US 12457388

I'll search for authoritative information on this patent number, including the litigation data mentioned in the record.

Let me dig into the claim text and check for any Federal Circuit activity specific to this patent.

US 12,457,388 B2 — Analyst Summary

Sources: the authoritative patent text you supplied (Google Patents, patent/US12457388B2/en), plus USPTO/PTAB-adjacent docket records surfaced in search (Stanford NPE Litigation Database, Justia Dockets, UniCourt, Ex Parte complaint analysis). No Federal Circuit 2026 docket naming this patent was found (see below).

1. Bibliographic data (from the authoritative patent record)

Continuation chain (verbatim from the record):
18/800,005 ← cont. of 18/384,835 (filed 2023-10-27, now US 12,063,421) ← cont. of 17/401,078 (2021-08-12, US 11,831,964) ← cont. of 16/435,051 (2019-06-07, US 11,095,952) ← division of 14/089,645 (2013-11-25, US 10,321,198) ← cont. of 12/981,925 (2010-12-30, US 8,595,069) ← cont. of 11/968,117 (2007-12-31, US 8,566,164). Also related to Ser. No. 11/736,544 (filed 2007-04-17), incorporated by reference.

2. Abstract (verbatim, condensed)

In a computer-implemented method, a selected online advertisement is automatically directed to, or online activity is automatically tracked and recorded from, an online user interface device which corresponds, at the time the advertisement is directed or the activity is tracked, to a first online access identifier. The advertisement is directed or the activity is tracked based on information from a user profile that references or includes a first set-top box identifier. That information is derived at least in part from automatically collected data related to user behavior with respect to at least one television advertisement delivered to a set-top box, which set-top box corresponds, at the time of that user behavior, to the first set-top box identifier. The first online access identifier is associated with the first set-top box identifier, by being referenced or included in the user profile, by being associated in a database, or otherwise.

3. Plain-language overview of the independent claim(s)

⚠️ Important limitation on this section: The authoritative text you supplied contains the abstract, summary, definitions, and description, but the literal claims section was not included in the fetched document (the text ends mid-"FIG. 1" description). I therefore give the overview below from the patent's own Summary/Abstract language and from the complaint-analysis description of asserted claim 1. Treat the claim numbering and exact wording as unverified. Google Patents/Justia can supply the literal claim set.

The patent is a cross-media ad tracking/targeting invention in the Intent IQ / AlmondNet family. It addresses the known technical problem that dynamic (DHCP-assigned) IP addresses make it hard to tie a television device to an online device over time, especially without collecting personally identifiable information (PII).

Independent claim 1 (method) — plain language, per the asserted-claim description:

1. Receive a notification at a computer system, where the notification includes or references a first set-top box (STB) identifier, and the notification results from and signifies that a television advertisement was presented using the STB corresponding to that identifier.
2. Using that notification, automatically cause a first action with respect to an online user interface device (e.g., a computer/phone) — namely directing a selected online ad to it, or tracking/recording the user's online activity from it — where the online device corresponds, at the time of the action, to a first online access identifier.
3. The first online access identifier and the first STB identifier are associated — either by being referenced/included in the same user profile, by being associated in a database, or otherwise. (In the parent '198 patent this association is expressly made "without using personally identifiable information"; the '388 abstract states the association element without reciting that PII limitation, so I cannot confirm from the supplied text whether the PII-negation language carries into '388 claim 1.)

Likely companion independent claims. The '388 family pattern (method / system / article-of-manufacture or computer-readable medium) strongly suggests additional independent claims of a system and a non-transitory computer-readable medium performing the same steps — e.g., the sibling '069 patent lists independent claims 18 (system) and 25 (article). I did not verify the exact independent claim numbers or text for '388.

Notable dependent-claim themes described in the spec: linking the online ad/tracked activity to the delivered TV ad in a database; STB identifier as a tag/cookie or STB IP address; online access identifier as an IP address or cookie; observed behavior including user commands during the ad (skip, fast-forward, mute, change channel); associating identifiers via a provider, a common IP address (shared modem/LAN), or a database of STB identifiers + IP addresses + timestamps; and mobile online devices that intermittently leave the home connection.

4. Enforcement status and 2026 CAFC check

District court litigation (found):

• Intent IQ, LLC v. Samba TV, Inc., No. 1:25-cv-01084-MN, D. Del., filed 2025-08-28, Judge Maryellen Noreika (Patent, 35 U.S.C. § 271).
  • Stanford NPE Litigation Database lists US 12,457,388 among the patents-in-suit in this case (alongside 10,321,198; 11,949,962; 8,677,398): https://npe.law.stanford.edu/patent/[12457388](/patent/12457388) and https://npe.law.stanford.edu/case/[307865](/patent/307865)
  • Docket: https://dockets.justia.com/docket/delaware/dedce/1:2025cv01084/90425
  • Samba TV filed a motion to dismiss for failure to state a claim (D.I. 19, ~Feb. 2026); briefing through March 2026 (reply D.I. 23; request for oral argument D.I. 24). The case was then consolidated into lead case 1:24-cv-00831-MN (order dated 2026-07-31), with all future filings in the lead case.
  • Docket source (UniCourt, case summary): https://unicourt.com/case/pc-db5-caseguaee43baca23b-[2203774](/patent/2203774)

Federal Circuit (2026): No CAFC 2026 docket or appeal naming US 12,457,388 was found in my searches. I cannot confirm a Federal Circuit appeal of this patent as of the current date; absence of a search hit is not proof that none exists. Separate Intent IQ family IPRs exist (e.g., IPR2022-01315 Roku; IPR2022-01505 Samsung; IPR2024-00666 re US 10,715,878), but those concern different patent numbers and are not '388 proceedings.

5. Explicit uncertainties / conflicting data

1. Issue-date conflict (flagged, not auto-corrected). An auto-generated complaint analysis for Intent IQ v. Samba TV (Ex Parte, doc. 15) states: "U.S. Patent No. 12,457,388 ... issued April 2, 2024 (Compl. ¶36)." This contradicts the authoritative patent record, which shows US 12,457,388 granted 2025-10-28. I report both literally. The 2025-10-28 grant also postdates the 2025-08-28 complaint's initial "Report to the Commissioner" (which listed only US 8,677,398; 10,321,198; 11,949,962), suggesting '388 was added to the case by later amendment — hence the internal inconsistency in that commentary.
2. Claim text not verified. The literal independent claims of '388 were not in the supplied text; my overview in §3 is reconstructed from the abstract/summary and a third-party complaint summary. Both may include paraphrase.
3. Ex Parte analysis also mislabels the '198 patent as "'398" in its prose (the '398 is actually US 8,677,398). Reproduced as-is; do not treat as an identifier for '388.
4. Patent number formatting. The record renders this patent both as US12457388B2 (no commas) and as US 12,457,388 B2; treated as one and the same.

Authoritative sources: https://patents.google.com/patent/US12457388/en • https://npe.law.stanford.edu/patent/12457388 • https://dockets.justia.com/docket/delaware/dedce/1:2025cv01084/90425

I'll research this patent's litigation history using the sources you specified.

Found the primary case. Let me dig deeper for details and any additional suits.

Let me verify whether this patent is asserted in any other cases and confirm current case status.

Bottom line

I found one (1) litigation that asserts U.S. Patent No. 12,457,388 (US12457388B2). No other suits, no Federal Circuit appeal, and no PTAB proceedings involving the '388 patent were identified in the sources searched.

Case asserting US 12,457,388

Procedural history (as retrieved)

• Aug. 28, 2025 — Intent IQ files complaint for patent infringement against Samba TV, Inc. (D.I. 1). The Rule 3 report to the Commissioner filed that day listed only three patents: US 8,677,398 B2; US 10,321,198 B2; US 11,949,962 B2 — i.e., the '388 patent was not in the original pleading (consistent with the '388 patent issuing Oct. 28, 2025, after the original filing).
• Sept.–Nov. 2025 — Stipulations extend Samba TV's response deadline (to Oct. 20, 2025, then Nov. 19, 2025).
• Dec. 31, 2025 — Docket reflects an amended complaint and a new Patent/Trademark Report to the Commissioner; this is when the '388 patent appears to have entered the case. The Stanford NPE Litigation Database case record for Intent IQ, LLC v. Samba TV, Inc. lists the asserted patents as 10,321,198; 11,949,962; 12,457,388; 8,677,398, and the complaint-analysis source describes the '388 patent as count/asserted against Samba TV's "equipment and platform components, including its identity graph and content recognition technology" (accused products identified as the Samba TV platform, Samba TV Identity, Samba TV Advertising ID, household identity graph, Screen6 cross-device graph, Samba Automatic Content Recognition, and Omniscreen attribution).
• Feb. 4, 2026 — Samba TV files a Motion to Dismiss for Failure to State a Claim (D.I. 19), with opening brief (D.I. 20).
• Mar. 4, 2026 — Intent IQ files answering brief in opposition (D.I. 22). Mar. 11, 2026 — Samba TV files reply (D.I. 23) and a Request for Oral Argument (D.I. 24). The motion remains undecided on the record retrieved.
• July 31, 2026 — Order (D.I. 25) granting motions to consolidate and sua sponte consolidating this case and others (lead case C.A. No. 24-831-MN, Intent IQ v. LiveIntent); all future filings to be made only in the lead case. Docket remark: "These case[s] have been consolidated with Lead C.A. No. 24-831-MN."

Sources

• Stanford NPE Litigation Database, patent page for 12457388: https://npe.law.stanford.edu/patent/12457388 — lists exactly one case.
• Stanford NPE Litigation Database, case record: https://npe.law.stanford.edu/case/[307865](/patent/307865)
• Justia dockets: https://dockets.justia.com/docket/delaware/dedce/1:2025cv01084/90425
• UniCourt (docket entries incl. D.I. 19–25): https://unicourt.com/case/pc-db5-caseguaee43baca23b-[2203774](/patent/2203774)
• Compliant-analysis summary identifying the '388 patent as asserted (be aware this source contains an erroneous issue date — see caveat below): https://ai-lab.exparte.com/case/dct/ded/1:25-cv-01084/doc/analysis/15
• Litigation links embedded in the Google Patents record itself: Unified Patents Delaware case page https://portal.unifiedpatents.com/litigation/Delaware%20District%20Court/case/1%3A25-cv-01084 and a Darts-IP "Family has litigation / first worldwide family litigation filed" notation.

Related Intent IQ matters that do not appear to involve the '388 patent

These are part of the same plaintiff's campaign and are relevant context, but I found no confirmation that U.S. 12,457,388 is asserted in any of them:

• Intent IQ, LLC v. LiveIntent, Inc., 1:24-cv-00831 (D. Del.) — lead consolidated case.
• Intent IQ, LLC v. Connatix, 1:24-cv-01296 (D. Del.); v. Zeta, 1:25-cv-00579; v. Criteo, 1:25-cv-01266 (all D. Del., consolidated for pretrial).
• Intent IQ, LLC v. tvScientific, Inc., 1:26-cv-00089 (D. Del., filed Jan. 27, 2026) — report lists only US 8,677,398 B2.
• Intent IQ, LLC v. Optable Technologies, Inc., 2:26-cv-00609 (E.D. Tex., filed July 24, 2026).
• Intent IQ, LLC v. Snap Inc., 2:26-cv-06562 (C.D. Cal., transferred from W.D. Tex. 7:25-cv-00181).
• PTAB: LiveIntent, Inc. v. Intent IQ, LLC, IPR2025-01317 (challenges US 8,677,398); Viant Technology LLC v. Intent IQ, LLC, IPR2026-00330 (challenges US 11,949,962). Neither challenges the '388 patent on the records retrieved.

Confidence, caveats, and limitation of results

1. Only one case for the '388 patent. The Stanford NPE Litigation Database patent page for 12457388 lists a single case (Samba TV), and the Google Patents record for US12457388B2 carries only the Delaware 1:25-cv-01084 litigation link. I did not find a second suit, an appeal, or an IPR directed at this patent.
2. Dates are as-retrieved from docket aggregators, and there is a chronology conflict. The records I retrieved include docket entries dated later than the "current date" (April 26, 2026) stated in your instruction — specifically the July 31, 2026 consolidation order and aggregator "last updated" stamps in August/September 2026. I am reporting them literally rather than correcting or discarding them; you should treat the post-April 2026 entries as the freshest ground truth per the retrieval.
3. Date error in one secondary source. The complaint-analysis page describes the '388 patent as "issued April 2, 2024." That is inconsistent with the Google Patents record (filing date 2024-08-09; granted/publication date 2025-10-28; pre-grant publication US20250142177A1 on 2025-05-01) and with the fact that the original Aug. 28, 2025 Rule 3 report listed only three patents. I treat the Oct. 28, 2025 issuance as correct and the secondary source as erroneous.
4. No CAFC or PACER first-hand confirmation. I did not retrieve a PACER or CAFC docket directly; the procedural detail above comes from Justia/UniCourt/open-public-records aggregations of the D. Del. docket. Outcome data (e.g., resolution of the motion to dismiss) was not available on the records retrieved.
5. "Family has litigation" note. The Google Patents page flags family-wide litigation (Darts-IP) — that is a family-level flag covering the wider AlmondNet/Intent IQ portfolio, not necessarily additional suits on this specific patent. I found no family-level case beyond the one above that specifically names U.S. 12,457,388.

If you need the operative amended complaint's '388-specific counts, the D.I. 19 motion to dismiss papers, or a PACER-authenticated docket for a validity/standing opinion, those documents would need to be pulled directly from PACER (D. Del., 1:25-cv-01084, now filed under lead C.A. No. 24-831-MN).

I'll verify the structured data against public sources before writing anything up.

Let me check specifically for PTAB activity on the '388 patent itself and the related litigation.

Now let me search specifically for PTAB dockets keyed to this patent number.

PTAB / AIA Trial Report — US 12,457,388 B2

Bottom line up front: The USPTO Open Data Portal returns zero AIA trial proceedings against US 12,457,388. Web verification (PTAB E2E documents surfaced via search, Docket Alarm PTAB dockets, and third‑party PTAB trackers) found no petition, institution decision, or FWD keyed to patent number 12,457,388 as of 2026-09-28. That is a real and unusual result: the patent owner's family is one of the most heavily IPR'd portfolios in ad tech (26 PTAB cases faced per one tracker), and this patent is already being asserted. I flag the caveat that ODP ingest can lag and a newly filed, not‑yet‑indexed petition cannot be excluded.

Proceedings overview

Total AIA trial proceedings on US 12,457,388: 0 — no active proceedings, no claims invalidated, no claims sustained, no settlements, no institution denials; the patent is entirely untested at the PTAB, which for a defendant means the patent is not softened by any PTAB record, but also that no § 315(e)(2) estoppel, no claim cancellation, and no adverse PTAB claim construction is available to you — you are the first mover, and every invalidity ground (including system art and § 112) remains on the table in district court.

Proceedings on US 12,457,388

None on file

• Type: n/a
• Filed: n/a
• Status: No AIA trial proceeding (IPR / PGR / CBM) identified for US 12,457,388 as of 2026-09-28.
• Judge panel: n/a
• Petition grounds: n/a
• Institution decision: n/a
• Final Written Decision: n/a
• Settlement / termination: n/a
• Appeal: n/a
• Defensive value: Because there is no FWD and no institution decision, there is no § 315(e)(2) estoppel running against any petitioner and no cancelled claim you can cite. Conversely, you cannot point to a PTAB judgment of invalidity as leverage. The only hard, verifiable facts are: (i) the patent issued 2025-10-28; (ii) its PGR window (9 months post‑grant) closed on or about 2026-07-28 with no PGR on file; (iii) it is asserted in Intent IQ, LLC v. Samba TV, Inc., D. Del. 1:25-cv-01084 (filed 2025-08-28), now consolidated for pretrial purposes into lead case 1:24-cv-00831-MN (order 2026-07-31). Any IPR you file would be the first art‑based challenge in the family branch that ends in the '388.

Family‑adjacent PTAB activity (NOT proceedings on the '388 — do not conflate)

These are proceedings against other Intent IQ / AlmondNet patents, several of which share the '388's specification lineage (the '388 is a direct continuation chain through 12/063,421 ← 11,831,964 ← 11,095,952 ← 10,321,198 ← 8,595,069 ← 8,566,164). They are useful for pattern evidence only. None of them estops anyone as to the '388, and none of them cancels any claim of the '388.

Detailed entries where the record matters most:

IPR2024-00421 & IPR2024-00422 — Viant Technology (joined) v. Intent IQ, LLC (U.S. Pat. No. 7,861,260)

• Type: Inter Partes Review (joined, combined FWD)
• Status: Final Written Decision, claims unpatentable
• Judge panel: Nabeel U. Khan, Scott B. Howard, David Cotta; Opinion by Howard, APJ
• Petition grounds: § 103 challenges to claims 1‑152 of the '260 patent (Viant asserted "considerable overlap between the prior art applied" across the two cases; identical expert declarations)
• Hearing: joint hearing 2025-08-28
• FWD: 2025-10-14. Quote: "we determine that Petitioner has shown by a preponderance of the evidence that the challenged claims are unpatentable." The Board also sua sponte struck Section II of Petitioner's reply claim‑construction brief.
• Defensive value: Shows the family's patents are not invulnerable when a petitioner survives institution on a different family branch — the Board will cancel a very large claim set (1‑152) on a well‑built § 103 record. Not estoppel‑relevant to the '388, but useful template/art‑citation research.

IPR2025-01317 — LiveIntent, Inc. v. Intent IQ, LLC (U.S. Pat. No. 8,677,398)

• Type: Inter Partes Review
• Filed: 2025-07-18
• Status: Director Discretionary Decision Deny (2025-11-20) — no institution; post‑institution fee refund approved 2026-01-12
• Defensive value: This is the case to study. Intent IQ successfully argued discretionary denial under the 2025-03-26 Director Memorandum on Interim Processes for PTAB Workload Management, emphasizing "settled expectations" of a family "in force for over a decade," weak merits, and an apparent RPI defect. A defendant filing on the '388 today should expect the same playbook — although the '388's recent issuance cuts against a settled‑expectations argument in a way that did not apply to the '398.

IPR2026-00330 — Viant Technology LLC et al. v. Intent IQ, LLC (U.S. Pat. No. 11,949,962)

• Type: Inter Partes Review
• Filed: 2026-04-02
• Status: Tracker reports Discretionary Denial; institution decision date reported 2026-08-06
• Record: Petitioners and their counsel jointly identified (Apploved/Applovin, Viant, Mediaocean power of attorney filings 2026-04-02); Patent Owner's preliminary response argued settled expectations (the '962 family claims priority to 2011-08-03; oldest enforceable family patent 8,683,502 issued 2014-03-25), weak merits, and unnamed RPIs.
• Defensive value: Confirms the current Director/Board posture is hostile to serial IPRs on aged Intent IQ family patents, especially where a parallel D. Del. action is live.

CAFC 26-1701 — Intent IQ, LLC v. Viant Technology LLC

• Type: Appeal from PTAB, originating case IPR2025-00128
• Filed: 2026-04-20
• Status: Pending; no judge assignment reported as of the source's last update
• Defensive value: The patent owner is appealing adverse FWDs, not just defending. Signals it will fight any claim‑cancellation outcome.

Strategic summary

1. Claim status of the '388. There is no PTAB record, so no claim of 12,457,388 is canceled, none is sustained, and all are untested at the Board. The patent issued 2025-10-28 and has an anticipated expiration of 2027-12-31 (20 years from the 2007-12-31 priority date, subject to any PTA), per the Google Patents bibliographic data — i.e., a very short remaining life. That changes the calculus materially: a petition filed in late 2026 would likely reach an FWD at or after expiry, which is itself a discretionary‑denial and a Sotera/efficiency argument for the patent owner and a "why pay?" argument for a defendant choosing between IPR and a § 282 invalidity case.

2. Estoppel landscape — clean slate. § 315(e)(2) estoppel attaches only to a petitioner (and its privies/RPIs) that obtains an FWD; there is no FWD, indeed no IPR, on the '388. No estoppel exists against anyone as to this patent. All prior‑art grounds remain available in the D. Del. case, including things IPR cannot reach — system/prior‑public‑use art, on‑sale art, and § 112 written description/enablement/definiteness. Note also that the family IPRs listed above were directed at different patent numbers; IPR estoppel is patent‑specific, so a LiveIntent or Roku or Samsung FWD on the '398, '962, or '878 does not bar any ground against the '388 — though those FWDs and institution denials are credible art/argument sources, and the institution denials tell you what the Board found unpersuasive.

3. Pattern signals. (a) Multiple IPRs by the same petitioners across the family: Roku filed at least five (all denied institution in 2022‑2023); Samsung at least four (denied); Meta at least two; FreeWheel/Viant at least five, including IPR2026-00330; LiveIntent one. This is a serial, coordinated defense campaign, not one-off filings. (b) The patent owner publicly touts licensing — in IPR2026-00330 it asserted the '962 patent "has been licensed to companies including Google, Microsoft, Samsung, and Amazon," which is a settled‑expectations/industry‑adoption story the Board has credited. Expect the same for the '388. (c) Defensive aggregator involvement: none found. The "Unified Patents" reference on the Google Patents page for the D. Del. case is a litigation‑data citation to Unified's public litigation dataset, not evidence Unified filed or funded an IPR on the '388. I found no Unified Patents (or other aggregator) IPR on this patent. (d) Litigation clustering: the Samba TV case, 1:25-cv-01084, is consolidated for all pretrial purposes with Intent IQ v. LiveIntent (1:24-cv-00831-MN) and with Connatix (24-1296), Zeta (25-579), and Criteo (25-1266) by Judge Noreika's 2026-07-31 order, with all future filings to be made in the lead case. That gives you co‑defendants, a shared invalidity fund, and a common interest privilege structure — the classic setup for a jointly funded IPR.

4. Which patents the Samba case actually puts at risk. Public case records list 10,321,198; 11,949,962; 12,457,388; 8,677,398 as the asserted patents. Note the defensive asymmetry: three of those four have already been challenged at the PTAB by other defendants, while the '388 has not been challenged at all. If you are evaluating which claim set to attack first, the '388 is both the newest and the least‑tested — and the least hemmed in by earlier Board reasoning.

Recommended next steps

1. Verify the § 315(b) clock for the '388 before anything else. The '388 issued 2025-10-28, after the 2025-08-28 original complaint in 1:25-cv-01084. A § 315(b) one‑year bar runs from service of a complaint alleging infringement of that patent — so it most likely runs from service of the amended complaint (docket reflects an amended complaint filed on or about 2025-12-30). Pull the actual service date; if it is late December 2025, your IPR deadline is late December 2026. Do not guess — the bar is statutory and unappealable.
2. There is no FWD to quote and no cancelled claim to cite. Say so in any demand‑response or IPR memo: "No PTAB proceeding has ever been instituted on U.S. 12,457,388; no claim has been canceled; every ground is available." If a demand letter from Intent IQ cites claims you believe are invalid, the leverage is the expiry (2027-12-31) and the untested nature of the claims — not a PTAB judgment.
3. No joinder route exists. Because no IPR has been instituted on the '388, there is nothing to join (35 U.S.C. § 315(c)). A first IPR on this patent must stand alone and must survive the Director's discretionary‑denial regime that just defeated LiveIntent (IPR2025-01317, denied 2025-11-20) and Viant (IPR2026-00330, denial reported 2026-08-06). Build your petition to address the 2025-03-26 Director workload memorandum factors head‑on: distinguish the '388 from the decade‑old family members (new claims, new limitations, recent issuance → no settled expectations), and consider a Sotera‑style stipulation to neutralize Fintiv given the consolidated D. Del. pretrial schedule.
4. Do not expect institution to be easy. The family's record shows the Board denying institution far more often than granting it where Intent IQ litigates the preliminary response hard. Where institution was granted and a full trial ran on a different family branch (IPR2024-00421/-00422, '260 patent), all challenged claims fell — the merits record is winnable with the right art and expert.
5. Preserve § 112 and system art in the district court case regardless. With no IPR estoppel, a defendant can and should press grounds IPR cannot hear. Note that Samba TV filed a Rule 12(b)(6) motion to dismiss (2026-02-04), with briefing through 2026-03-11 and a request for oral argument — the grounds are not fully public in the sources I could verify, but a § 101 eligibility attack on this family's claims is the obvious candidate and should be coordinated with any IPR strategy.
6. Calendar the exhaustion of PGR. The 9‑month PGR window closed ~2026-07-28 with no PGR filed, so IPR (and § 282) are the remaining vehicles; § 112‑based cancellation at the Board is no longer available via PGR.

Sources and verification notes

• Structured PTAB list: USPTO Open Data Portal "PTAB proceedings on file" block in the prompt — canonical and authoritative for the count of zero. I did not find a web source contradicting it.
• Patent bibliographic, priority, continuation chain, and anticipated expiration: https://patents.google.com/patent/US12457388/en (fetched 2026-09-28). The cross‑reference paragraph lists the full chain: 18/384,835 → 17/401,078 → 16/435,051 → 14/089,645 → 12/981,925 → 11/968,117 (2007-12-31 filing).
• IPR2024‑00421/-00422 FWD (2025-10-14, Paper 41, '260 patent, claims 1‑152 all unpatentable; panel Khan, Howard, Cotta): PTAB document hosted at Docket Alarm, https://www.docketalarm.com/cases/PTAB/IPR2025-01317/LiveIntent_Inc._v._Intent_IQ_LLC/docs/10-31-2025-Petitioner/Exhibit-1023-...pdf (Ex. 1023 in IPR2025-01317). Confirm against PTAB E2E / PTAB Decisions.
• IPR2025‑01317 file entries (filed 2025-07-18; Paper 9 Director Discretionary Decision Deny 2025-11-20; Paper 11 refund 2026-01-12): Docket Alarm PTAB docket search results for patent 8,677,398.
• IPR2026‑00330 details (filed 2026-04-02; parties Viant/Apploved‑Applovin/Mediaocean; '962 patent; institution decision date 2026-08-06; status reported as discretionary denial): third‑party PTAB trackers (docketalarm sandbox; ipverse.greyb.com). Treat the Aug 2026 institution decision as reported‑but‑unverified; pull the paper from PTAB E2E.
• CAFC 26‑1701 (Intent IQ, LLC v. Viant Technology LLC, filed 2026-04-20, from IPR2025‑00128): third‑party CAFC docket aggregator; not verified against PACER or CourtListener. Confirm via CourtListener docket search or PACER.
• Family IPR list and statuses (Roku, Samsung, Meta, FreeWheel, Viant, LiveIntent): ipverse.greyb.com PTAB case list for Intent IQ and PTAB documents surfaced in search. Individual statuses are secondary‑sourced; verify each at PTAB E2E (https://ptacts.uspto.gov/ptacts/) and PTAB Decisions (https://www.uspto.gov/patents/ptab/decisions).
• District court: Intent IQ, LLC v. Samba TV, Inc., 1:25-cv-01084 (D. Del., filed 2025-08-28; Judge Maryellen Noreika; consolidation order 2026-07-31 into lead case 1:24-cv-00831-MN); asserted patents 10,321,198 / 11,949,962 / 12,457,388 / 8,677,398 per Stanford NPE Litigation Database.

Explicit uncertainty statement: I could not locate any PTAB proceeding, institution decision, FWD, or appeal for US 12,457,388 in any indexed source. I am not aware of any facts that would let me state more than that. The '388's '164/'069/'198-lineage siblings have a dense PTAB history; the '388 itself has none. One caution: an AI‑generated complaint summary circulating for 1:25-cv-01084 misstates the '388's issue date as "April 2, 2024" (that is the '962 patent's date). The authoritative grant date is 2025-10-28.

I'll research this systematically. Let me start with searches on the assignment record, the assignee, the inventors, and the litigation.

Let me dig into the assignment records specifically, plus the corporate history of Intent IQ / AlmondNet.

Ownership & Assignment Analysis — US 12,457,388 B2

Data caveat up front (important): I attempted to pull the recorded assignment data from the USPTO Assignment Center (assignmentcenter.uspto.gov / assignment.uspto.gov) and could not retrieve the reel/frame records for this patent in this session — the Assignment Center search is a JavaScript application and returned no indexable records to my tools. I therefore cannot cite verified reel/frame numbers for US 12,457,388, and I will not fabricate any. Everything below is drawn from the patent front matter (authoritative text supplied), Google Patents legal events, and court/corporate records, and is labeled as to source. Where the Assignment Center data gap matters, I flag it.

Inventors

Pattern note: No evidence of inventor departures preceding a sale. This is the inverse of the classic fire-sale tell — the named inventor (Shkedi) is the founder and remains the controlling principal of the assignee family, and the patent was re-filed as a continuation as recently as 2024, i.e., the inventor is actively prosecuting and litigating his own portfolio. That pattern points away from a portfolio dump.

Original assignee

• Intent IQ, LLC (Long Island City / New York, NY), per Google Patents, listed as both Original Assignee and Current Assignee.
• Corporate parent: AlmondNet, Inc. — confirmed by the Rule 7.1 Disclosure Statement filed in Intent IQ, LLC v. Samba TV, Inc., 1:25-cv-01084 (D. Del.), Aug. 28, 2025, "identifying Corporate Parent AlmondNet, Inc. for Intent IQ, LLC."
• Line of business: privacy-first identity resolution / ad-tech. Intent IQ delivers an "identity device graph" for cookieless, MAID-less and CTV environments (identity resolution, first-party ID clustering, bid enhancement, attribution). Sister company Datonics operates a data marketplace; AlmondNet is the group holding/pioneer entity.
• Does it ship a product embodying the claims? Yes, on the evidence. The asserted claims of the '388 family cover recording online activity / targeting based on TV-ad exposure mapped to an online device — which is the function of Intent IQ's identity graph and attribution products. The company state at the Samba TV complaint level is that the accused Samba components ("identity graph," "cross-device graph," "content recognition," "attribution") are the mirror image of what Intent IQ itself sells.
• Status: Operating, not dissolved, not in bankruptcy. It is a serial but commercially active licensor: Intent IQ/AlmondNet publicly state their portfolio is "licensed by some of the largest companies in the world, including Google and … Microsoft," and (April 2026) a license executed with Amazon resolving the AlmondNet v. Amazon dispute.

Assignment timeline

The public record for this specific patent shows no post-issuance assignment activity and no chain of successor LLCs. Google Patents legal events for US 12,457,388 list only: filing (2024-08-09), publication of US 2025/0142177 A1 (2025-05-01), and grant (2025-10-28). The patent names Intent IQ, LLC as both original and current assignee — i.e., ownership has not moved out of the original assignee.

Because the Assignment Center records could not be retrieved, I cannot supply reel/frame entries, execution dates, or the assignment correspondent of record. What I can document from the family record:

• Family assignment practice (related applications, not this one): In adjacent AlmondNet/Intent IQ applications, the recorded assignment instrument runs "from Roy Shkedi (inventor) … to AlmondNet, Inc., a Delaware corporation" (assignment text quoted in prosecution documents for the media-properties family, e.g., application Ser. No. 11/763,286 / 13/491,454; attorney docket "14-EP-CON3"). This establishes that the inventors' original conveyance in this corporate family was inventor → AlmondNet, Inc. I could not confirm whether the '388 line records a second AlmondNet, Inc. → Intent IQ, LLC assignment or is held directly by Intent IQ — that is exactly the record the Assignment Center lookup would have settled. Flagging as unresolved.

Correspondent observation (prosecution, not assignment): The recurring attorney of record across the Intent IQ / AlmondNet portfolio is Louis J. Hoffman, P.C. (Louis J. Hoffman, Reg. No. 38,918), Customer Number 26362, 14301 North 87th Street, Suite 312, Scottsdale, AZ 85260 (later 7689 East Paradise Lane, Suite 2, Scottsdale, AZ 85260). He appears as the filing attorney on the family applications (e.g., the '962/'198 lineage and the media-properties family). This is a repeat prosecution correspondent across the whole family — but I have not confirmed he is the assignment correspondent of record, so I do not score it as the "repeat assignment correspondent" NPE tell below.

Timeline diagram

timeline
    title Ownership of US 12457388
    2007 : Priority application filed by Shkedi
    2013 : Parent continuation filed
    2024 : This continuation filed Aug 9
         : Intent IQ files ad tech suits
    2025 : Patent issued Oct 28
         : Suit filed against Samba TV
    2026 : Amazon license resolves dispute

NPE / troll-pattern signals

1. Shell-entity transfer — Not present (on available records). No recorded move from an operating assignee to a licensing-only LLC. The patent's original and current assignee are the same entity (Intent IQ, LLC), which is a product company with an R&D center, paying customers, and a Rule 7.1 corporate parent (AlmondNet, Inc.). I could not verify the Assignment Center reel/frame, so this is a qualified "not present."

2. Known asserter in the chain — Not present. Intent IQ / AlmondNet do not match the named NPE lists (Acacia, Marathon, IV, IPNav, Wi-LAN/Conversant, Vringo, Pendrell, Round Rock, MPHJ, Lumen View, Spangenberg entities, etc.). The Stanford NPE Litigation Database classifies the asserter in the Samba TV case as a "Product company" (category 8), not an NPE — see npe.law.stanford.edu/case/307865 and npe.law.stanford.edu/patent/12457388. Caveat: it is a high-frequency plaintiff (IP Verse reports ~26 PTAB cases faced and numerous district-court suits), which is a behavioral overlap with NPEs even though the corporate classification is operating company.

3. Repeat correspondent across the chain — Unclear. The same firm, Louis J. Hoffman, P.C. (Customer No. 26362), recurs across Intent IQ/AlmondNet prosecution filings — but the signal as defined requires recurrence on assignment recordings, which I could not retrieve. One recurring attorney is suggestive of centralized portfolio management, not proof of NPE shell-layering.

4. Cascading transfers — Not present. No evidence of multiple consecutive assignments through chained LLCs in <24 months. Ownership appears continuous within Intent IQ/AlmondNet.

5. Pre-litigation transfer — Not present. The '388 patent was asserted in the amended Samba TV complaint (D. Del., consolidated into lead C.A. 24-831-MN) without any intervening ownership change; the accuser is the same entity that filed the continuations.

6. Bankruptcy fire-sale — Not present. No Chapter 7/11 of the assignee; AlmondNet/Intent IQ are actively operating and licensing (Amazon license, April 2026).

7. Privateering — Not present. There is no operating-company-to-NPE hand-off; AlmondNet/Intent IQ assert on their own behalf. (Note: the group's public posture — asserting foundational ad-tech patents against Roku, Samsung, Amazon, Samba TV, LiveIntent, Criteo, Adform, Snap — is assertive licensing, but from the practicing entity itself, not via a proxy shell.)

8. Defensive aggregator — Not present. The chain does not terminate at RPX, AST, LOT, Unified, or OIN.

Verdict

Operating-company assertion.

Intent IQ, LLC is the original and current assignee on the face of US 12,457,388, is the product entity of the AlmondNet Group (parent confirmed by the Rule 7.1 statement in 1:25-cv-01084, filed 2025-08-28), ships identity-resolution/attribution technology that reads on the asserted claims, licenses its portfolio to Google, Microsoft, Samsung and Amazon, and sues ad-tech competitors (Samba TV, LiveIntent, Criteo, Adform, Snap, Roku, Samsung) rather than end-user strangers. The Stanford NPE Litigation Database scores the asserter as a "Product company," not an NPE. The one caveat is the unverified assignment reel/frame record — I could not confirm from the USPTO Assignment Center whether a separate AlmondNet→Intent IQ intra-group assignment exists, so I do not claim a clean chain-of-title citation; I only claim that the recorded owner and the litigant are the same operating entity.

Verify against the source: USPTO Assignment Center — https://assignment.uspto.gov/patent/index.html#/patent/search (search "12457388" and by application "18/800,005"). Cross-check: Google Patents legal events (https://patents.google.com/patent/US12457388/en), Stanford NPE DB (https://npe.law.stanford.edu/patent/12457388), and the D. Del. docket 1:25-cv-01084.

I'll research US patent 12457388 and its cited prior art. Let me search for it.

Let me look for the front-page citations of US12457388 and related prior art references.

The searches are noisy. Let me target the patent's front-page citation list and prosecution record directly.

Prior-Art Analysis — U.S. Patent No. 12,457,388 B2

Important caveat up front: My retrieval of the official front-page "References Cited" list for US12457388B2 — i.e., the exact (56) citations printed on its face — was blocked (tool-step limit reached mid-search). Per your strict rule, I will not fabricate citation entries. Below I separate (A) what is verified from the record of this patent family, and (B) what I could not confirm. Where I am uncertain of a title or number, I say so explicitly.

1. Patent identification (verified from the document you supplied)

Litigation/PTAB context (verified): The patent is asserted in Intent IQ, LLC v. Samba TV, Inc., No. 1:25-cv-01084 (D. Del.), together with US 10,321,198; 11,949,962; and 8,677,398. The '198 ancestor (US 10,321,198) was challenged at the PTAB by Roku (IPR2022‑01315) and Samsung (IPR2022‑01505). This matters because the prosecution/IPR records of the ancestor are the best evidence of what the family's examiner and challengers treated as prior art.

2. What the claims cover (for the §102 comparison)

Consistent with the abstract and summary you supplied (and with the ancestor '198 claims quoted in the Roku record), the independent claims are directed to a computer‑implemented method/system in which:

• (a) a selected online advertisement is automatically directed to — or online activity is automatically tracked and recorded from — an online user interface device that corresponds, at that time, to a first online access identifier;
• (b) the directing/tracking is based on information from a user profile that references or includes a first set-top box (STB) identifier, which information is derived at least in part from automatically collected data on user behavior with respect to a TV advertisement delivered to a set-top box; and
• (c) the first online access identifier is associated with the first STB identifier (via the profile, a database, a common IP address, or otherwise).

Dependent-claim themes in the family include: linking the online ad/tracked activity to the delivered TV ad in a database; interaction/command‑based observed behavior; cookies/tags vs. IP addresses as identifiers; associations received from a TVP/ISP or established by common IP address; and use of an STB/ad database with timestamps. (Claim numbering in US 12,457,388 should be read from its own printed claims — see caveat in §5.)

3. Verified prior-art references (from the family's prosecution record)

These are the references the USPTO examiner actually cited against the ancestor application 14/089,645 that issued as US 10,321,198 — and which therefore appear in the file history and on the corresponding family front pages:

Key takeaway for §102: Blasko is the only reference in the verified record that the examiner asserted under §102 anticipation, and it was successfully distinguished — the Board held Blasko does not disclose associating an STB identifier with an online-device identifier without PII. Any §102 theory against US 12,457,388 would have to overcome that same finding unless it identifies a different, single reference that expressly discloses non‑PII cross‑device association plus (b) profile derived from TV-ad viewing/interaction behavior.

4. IPR references (challenger-side; §103 grounds, not §102)

From the Roku petition exhibits retrieved:

• Madhavan (Ex. ROKU‑1008) — described in the petition as disclosing a user profile containing demographic, geographic and behavioral data, linked to events/activities tables and category tables. (I could not verify the exact publication number/title/date from the returned snippets — treat the citation as unverified.)
• Whitehead (Ex. ROKU‑1001/1003 context) — described as disclosing targeted advertising selected at the local-area-network level (ads selected based on RInyte/IP‑access‑session information across a PC, an Internet‑connected STB, and a mobile handset sharing a common IP address), including associating device identifiers based on IP access session / common IP address B plus timestamp.

Roku's ground was §103 (Madhavan + Whitehead), arguing the combination renders obvious the association of the first online access identifier with the first STB identifier and the use of a database for that association. These were not advanced as §102 anticipatory references, and I have not verified their bibliographic details (numbers, dates, titles) directly.

5. Citations I could NOT verify (so I do not assert them)

The following appeared only as related-family mentions in Google Patents "Cited By"/family listings, not as confirmed (56) prior-art citations of US 12,457,388. They are Shkedi/Intent IQ family members or older Blasko-era publications, and their citation status on this patent's face is unconfirmed:

• US 8,997,138 B2 (Intent IQ, 2015-03-31) — "Correlating online behavior with presumed viewing of television advertisements" (same family line — likely not prior art)
• US 7,747,745 (Shkedi) — "Media properties selection method and system based on expected profit from profile-based ad delivery" (same inventor family)
• US 2002/0019769 A1 — "System and method for establishing incentives for promoting the exchange of personal information and targeted advertising"
• US 2002/0032603 A1 — "Method for promoting internet web sites"

I cannot confirm these are on US 12,457,388's front page, so I am not characterizing them as its cited prior art.

6. Bottom line

1. Only one reference in the verified record (Blasko, US 2001/0049620 A1) was ever asserted under §102, and it was distinguished on the "no non‑PII association of STB identifier ↔ device identifier" ground.
2. Ertugrul (US 2008/0222283 A1) and the Roku references (Madhavan, Whitehead) were §103 combination references, not §102 anticipatory art.
3. Because US 12,457,388 is a continuation that issued in 2025 with a 2007-12-31 priority date, additional art in the 2001–2007 window could exist that is not reflected in the ancestor's prosecution. I could not retrieve the patent's own front-page (56) list, so a complete anticipatory-reference mapping is not possible from the data I obtained. To finish rigorously you should pull the printed (56) references from the granted PDF of US 12,457,388 and the certified file wrapper of 18/800,005, then map each to the claim language.

Recommended next step: retrieve the granted-PDF "References Cited" section and the POSITA/claim-construction record from Intent IQ v. Samba TV (1:25-cv-01084) — those will supply the definitive reference list and any §102 stipulations.

Would you like me to attempt another retrieval pass focused specifically on the granted-PDF front page and the Samba TV complaint exhibits?

I'll ground this analysis by pulling the actual prior-art/family data for this patent before reasoning.

Let me pin down the exact prior-art references cited on this patent and in the litigation/IPR record for the family.

Now let me retrieve the actual claim set of US 12,457,388 (and its immediate parent US 12,063,421) so the § 103 analysis is claim-specific.

Obviousness Analysis of US 12,457,388 B2 under 35 U.S.C. § 103

0. What I used, and one important sourcing caveat

Caveat on the "Prior Art section of this page." The text I have of the Google Patents page for US 12,457,388 B2 does not reproduce the bibliographic prior‑art tables — there is no "(56) References Cited" list and no "Similar Documents / Cited By" table in the fetched text. What the page does give under the prior‑art heading is:

• Prior art keywords: stb; online; television; user; identifier
• Prior art date: 2007‑12‑31
• Priority date: 2007‑12‑31; Filed: 2024‑08‑09; Granted/Published: 2025‑10‑28
• A reference to "Family has litigation" (Darts‑IP family 40800349) and a Delaware District Court case, 1:25-cv-01084

Because the page's reference lists are absent, I grounded the analysis in the reference lists and prior‑art record I could verify for this exact specification family: (a) the "(56) References Cited" list printed on the face of the immediate ancestor US 10,321,198 B2 (the '198 patent) as reproduced in district‑court filings; (b) the ex parte appeal record for the '198 application (Appeal 2017‑000325); and (c) the IPR petitions against sibling patents of the same specification (IPR2022‑01315, IPR2022‑01505). Where I could not verify a reference number I say so rather than guessing.

Also note: the application's own front‑page claim set was cut off in the fetched text. I therefore analyze the claim scope from (i) the '388 abstract and summary, which are reproduced in full on the page, and (ii) the corresponding, verified claims of the ancestors in the same continuation chain (the '198 patent's claim 1 and dependent claims 9, 10, 17, 20, 21, 33, as recited in the record). If a limitation‑by‑limitation chart against the literal text of '388 claim 1 is needed, the granted claims should be pulled directly — I flag this rather than invent claim numbers.

1. Governing legal framework

Statute. Priority is 2007‑12‑31 (via 11/968,117 → 8,566,164 → 8,595,069 → 10,321,198 → 11,095,952 → 11,831,964 → 12,063,421 → 18/800,005). The application was filed 2024‑08‑09 as a continuation of Ser. No. 18/384,835 (US 12,063,421). Assuming every claim is supported by the 11/968,117 disclosure (which the family's consistent continuation practice suggests, and which the page's "Anticipated expiration 2027‑12‑31" reflects — i.e., 20 years from the 2007 filing), the claims are entitled to the 2007 priority date and are governed by pre‑AIA § 103(a). If any claim adds matter not supported by the 2007 disclosure, the AIA version of § 103 and AIA § 102 would apply instead — that is the threshold question a validity challenge should resolve first.

Consequence for prior art. Under pre‑AIA law:

• References published before 2007‑12‑31 are § 102(a)/(b) art (e.g., Blasko, pub. 2001‑12‑06).
• References with an earlier effective filing date but published after 2007‑12‑31 can be § 102(e) art (e.g., Ertugrul, US 2008/0222283 A1, pub. 2008‑09‑11 — qualifies only if its effective filing date predates 2007‑12‑31; Madhavan, US 8,180,674, filed 2006‑10‑03).
• § 103(c) caveat: subject matter that qualifies only under § 102(e)/(f)/(g) and was commonly owned at the time of invention is disqualified as prior art for § 103. Much of the "References Cited" list on the family's front pages is Shkedi's own prior work (e.g., US 7,861,260; US 2008/0259906; US 2005/0086112; the incorporated‑by‑reference application Ser. No. 11/736,544). Those references are weak or unavailable as § 103 art because, depending on inventive entity, they are either not "by another" or fall within § 103(c). The strongest § 103 case therefore rests on third‑party art (Blasko, Ertugrul, Madhavan, Whitehead, Vallier, Baum, Damick, etc.), not on the inventor's own chain.

POSITA. A person of ordinary skill in the art in December 2007 would have been a practitioner (or team) with a bachelor's degree in EE/CS and ~2–3 years of experience in network advertising delivery, IP networking (DHCP/NAT, cookies/tags) and/or interactive television/set‑top‑box middleware — consistent with how the family's own specification frames the field.

2. The claim scope to be analyzed

The '388 abstract/summary recites, in substance, the following limitations (call them L1–L7):

The '198 patent's claim 1 was adjudicated in IPR as containing, in part (c), the express "without using personally identifiable information" requirement — a limitation the PTAB found dispositive in the applicant's favor on the Blasko record (see §6 below). Whether the '388 claims carry an equivalent PII‑free limitation materially changes the strength of any § 103 ground, and should be confirmed against the literal claim text.

3. Verified prior art available in the record

4. Ground 1 — Blasko in view of Ertugrul (and further in view of Howcroft)

This is the combination the USPTO itself actually made. In the '198 application, claims 50–60 and 63–80 were rejected as anticipated by Blasko, and claims 61–62 as obvious over Blasko in view of Ertugrul.

Mapping. Blasko provides L1, L3, L4, L5, L6: profile vectors derived from observed behavior, including a user's television viewing via an STB, used to select advertisements. Ertugrul supplies the missing structural spine — L2 and L7 (association of a first online access identifier with a device/STB identifier via ISP‑side recognition of a common local network), L8 (database records / tags / cookies as the indicia of association), and the storage mechanics of L10.

Motivation to combine. Both references are in the same field (network‑based targeted advertising across devices) and address the same problem: Blasko determines what ad to serve from cross‑device profile data but does not specify how the devices are reliably associated; Ertugrul teaches precisely that association mechanism, including ISP‑observed common‑network association and cookie/tag persistence. Combining them is the KSR "known technique to improve a similar device in the same way" and "combining prior‑art elements according to known methods to yield predictable results" — the expected result being correctly addressed cross‑device advertising. Ertugrul also supplies the express privacy rationale: forming the association from UIDs/IP‑level context rather than user identity.

The known weakness. The Board, in Appeal 2017‑000325, reversed the Blasko rejection because Blasko's correlation server uses PII (via opt‑in) to correlate profile vectors, and therefore did not disclose associating a set‑top‑box identifier with a device identifier without using PII (as then required by claim 50). Under pre‑AIA § 103, however, that gap is closeable: Howcroft (US 2008/0040742) was applied by the Office in parallel prosecutions of this family for exactly this limitation — "electronic indicia of the electronic association include no personally identifiable information" — with the articulated motivation of "not being able to identify a particular user." So the correct § 103 formulation is Blasko + Ertugrul + Howcroft, not Blasko alone. This is also the KSR "finite number of identified, predictable solutions" rationale: the privacy/non‑PII association was one of two recognized design choices (identified vs. anonymous) with a known benefit.

Bottom line for Ground 1: strong for L1–L8 and L10 if the claims do not import a specific non‑PII requirement that the combination cannot meet; medium strength on L9 (see Ground 2/Vallier), which Blasko does not squarely address.

5. Ground 2 — Madhavan in view of Whitehead (and further in view of Vallier)

This is the combination Roku advanced in IPR2022‑01315 against the sibling '878 patent, and it maps well onto the recording/tracking claims of this family.

Mapping.

• L1–L6 (Madhavan): Madhavan "enabl[es] automatic matching of advertising information to events or actions and interests of users"; user actions on media device 140 include user interaction with television advertisements inserted in linear broadcast and VOD via an STB; targeted ads are selected based on usage from a PC, an internet‑connected STB, and a mobile handset at times 2504/2508/2510 and sent at time 2512.
• L7, L10 (Whitehead): Devices coupled to a common residential LAN — "PCs, PDAs, smart phones, cellular telephones, set‑top boxes" — are associated using non‑PII information, e.g., "a single IP address assigned to the home router" plus timestamp; "a single IP access session can involve multiple devices either over time or at about the same time"; video‑service usage revealing marketing information is collected "during the IP access session and used later in that session to send marketing content addressed to the IP access session IP address." This is the common IP address association branch of L10, and — because it is timestamped and session‑bounded — the STB/ad database with times and dates branch as well.
• L8 (Madhavan/Whitehead): A database (Madhavan's database 200/user tables 210; Whitehead's tables/records at blocks 300–360 of FIG. 5) storing the device‑identifier association is an obvious and disclosed implementation.
• L9 (Vallier): Monitoring of TV viewing habits including skipping, rewinding, re‑watching, click‑through interactive commercials, and tune‑in/record commands supplies the "user interaction responsive to the television advertisement" and "user commands executed during presentation" limitations.

Motivation to combine.

1. Same field, same problem: both are network‑based content delivery and targeted advertising; Madhavan manages content/ad selection at the LAN level over multiple devices, and Whitehead teaches how such devices can be associated anonymously.
2. Articulated advantage — privacy: Whitehead itself states that "anonymous marketing techniques can be considered more desirable, from a privacy standpoint, than non‑anonymous marketing techniques," and that an IP access session "can be treated anonymously." Where the prior art supplies the very reason to make the substitution (non‑PII association), the KSR "known technique to improve a similar device" rationale is met.
3. Predictable, beneficial result: the combination yields improved content management that accommodates the viewing habits of all devices in a common LAN, with reduced privacy exposure — a predictable engineering result.
4. Vallier's addition produces "more effective targeted advertisements" by capturing ad‑relevant engagement signals (skip vs. watch, rewind, interactive click) — the classic "improving a known device ready for improvement" rationale.

Timing note. Madhavan (filed 2006‑10‑03) is § 102(e) art. Vallier (US 9,495,682) must be checked against the 2007‑12‑31 date on its earliest effectively filed disclosure, and it may also be vulnerable to a § 103(c) common‑ownership argument if it is within the same corporate family; the petitioner in IPR2022‑01315 treated it as prior art, but that is a contested point worth verifying against the claim of priority actually at issue.

6. Ground 3 — Madhavan in view of Baum, Harrington, and Damick (plus Rankin; plus Varghese)

Roku's IPR2022‑01315 Ground 1 was Madhavan + Baum + Harrington + Damick for the base claims, Ground 2 added Rankin, and Ground 3 added Varghese — with the same "reasons to combine" structure (same field; addressable/IPTV ad delivery and household device identification improve one another; predictable results). For the '388 claims, this ground is most useful for the dependent limitations:

• Baum (US 7,843,923) — addressable/IPTV advertisement delivery infrastructure relevant to "directing a selected online advertisement" and to the ad server's role.
• Damick (US 2006/0242294) — advertising management/device identification.
• Rankin (US 2003/0207685) — network address translation / device‑address management, relevant to the common IP address and dynamic‑address branches of L10.
• Varghese (US 7,908,645) — relevant where claims recite particular reporting or measurement of ad exposure.

Because '388 is a continuation, this ground is complementary to Ground 2 rather than independent: it supplies alternative artisan‑recognized mechanisms for the same L7/L10 association and for the ad‑delivery mechanics.

7. Dependent‑claim combinations to close each remaining gap

8. Why a POSITA would have combined these (consolidated motivation)

Applying the KSR/MPEP rationales the Board itself recited in the IPR petitions:

1. Predictable combination of known elements — each reference contributes a mechanism (profile collection, ad selection, device association, command capture) already known in the art, producing the expected result of better‑targeted advertising and measurable ad effectiveness.
2. Known technique improving a similar device in the same way — non‑PII, IP‑session‑based cross‑device association was a known technique (Whitehead, Ertugrul) applied to a known cross‑device advertising system (Madhavan, Blasko).
3. Finite number of identified, predictable solutions — the artisan faced essentially two choices for associating a PC and an STB: with PII or without. Whitehead, Ertugrul, and Howcroft each expressly identify the anonymous option and its privacy benefit, giving a reasonable expectation of success.
4. Articulated design incentive — advertisers in the record repeatedly state the commercial benefit of measuring whether a TV ad produced an online response (the effectiveness‑measurement rationale appearing on the '388 page itself: tracking "can allow an advertiser to evaluate the effectiveness of the television advertising"). That is precisely the problem L1/L5/L8 address, giving a strong motivation independent of hindsight.
5. Technological convergence — by 2007 the same physical loop (cable/DSL plant) carried both television and Internet service, and ISPs/TVPs commonly observed both (the specification itself relies on "a common ISP/TVP" and on DHCP‑assigned modem addresses), so the artisan had both the data and the incentive to correlate.

9. Counterarguments, weaknesses, and what I do not know

(a) The strongest non‑obviousness argument is the PII‑free association. The PTAB's decision in Appeal 2017‑000325 did not merely find a gap in Blasko — it held that Blasko's own correlation mechanism requires PII and that the alternative (a single, non‑correlated profile vector) teaches away from cross‑device association without PII. Any '388 claim containing a "without using personally identifiable information" limitation must therefore be met by Ertugrul and/or Howcroft, not by Blasko alone. If a claim requires non‑PII association established contemporaneously by a common IP address assigned by a common ISP/TVP, that is still covered by the Whitehead ¶0080/¶0069 disclosure plus Ertugrul ¶0086 — but expect the patent owner to argue that IP address/time correlation is not an "association" of identifiers.

(b) L5 ("automatically collected data related to user behavior with respect to at least one television advertisement"). Blasko is primarily about programming/profile vectors; whether his system collects data tied to a specific delivered television advertisement (as opposed to a profile generally) is the point the applicant attacked, and it is the point the petition record supports best through Madhavan (media device 140 + user interaction with TV ads) and Vallier (commercial‑level viewing/skip data). Ground 1 alone is weaker than Ground 2 on this limitation.

(c) I do not have verified outcomes for IPR2022‑01315 (Roku v. Intent IQ) or IPR2022‑01505 (Samsung) — i.e., whether institution occurred, and whether any final written decision held the sibling claims unpatentable. I also do not know the disposition of the Delaware case 1:25-cv-01084 listed on the page. I am not asserting any litigation conclusion; the IPR petitions are cited only as evidence of how a skilled artisan and the Office actually structured the combinations, not as adjudicated invalidity.

(d) Secondary considerations. This family has a 2007 priority date and commercial adoption, which the patent owner can marshal as evidence of long‑felt need and industry recognition (the page's "Family has litigation" flags a crowded enforcement program). Any obviousness case should address: (i) whether the claimed combination solved a problem others had failed to solve (the specification's own framing suggests the cross‑media + dynamic‑address problem was known but unsolved); and (ii) whether the asserted nexus is to the claimed association feature rather than to the ad network's general commercial success.

(e) Self‑collision risk. Do not build the case on Shkedi's own earlier patents/publications (e.g., US 7,861,260; US 2005/0086112; the incorporated '544 application). Beyond the § 103(c) common‑ownership disqualification, several are incorporated by reference into the specification, which both weakens them as prior art and invites the argument that they are part of the invention's own disclosure.

10. Bottom line

• Ground A (primary): Blasko (US 2001/0049620 A1) + Ertugrul (US 2008/0222283 A1) + Howcroft (US 2008/0040742) renders obvious the core method of directing a selected online ad to, or automatically tracking/recording online activity from, an online user interface device based on a user profile that references an STB identifier and is derived from data about user behavior toward a delivered TV ad, with the online access identifier associated to the STB identifier (including non‑PII association and database/tag‑or‑cookie indicia). Motivation: same field, known cross‑device advertising + known ISP‑observed device association + an expressly recited privacy rationale.
• Ground B (reinforcing, and best on the dependent limitations): Madhavan (US 8,180,674) + Whitehead + Vallier (US 9,495,682) — and for dependent claims + Baum, Damick, Rankin, Varghese. Motivation: LAN‑level content/ad management improved with common‑IP‑plus‑timestamp non‑PII device association, further improved with commercial‑level viewing/skip/interaction monitoring.
• Mapping to limitations: Ground A carries L1–L8 and L10; Ground B carries L5, L8, and L9 (interaction and executed commands). Claims reciting the non‑PII association and the common‑IP address/TVP‑supplied association are the fulcrum, and those elements are met by Ertugrul/Howcroft and Whitehead respectively — not by Blasko, which is where the family has previously prevailed.

Verification steps I recommend before filing anything: (1) pull the literal granted claims of US 12,457,388 B2 and confirm whether the PII‑free association limitation is present and how it is worded; (2) confirm the effective filing dates of Ertugrul US 2008/0222283 A1 and the Whitehead publication (the latter's number was not verifiable from my sources); (3) confirm the pre‑AIA/AIA status by checking that every '388 claim is supported by 11/968,117; and (4) confirm the final written decisions in IPR2022‑01315 and IPR2022‑01505, which I could not verify.