply motivation to combine.
4. No change in principle of operation. Adding a profitability gate to Haveliwala's profile-priced ad server changes which placements are made, not how the system operates — the classic "predictable result" scenario.
5. Finite, pre-identified solutions. Given profile value V and media price P, the options are: buy (P<V), don't buy (P≥V), or buy with a margin buffer. That is a closed set of predictable responses — KSR's "obvious to try."
6. Express examiner motivation exists for the key sub-limitation. For the time-period/spending-limit dependent claims, the record contains a written, adopted motivation for Haveliwala + Boubek (see Ground 2), which defeats a "no articulated reason to combine" defense there.

---

5. Anticipated rebuttals to an obviousness case (and how they fare)

---

6. Claim-by-claim summary map (condensed)

---

7. Confidence and uncertainty ledger

• High confidence: Merriman US 5,948,061 content and identity; Haveliwala US 8,321,278 / US 2005/0222989 A1 content (third-party-website embodiment; advertiser bidding on user profiles; CTR/performance-adjusted "effective price"; category-based profiles); the examiner's Haveliwala + Boubek reasoning as quoted in the 13/491,454 record; the '745 specification's own admitted Boomerang/BT-company prior art and its profit equations; that pre-AIA § 103 and KSR govern.
• Medium confidence: Boubek's and Bandura's full disclosures (I have the examiner's characterization, not the references); Herz's disclosure; the Merriman–Jaye / Roth–Jaye motivation rationale as reported in IPR2022‑01319 papers (I have Petitioner's reply, which is advocacy, not a Board finding).
• Explicitly unverified: patent numbers for Roth, Jaye, Minor, Coleman, Boubek, Herz, Kublickis, Jaschke; the Eldering-type reference for the different-medium claims; and the '745's own structured "References Cited" list, which did not render in my retrieval. Per the operating rules, I do not supply numbers I could not confirm.
• Cross-check note (possible prior-section inconsistency, flagged not corrected): the earlier "Patent summary" and "Litigation summary" sections list slightly different asserted-patent sets for AlmondNet v. Amazon, No. 6:22-cv-01204 (the former includes 8,775,249 / 8,959,146 / 10,715,878; the latter adds 7,861,260 and 7,979,307). This does not affect the obviousness analysis but should be reconciled against the complaint if the litigation posture matters.
• Legal-status caveat carries forward: "Active / expires 2028-05-13" remains a Google Patents assumption, not a legal conclusion.

---

Bottom line

On the art available on this record, claim 1 (and its apparatus/system counterparts 24 and 37) is a strong § 103 candidate over Merriman + Haveliwala, with the media-price ceiling supplied either by Boubek's spending-limit teaching or by the '745's own admission that the problem is unprofitable ad space. The time-limited, tag-based dependent claims (4, 17, 20, 26, 33, 34, 45) are the most vulnerable because the examiner's Haveliwala + Boubek combination is already on paper with an express margin-based motivation. The categorization claims (7, 14, 30, 42), the CPC/CPM translation claims (9, 36), and the look-up-structure claims (11, 28, 39) are also well met by Haveliwala alone. The genuinely contestable limitations are the "third-party second media property" authorization framing and the different-medium claims (23, 35, 46) — the latter only because I could not verify a pre-2006 addressable-TV reference in this session.