Almost every company in this issue told you what it was this week, and almost every one got contradicted by its own paperwork. Fox said it bought Roku's ad expertise, then lost Roku's ad chief the morning the deal closed. Netflix told buyers it was becoming "formidable", and its ad sales chief is now expected to leave. Netflix called two games "top tier" and shut both studios within a month. Perion called itself "an advanced technology leader" and sold its operating business for about $192 million. Fox Sports says college football viewing is up on a game count Nielsen says went down. The Trade Desk called its first revenue decline "meaningful progress." WPP lectured rivals about skipping pitches, OpenAI talked about trust while a study priced its ad visits at a third of Google's, and one company, American Express, actually checked out.

A company's description of itself is a press release. The filing, the org chart and the count are the story. This week's File is Netflix: what's documented, what's single-sourced, and what the engagement numbers say about the ad miss. ADOTAT+ members get two more behind the wall: The Trade Desk's third quarter, line by line, and what Nicolle Pangis's expected exit means for everyone who buys Netflix.

🎧 Listen: The Weekly Reckoning, Episode 30. Every story in this issue in about 14 minutes, starting with Netflix.

On June 15, Lachlan Murdoch told analysts Roku had "unique expertise in performance marketing." The deal closed Friday, Nov. 6. The same morning, a staff memo from Paul Cheesbrough and Anthony Wood said Charlie Collier, president of Roku Media and the man who ran Roku's ad business for four years, would not be making the trip. Neither would general counsel Chris Handman. Fox cleared out two of its own while it was at it: Tubi CEO Anjali Sud and AdRise's Robyn Peterson. The new Roku CEO is Mustafa Ozgen, who ran devices and technology. The ad business Murdoch praised now reports to the executive who built the hardware.

Roku's own filings show what the expertise was producing. Third-quarter ad revenue rose 23% to $727.4 million, but Roku delivered 33% more video impressions and earned 7% less for each one. Over nine months, impressions rose 42% and price per impression fell 11%. One person familiar with Roku's ad business said its sellers "took whatever" they could get. People at Fox tell ADOTAT Roku's PR and marketing team will likely be cut, which Fox has not confirmed. Fox also inherited a pitch to drugmakers built on what the TV saw you watch, and Roku and DeepIntent didn't answer our questions about it. Murdoch paid for expertise. The org chart says he kept the real estate.

Six months ago, Netflix told Madison Avenue it was done proving it was a durable player and was becoming "a formidable one." The executive who sold that to U.S. and Canadian buyers, VP of advertising Nicolle Pangis, is now expected to leave, according to multiple people familiar with the matter. One person told ADOTAT The Trade Desk is trying to recruit her to run its CTV sales and wants to use her to push Netflix to name it exclusive. That's one person's account, not settled fact. Nobody we spoke to blames her. A longtime ad tech executive called her "a calming influence who helped steady the team" and pointed at the org chart instead: an ad product organization that does not report to sales. This week's File, below, has what changed since "formidable."

The Trade Desk reported third-quarter revenue of $736 million on Monday, down 0.5%, its first year-over-year decline since it went public in 2016. Jeff Green called it "meaningful progress." It cleared the floor the company set for itself in August by $86 million, and shares jumped. What the beat was actually made of, including a one-time partner deal and an accounting change the company won't size, is in the ADOTAT+ File below the wall.

On July 16, co-CEO Greg Peters told investors FIFA World Cup: Launch Edition and Unhinged had "really solid numbers that put it in the top tier." On July 31, Refactor Games, which built the FIFA game, laid off about 85% of its staff, a cut its part-owner Delphi Interactive made, not Netflix. On Aug. 13, 28 days after the call, Netflix closed Night School Studio, which made Unhinged, and Moonloot. In October the Wall Street Journal reported an internal memo: only 3.5% of subscribers play Netflix games in a month. The Q3 shareholder letter didn't mention games at all, except a GTA VI sneak peek.

Here's the part the gaming press skipped: a Netflix game has never carried a single ad. Every minute spent playing produced zero impressions, at the same time Netflix cut its ad guidance and Nielsen showed its share of U.S. TV viewing slipping from 8.8% to 7.8% while YouTube hit 14.2%. Netflix kept the party and kids' games, the ones you play with the TV already on. It didn't learn that people won't play. It learned that playing doesn't move the number it reports.

Affle is paying $11.70 a share, about $420 million, for a company holding $228.3 million in cash and no debt. Back out the cash and Affle pays about $192 million for the operating business, roughly 3.7 times the EBITDA guidance Perion has since withdrawn. First-half adjusted EBITDA was $3.2 million. Perion spent roughly $685 million on acquisitions and its own stock trying to become something else. Craig-Hallum cut the stock to Hold at exactly $11.70, and the consensus target still sits above the deal price. The analysts think Perion is worth more than its own board accepted.

The family tree explains the price. SweetIM's browser add-ons, which Malwarebytes still flags as "PUP.Optional.Perion." Conduit's toolbars. Undertone, which by its own account invented the pop-under. Content IQ's roughly 80 clickbait sites, shut down in under two days once Sincera looked. Then Bing fell from 35% of revenue to under 5%. Perion still pays out 57% of revenue in traffic acquisition costs. And its merger filing shows Affle and Perion under NDA on Feb. 9, nearly nine months before the announcement and a month before Canaccord, later Perion's sell-side banker, put a $14 target on the stock. Affle didn't overpay for a product. It declined to pay for one.

Nielsen has college football's average audience down 11% through Week 9, at 1.6 million viewers, across 402 games against 410 last year. Fox Sports' Michael Mulvihill says total viewing is up 2%, because more games are on national TV and the higher volume "negates the lower average viewership." Multiply Nielsen's own numbers and total viewing should be down roughly 12.7%. For both figures to describe the same games, broadcasts would have to run about 17% longer in a single season. Nobody has said that happened.

The likely answer: Fox is counting a different slice. The big five networks aired 131 games through Week 8, and in The Athletic's numbers Fox is the only one of the five that's up. Meanwhile Nielsen's new co-viewing methodology lifted baselines by about 3.4%, which puts the sport closer to down 14% on last year's ruler, and the middle tier the 24-team playoff is meant to sell fell as much as 22.7%. Fox Sports, Nielsen and the College Football Playoff were sent questions Nov. 9. The only network that went up is the one doing the arithmetic.

On WPP's third-quarter call, CEO Cindy Rose took on what she called "this pitchless CEO handshake thing." Agencies winning accounts CEO to CEO, around the CMO and without a review, don't do the industry any favors, she told investors. She didn't name Publicis. She didn't have to: PepsiCo moved to Publicis without a pitch, following Microsoft in April and LVMH's Asia-Pacific account in September.

Publicis answered with an open letter on Instagram. Claiming it wins on a handshake, the letter said, is "an insult to clients' intelligence." Then it went for the one pitch WPP chose not to enter, Coca-Cola's North American media review: "it's not an endurance problem; it's a capability one." The caption read, "we're not angry, we're just disappointed," and the letter was signed "Your friends at Publicis." The most-liked replies backed neither side. Both holding companies say clients are smart. Neither one asked a client.

OpenAI's head of monetization, Benji Shomair, said Monday that trust guides its ads business, and that partners sometimes want things OpenAI "just can't" do. The same week, ChatGPT's new "visual ads" started showing up in the gap while it generates an image you asked for, and OpenAI opened a self-serve Business Hub in beta for brands to see how people engage with them inside ChatGPT.

The receipts are less inspiring. A working paper by Maximilian Kaiser and Christian Schulze, still under review, found a paid ChatGPT visit earned $0.89 in revenue per session in U.S. traffic, against $2.45 for a paid Google Search visit. Paid ChatGPT traffic converted at 0.37%, against 1.17%. Among 87 U.S. sites buying both, 97% got less revenue per visit from ChatGPT. By the authors' math, a ChatGPT click is worth about 26 cents against a typical 72-cent search click. Trust is the pitch. The receipts say the click is worth about a third.

American Express named Jess Ling chief marketing officer, effective Nov. 30, and credited her with bringing greater discipline to media spending. ADOTAT checks words like that for a living, and the first numbers looked bad: marketing expense rose 11% last quarter to $1.77 billion, in the same quarter Amex booked a $336 million restructuring charge to cut about 5% of positions. Amex spent $6.3 billion on marketing in 2025, up about 75% since 2019.

Then the rest of the filings came in. Amex moved money away from cheap cash-back cards toward fee-paying ones: 76% of new accounts last quarter were on fee-paying products, net card fees rose 17%, and after Platinum's fee went up 29% to $895, the card kept 99% of its spending. Ling spent 17 years in B2B and revenue marketing before she ran brand. At Amex, discipline doesn't mean spending less. It means spending more and being able to say what came back. Amex didn't promote a poet. It promoted an accountant with taste.

Read together, this is one story told ten times. Fox bought a sales team's reputation and kept the hallway. Netflix sold scarcity and built pipes that sell it cheaper. The Trade Desk set a floor and called clearing it progress. Perion bought identities for ten years and got priced on the oldest one. Affle buys names. Fox Sports picks the denominator. WPP and Publicis argue about which of them respects clients more. OpenAI sells trust at a third of the price. Nobody lied in a press release this week. They just stopped reading their own filings at the good part. The one exception is American Express, where the numbers say what the press release says. Apparently it can be done.

In April, Netflix said its ad business was on track for $3 billion in 2026, roughly double 2025, with more than 4,000 advertisers and the ad tier driving about 60% of sign-ups where it's offered. The Wall Street Journal has reported Netflix is aiming for about $9 billion in global ad sales by 2030. That's what "formidable" meant: not a nice place to test streaming, but one of the biggest ad businesses in television.

This File puts what ADOTAT reported this week next to Netflix's own numbers. It separates what's on the record from what one person told us.

In October, Netflix cut the $3 billion target to about $2.7 billion. Its own explanation was delays in product rollouts and partner integrations, not a sales collapse. Peters called the company "a bit behind our aggressive target for the year." Ads are still growing more than 80% this year by Netflix's own outlook. But "on track for $3 billion" became "about $2.7 billion" inside a single year, and the third-quarter shareholder letter gave no standalone ad revenue figure at all. The cut is the only ad number buyers got.

Ad revenue rides on viewing hours, and Netflix's viewing hours have flattened. Global engagement grew about 2% in the first half of 2026, on roughly 97 billion hours viewed. In July, Netflix discontinued its twice-yearly What We Watched engagement report, and the Wall Street Journal says it will now report engagement once a year. When the number stops flattering you, you publish it less often.

Meanwhile the supply keeps coming. The ad tier launches in 15 more countries in 2027, starting in March with nine in Europe, the Middle East and Africa plus Colombia, Peru and New Zealand, with ads in video podcasts and vertical video on the way. More inventory, flat hours, falling prices. That's not a formula for "formidable."

Netflix launched ads in 2022 at about $65 CPM. By 2023 it was roughly $39 to $45. By 2024, advertisers were telling Adweek some inventory would go for $20 to $30. By this spring, reports put Netflix in the low $20s, competitive with ordinary CTV, with roughly half of non-live ad revenue running through third-party DSPs. ADOTAT laid out the spread in February: $45 to $65 direct, $20 to $30 through DSPs. That spread taught buyers that "premium" was negotiable.

Jack Myers put it in one line in The Myers Report this week: "Netflix sits in the most expensive position in the CTV market: the middle." In his reading, Netflix priced like a scarce network before it had scale, then cut price as supply grew, and never set a pricing anchor buyers accepted.

Netflix built the discount on purpose. It launched with Microsoft as its ad tech and sales partner, said in May 2024 it would build its own, and built the Netflix Ads Suite, with Magnite as the SSP and pipes out to The Trade Desk, Google's DV360, Yahoo and, since September 2025, Amazon's DSP, a deal that sent Trade Desk shares sliding. Pangis publicly framed programmatic as "a pipe to get to Netflix content", the same as a direct deal. Each pipe made Netflix easier to buy. Each one also made it easier to buy cheaper.

On Monday, The Trade Desk's third-quarter release listed Netflix first among its business highlights. Starting in November, advertisers can use first- and third-party data to target Netflix audiences through Kokai, with in-platform measurement "expected" later and no date given. That's the same release that reported The Trade Desk's first revenue decline as a public company. Read it next to what one person told us: The Trade Desk wants Netflix to name it exclusive. An exclusive would undo the multi-DSP strategy Pangis was hired to run, and hand Jeff Green back the claim Amazon took from him.

Jeremi Gorman, who built the ad business, left in October 2023. Peter Naylor, the sales chief who bridged Netflix's technologists and its marketers, left in 2024. Julie DeTraglia, who ran measurement, left in 2025. Jon Whitticom, the VP of ads product who helped build the Ads Suite, left in August, and the forecast cut that followed blamed product delays. On Nov. 2, Netflix hired Darshan Kantak from Google as VP of its ads platform. He reports to Amy Reinhard alongside Pangis, not to her. That's the separated org chart our source described, with a new name in the other box.

Count them: the original ad president, the original sales chief, the measurement chief, the product chief, and now, by multiple accounts, the replacement sales chief. In four years. Every exit came with a gracious statement. The constant is the structure.

Netflix's games retreat and its ad problem are one decision seen from two sides. Netflix reports viewing hours and sells ad impressions. Games produced neither. They ran at 3.5% monthly reach by the internal memo the Journal reported, about 10% short of Netflix's own target, and never carried an ad. When the hours flattened and the ad guidance came down, anything that took a subscriber off the show and put nothing in the ad server became a leak. Netflix plugged it.

On the record: the forecast cut and Netflix's reason for it, the engagement figures, the DSP partners, the executive departures, Kantak's reporting line, the games closures, and The Trade Desk's Netflix language in its own release. From multiple people: Pangis is expected to leave. From one person: The Trade Desk is trying to recruit her to run CTV sales, and wants her to push Netflix toward exclusivity. Not confirmed: whether Pangis has talked to The Trade Desk, what title is on the table, or whether any exclusive pitch has reached Netflix. Netflix and The Trade Desk have not answered our questions. The CPM figures come from published reports, because Netflix doesn't publish its prices.

Ask who owns your Netflix account on Jan. 1, and whether the commitments in your upfront survive a sales leadership change. Ask whether Netflix is discussing exclusivity with any DSP, and what happens to deals you buy through Amazon or DV360 if it is. Ask for a date, in writing, for the in-platform measurement The Trade Desk says is coming. Ask what Kantak's arrival changes on the product roadmap your sales rep has been selling. Ask how many hours your audience actually watched, since Netflix now reports that once a year. And ask for the direct price and the programmatic price side by side, because the gap between them is the real rate card.

One more data point. On Nov. 10, Netflix launched a new brand campaign, "Find The Thing You Never Knew You Needed," built to get subscribers watching again. iSpot estimates Netflix spent about $105 million on national TV ads in the past year. The company that sells ads against viewing hours is now buying ads to get the hours.

Netflix promised buyers formidable. What it delivered was available. Those are both worth money, but they are not worth the same money, and the person who had to explain the difference is the one expected to leave.

This issue draws on ADOTAT's reporting published Nov. 9 to Nov. 12 and the sources in each piece, plus company releases, filings and calls, trade coverage of WPP, Publicis and OpenAI, the Kaiser and Schulze working paper (still under review), and iSpot data reported by MediaPost. Pangis's expected departure comes from multiple people and The Trade Desk recruitment from one person; Netflix has confirmed neither. The Roku PR and marketing cuts are expected, not confirmed. The Perion net price and the college football figures are ADOTAT's arithmetic. Refactor's layoffs were attributed to Delphi Interactive, not Netflix, and nothing here alleges Affle took part in Discover-Tech's spyware activity. Roku, DeepIntent, Fox Sports, Nielsen, the College Football Playoff, Netflix and The Trade Desk were sent questions before the underlying stories ran; any answer will run as a dated update. Pesach Lattin has done contract work for Fox5 in Las Vegas and has said publicly that Charlie Collier should be replaced, founded and sold the ad game network Future Games Network, ran a competing ad network in the pop-under era, and carries an American Express business card; Jack Myers is a longtime ADOTAT reader; sponsor Troutman Amin had no say over editorial. ADOTAT is reader-supported. No paid placement in editorial.

The Trade Desk raised its buyback to $1 billion. It didn't buy a single share last quarter. ADOTAT+ members get the second File today: The Trade Desk's first revenue decline as a public company, read line by line against its own guidance, its own call and its own cash.

The one-time partner deal inside the quarter. The accounting change that moves costs out of revenue, which the company won't size. The $1.65 billion it sat on. Plus a third File: who gains leverage when Netflix's sales chair moves, the conflict nobody has priced if The Trade Desk hires her, and the one reporting-line change that would tell you Netflix finally fixed this. Upgrade to ADOTAT+ to read both.