A word before Frank starts commenting: he's spent forty years inside measurement, at just about every company that ever counted a viewer, and now he shows up in the margins here with what he actually thinks. Get used to him.

I heard a theory.

It arrived three separate times in two weeks, from three people who do not work together, do not compete directly, and did not coordinate. Ordinarily that is the good kind of coincidence.

The theory runs like this. The foundational ACR and audio watermarking patents are hitting expiry at roughly the same moment the market is pricing ACR datasets as durable proprietary assets. ACR can only confidently measure against a schedule, so a large share of what is sold as connected TV measurement is linear measurement under a newer label. Together those produce a commoditizing dataset priced as a fortress. Alphonso is the live test, with a court-mandated tender closing September 11 and the entire dispute turning on what an ACR dataset is worth.

Commoditizing.

I have been hearing that word since the first Bush administration. Panels were commoditizing. Set-top data was commoditizing. Digital was commoditizing, and then digital ate everything and bought a building.

Everything in this business is commoditizing right up until the moment somebody tries to buy it. Then it is a strategic asset and the price went up.

The theory deserves a full hearing before it gets taken apart.

Automatic content recognition is a method. A television samples what is playing across the panel or through the speakers, reduces that sample to a signature, and matches the signature against a library of known content. That is a technique. Techniques are patentable, and a United States patent runs twenty years from its earliest non-provisional filing date.

The techniques at issue are old. Audio watermarking and fingerprinting were substantially solved in the late 1990s and the 2000s, by companies measuring television when television was one signal on one screen. The clocks on the earliest claims started decades ago and have run continuously since.

Set that against what the resulting datasets are now worth. An ACR footprint is underwritten as a durable proprietary asset. It appears in valuations as a moat, in acquisition prices as something that cannot be replicated, and in strategy documents as a structural advantage. Vizio's data business was material to what Walmart paid. LG's is material to what the Alphonso tender is fighting over. The market prices the footprint as though the ability to build one is permanently constrained.

If the constraint was the patent, the constraint is ending. That mechanism is familiar. It emptied the codec licensing business and the compression licensing business. A method goes generic, the cost of building collapses, and the incumbent discovers its moat was a calendar.

Moat. Nobody in this industry has ever seen a moat. Nobody here has seen a castle. What they have seen is a contract with a renewal date, and they call it a moat because moat sounds like a thing that maintains itself.

A moat is a hole full of water. That is the entire concept. A hole. Full of water. Somebody has to keep filling it, and if that person quits, it then called a ditch.

Ten years I have been asking people to describe their moat without using the word moat. Nobody has done it yet. You got yourself a ditch. Ditch.

The second half of the theory is sharper. ACR matches signatures against a library, and libraries exist for scheduled content because somebody publishes a schedule. Match a signature against a broadcast lineup and the answer is knowable. Attempt the same inside a streaming application and there is nothing to match against, because no one publishes what a given service delivered to a given household at 8:14 on a Tuesday.

Which produces the uncomfortable conclusion. A technology whose confident answers are limited to scheduled programming has been sold for a decade as the measurement layer of the post-schedule era. The thing that replaced appointment television is the thing ACR sees worst.

A commoditizing method, a narrowing use case, and a market pricing the result as a fortress with a tender closing in eleven days. It is a strong argument. It is also mostly wrong.

Consider who is advancing it.

One holds ACR patents. If the method is a scarce, defensible, litigable asset right up until the moment it is not, the holder has a direct interest in the market understanding that the window is closing.

One sells an alternative to ACR-based measurement. If ACR is structurally blind inside streaming applications, the dollars leaving it have to arrive somewhere, and the party explaining the blindness sells the destination.

One operates a business whose entire positioning is that value accrues above the raw signal layer. If the signal layer is commoditizing, margin migrates upward, which is the product.

Three companies. Three revenue models. Three distinct commercial reasons for the thesis to be true. One identical conclusion, delivered independently inside a fortnight.

That is not convergent evidence. Convergent evidence is unrelated interests arriving at the same answer. This is related interests arriving at a shared answer and appearing unrelated in the process. It is the mechanism by which an industry manufactures its own conventional wisdom, and it explains a great deal of what everyone in advertising technology knows without being able to say where they learned it.

None of the three appears to be running a play. Each of them believes it. That is the harder problem, because a proposition sincerely held by people it happens to pay is close to unfalsifiable inside the trade press and almost never gets checked against a primary document.

Independently. That is the word everybody reaches for. They arrived at it independently.

Independently, from three companies that read the same six newsletters, attend the same four conferences, and hire exclusively out of one another. That is not independence. That is a neighborhood.

I have watched this entire industry change its mind in a quarter because a guy said something on a panel in Boca and not one person in that room had a follow-up question.

Kowalczyk Ratio on consensus is one to one. One idea. One room it started in.

This issue is part one, and a test in progress rather than a verdict. What follows is what the public record supports as of today. What it does not yet contain is the post-trial docket in the Samsung litigation, or any answer from a manufacturer or a streaming service to the permission question that sits at the center of this. Part two runs next Tuesday with whatever returns, including anything that makes this issue wrong.

Begin with the only genuinely old asset in the stack. Nielsen's US 6,647,548, coded and non-coded program audience measurement, carries a priority date of September 6, 1996. On an ordinary twenty-year term that lapsed on September 6, 2016.

Nothing followed. No competing currency emerged. No entrant walked through the open door. If expiring foundational patents were the mechanism that commoditizes television measurement, a decade of evidence would exist by now. What exists instead is Nielsen.

The family the theory actually means is different. US 10,467,286 and US 11,386,908, methods and apparatus to perform audio watermarking and watermark detection and extraction, both listing a priority date of October 24, 2008. Ordinary benchmark: October 24, 2028. Two years out, and still subject to patent term adjustment and terminal disclaimer before that date means anything.

Above that family sits a further stack. Gracenote's audio fingerprinting patent US 9,286,902, 2013 priority, benchmark 2033. Nielsen's timestamp transition resolution patent US 11,562,753, 2017 priority, benchmark 2037. Audio sensor selection within a metering device. Watermark detection in magnetic field data, a different technical route entirely.

Alphonso's own contribution is two published applications, 2015/0331661 and 2015/0332669, efficient audio signature generation, May 2014 priority, benchmark 2034 if granted claims trace to that date. Published applications are not granted patents, and grant status has to be verified before anyone assigns them value.

For Vizio, Inscape and Samba, the public record supports no defensible expiry claim at all. Samba describes chipset-level deployment in its marketing materials. Marketing materials are not a patent portfolio.

A patent stack does not carry a single date. Continuations, divisionals and later implementation claims mean a 1996 asset can lapse while detection, extraction, metering, identity, timing and calibration claims remain enforceable for another decade. There is no scheduled morning on which ACR enters the public domain.

Twenty years from filing. That is the entire clock. It is printed on the front of the document, for free, and it has been sitting there the whole time.

I have now watched four separate companies build a strategy around a number you can read off a webpage.

And the word they all use is durable. Durable IP position. Durable advantage. My mother has a Corningware dish from 1974 that is durable. A patent is a countdown with a lawyer attached.

Kowalczyk Ratio on this one runs about eleven to one. Eleven people who said durable. One who looked up the date.

Every version of the theory arrives eventually at the same courtroom. Anonymous Media Research Holdings, LLC v. Samsung Electronics America, Inc., Eastern District of Texas, Marshall Division, No. 2:23-cv-00439-JRG-RSP. Two patents tried, US 10,719,848 and US 10,963,911. The jury found infringement of claims 1, 5, 9 and 13 of the '848 patent and claim 1 of the '911, rejected Samsung's invalidity defenses, and awarded $78,512,999.

Two things circulating about that verdict need straightening out.

The verdict was reported on September 26, 2025. October 6 is the date of a later press release publicizing a result that had already occurred, and it has since been repeated as though it were the decision date. It is not.

A jury verdict is not a paid recovery. Whether that award became an enforceable final judgment, and whether it was modified on post-trial motions, stayed, settled or appealed, requires confirmation from the district court docket and any Federal Circuit filings. Until that check is done, the accurate description is a $78.5 million jury verdict, not money Samsung has paid.

Nor were these patents an untested target. PTAB activity preceded the trial. Reporting indicates Samsung's inter partes review efforts against the two media measurement patents were denied institution in April 2025, and that Samsung's IPR involving the '911 patent, IPR2025-00558, was terminated following a joint motion. A separate Roku petition against the '848 patent was also denied institution. Any post-verdict PTAB proceeding would require its own docket confirmation.

The more consequential correction is technical. These are not audio watermarking patents. They are directed to media usage monitoring and measurement within an automatic content recognition system, and the litigation record discusses fingerprinting, identification, matching, databases and analytics as components of that architecture. The distinction is not monitoring versus fingerprinting. It is ACR-based media measurement versus patent claims directed at embedding or detecting an audio watermark, which are different things legally and technically.

Which means the verdict shows that Samsung lost a particular patent trial over particular claims covering a media recognition and measurement system. It does not show that the foundation under audio watermarking is failing. The case circulating as proof of one is a case about the other.

Marshall, Texas.

Four people have now forwarded me this verdict like it is a death certificate. Seventy-eight and a half million dollars. Terrific. A jury in Marshall wrote a number on a form. That is not a payment. That is a wish with a case number on it.

But go read what the patents actually cover. Media use monitoring. Monitoring. Not watermarking. Not fingerprinting. Monitoring.

The entire industry is standing here holding up a receipt from a different store.

Why join ADOTAT+

Because the free version tells you what's happening. The paid version tells you what actually happened, which is a different story most weeks.

You get the sourcing, not just the headline. Every investigation runs its full reporting behind the wall, the documents, the on-record quotes, the math that made a company stop answering emails.

You get Frank uncut. Free readers get the short version. Paid readers get the version where he actually finishes the thought.

You get to read things companies asked us not to publish. We published them anyway. That's most of the point.

Ads don't buy coverage here. We run them. They don't get a vote. Your subscription is what keeps that arrangement possible.

You're not funding a hobby. You're funding the two people who chase down thirteen unanswered questions until somebody finally answers eleven of them.

$99/month, or $600/year if you'd rather stop thinking about it. Cancel whenever. We won't be offended. We'll just assume you already knew the ending.

Built from USPTO records, the filings in Anonymous Media Research Holdings v. Samsung (E.D. Tex. 2:23-cv-00439-JRG-RSP), the Texas AG's ACR actions and the 2026 settlements, Vizio's and Samsung's published viewing data policies, Nielsen's Q2 2026 Gauge, Comscore's Q2 2026 results, and interviews conducted August 18 to 31, 2026. Expiry dates are ordinary twenty-year benchmarks from listed priority dates and are approximate. The $78,512,999 figure is a jury verdict reported September 26, 2025, not a final judgment or a paid recovery; judgment, appeal and PTAB status are unconfirmed against the dockets. LG Ads and Alphonso revenue figures are company projections, and no public document splits licensing from ad sales. The app-level permission finding reflects an absence of public evidence, not proof of prohibition. Frank Kowalczyk is fictional. Howard Shimmel appeared on The ADOTAT Show in May 2026. ADOTAT has no financial relationship with any company named here. Those companies were contacted where practicable and responses will be added as they arrive.