There is a particular species of press release that arrives in clusters, and if you have covered this industry long enough you learn to read the cluster instead of the release.

Last week The Trade Desk put out four of them in five days.

July 13. Penry Price joins the board, effective the ninth, expanding it to seven members. He chairs Compensation. He sits on Audit. Options over 54,502 shares, struck at $19.75.

July 15. Kristi Argyilan arrives as Chief Commercial Officer, taking data partnerships, identity, measurement, retail media, and governance. She reports directly to Jeff Green. She comes from Uber's ad business, and before that Albertsons Media Collective, Roundel, IPG Mediabrands. She also sits on LiveRamp's board. LiveRamp, as of this year, belongs to Publicis, which is a sentence we are going to have to come back to twice.

July 16. Vinny Rinaldi, out at Hershey, into client strategy, reporting to COO Vivek Kundra.

July 17. Ron Lamprecht as Chief Business Development Officer. A title that did not exist in June. Seven years as Director of Corporate Business Development at Amazon. Eighteen years at NBCUniversal before that, including EVP of Digital Enterprises. He also reports to Kundra.

Every one of them starts July 27. The same Monday. Not staggered, not phased, not whenever their notice periods happened to run out. One day.

The Trade Desk reports second quarter earnings on or around August 6.

I want you to hold those two dates next to each other for a second, because nobody in the trade coverage did. Ten days. Four senior executives, seated, announced, photographed, and describable in the present tense, ten days before Jeff Green walks onto a call where consensus sits at roughly $0.41 a share, flat against last year, on revenue the company itself guided to at least $750 million against a Street sitting closer to $771 million.

He told the analysts to come down. They have not fully come down. So the number is going to be flat and the guidance is going to be soft and the man is going to need something else in his hand when he opens his mouth.

Now he has four somethings. All of whom started a week and a half earlier.

That is not hiring. That is not even really a turnaround. That is set dressing, delivered to the theater ten days before curtain.

But go back to Price, because the trades wrote him up as a governance upgrade and he is nothing of the sort.

Here is what happened first. After the board resignations this spring, The Trade Desk notified NASDAQ that it was temporarily out of compliance with audit committee and compensation committee membership requirements.

Read that slowly. A publicly traded company informed the exchange that it did not have enough qualified directors to staff the committees whose entire function is to sit between the executives and the shareholders and ask uncomfortable questions about money. The audit committee. The audit committee, at the company currently disputing an audit.

And then in July a director arrives and immediately takes the Compensation chair and an Audit seat.

That is not a hire. That is a patch. Somebody had to go into that hole and Price went into it, and the industry press ran it as momentum because the press release did not include the words "regulatory non-compliance," which press releases famously never do.

Argyilan's announcement produced a 2.2 percent pop, to $19.37. Lamprecht landed at $19.12, down 77 percent year over year. As of this writing the stock sits near $18.51. Market cap around $8.7 billion, against a 52-week range of $16.98 to $91.45.

Four senior executives in five days, and the stock went down.

Sit with the market cap number, because it is genuinely difficult to hold in your head. In November 2024 this company was worth roughly $58 billion. When the Publicis memo landed in March it was around $13 billion. It is now $8.7 billion.

Eighty-five percent of the value. Twenty months.

At that level The Trade Desk trades near 6.6 times EBITDA and under 11 times earnings, which are not the multiples of a growth company having a bad stretch. Those are the multiples of a company somebody is running a model on at eleven o'clock at night with the door closed. Vista Equity and Quinti Capital have reportedly been working together on Criteo, which popped 24 percent on the news, because private equity has discovered that busted ad tech is now priced to move.

And note what everybody hired last week is getting paid in. Price's options are struck at $19.75, within three dollars of the 52-week floor. Whatever Argyilan and Lamprecht negotiated is priced off the same basement.

Which cuts two ways and you should hold both. If this company recovers, the board just bought an entire leadership team at the bottom tick, which is excellent capital allocation. If this company gets sold, the board just assembled a group of executives whose comp packages pay off in a change of control.

The org chart looks identical either way. That is not a clever line, that is the actual problem with reading this thing from the outside.

Now the part that reorganizes all of it.

An investment guru in Jeff Green's orbit who hung out with him recently told kme that he is worried about The Trade Desk.

And that it is the first time. Ever.

You have to understand how strange that is to type, because Jeff Green being worried is not like other CEOs being worried. Other CEOs are worried professionally. It's in the job description, somewhere between the earnings call and the offsite. Jeff Green built an entire public identity on a foundation of not being worried, and he built it in front of witnesses, on video, at every conference that would give him a microphone, for roughly a decade.

This is a man who told analysts he did not think Amazon had a DSP as he would define one. Twice. This is a man who stood at CES on January 7, 2026 and told the industry that 2026 would be the open internet's strongest year yet.

The Publicis memo landed seventy days later.

This is also a man who, in March, put roughly $148 million of his own money into open-market purchases of his own stock at around $24 a share, then published a letter on the trade publication he owns explaining that he was putting his money where his mouth was, that he had never been more certain the approach was right, and that this was the biggest purchase of his life. He was correct that it was the biggest purchase of his life. He was also, per his own telling, informed afterward that it ranked among the three largest insider buys ever recorded.

Thirteen days later, the memo dropped.

So when people around him say he's worried, the fact is not "a CEO is stressed." Every CEO is stressed, that's what the equity is for. The fact is that a specific man with a decade-long documented allergy to public doubt has apparently, privately, developed some.

And to understand why that matters, you have to understand what he was never worried about. Which requires going back considerably further than a bad quarter.

That is not a metaphor and it is not a sentiment. It is a logistics network. Bishop's storehouses. Welfare farms. Canneries. Deseret Industries, the church's own brand with its own label and its own distribution, built during the Depression so that members in need could eat without going to the government and without shame. The bishop assesses the need. The storehouse fills it. The food comes from the congregation, from people who tithed precisely so it would be there.

Jeff Green grew up in a house where the system was needed.

A single mother. Not much money. The cupboards stocked, according to people who knew the family then, with Deseret brand food. Which is to say the congregation was feeding this family. Which is to say the church was not an abstraction in that house, not a Sunday obligation or a social community at a comfortable remove. It was the reason there was dinner.

He mentions the childhood the way self-made men mention the parts of their past that explain too much: a throwaway line about wanting to play in the NBA, a laugh, then straight to the part where he wins. The poverty sits in his public biography the way a scar sits on a face, visible if you know to look, unremarked upon by the man who stopped noticing it decades ago.

Here is what that upbringing installed, and it matters more than any of the ad tech.

The thirteenth article of faith gets memorized by LDS children before they are old enough to parse it: we believe in being honest, true, chaste, benevolent, virtuous. Honest is first. Not aspirationally. As an identity claim. And church teaching is unusually specific about where it applies. Members are to be people of truth and integrity in all their words, dealings, business arrangements and everything else.

There is no carve-out in there for the programmatic supply chain. No footnote exempting fees charged on top of fees when the contract says cost plus nothing. That is what makes it a covenant rather than a mission statement.

A boy who grew up eating church welfare groceries did not learn the obligation of the powerful to the powerless from a sermon. He learned it from his own cupboards.

He went to BYU, which is not just picking a school, it is choosing as an adult to go further into the institution when you were finally old enough to go somewhere else. Then two years on a mission in the California Ventura Mission, knocking on doors in Thousand Oaks and Newbury Park and the Conejo Valley, practicing the discipline of walking up to strangers and telling them you have something true and important to say and meaning it.

He lives in Thousand Oaks now. The man came back to the exact place where he learned to knock on doors.

So when he built the company, he built the theology into it, and then he did something almost nobody does. He put it in a legal document.

September 2016, The Trade Desk's prospectus to the SEC. It contained a section heading, in capital letters:

We Are a Clear Box, Not a Black Box.

Beneath it, the promise: clients see their inventory costs, their data costs, the platform fee, detailed performance metrics. All of it visible. All of it disclosed.

That was the covenant. Written down. Filed with the federal government. And Wall Street priced it, because the transparency brand carried a multiple and the conviction carried a market cap.

He meant it. That is the most important sentence anyone can write about this story, and if you skip it you get the story wrong. He meant it in 2016 and he meant it on every earnings call after, in language that sounded less like guidance and more like liturgy.

But here is the mechanical problem with neutrality, which nobody thinks about until an auditor does. Neutrality is not a feeling. It has exactly one operational proof, and that proof is itemization. You cannot audit an incentive. You cannot inspect a motive. You can only read the bill.

Google hands you a number and asks for trust. Amazon hands you a number and asks for trust. The Trade Desk handed you thirty pages and asked for nothing.

The invoice was not the billing department getting carried away. The invoice was the clear box. It was the artifact. It was the only thing this company had that the walled gardens structurally could not copy, no matter what they spent.

An agency executive told Business Insider in 2019: you have always been able to trust what Jeff is saying, you are not getting one thing in the room and another outside of the room.

That executive was not wrong. In 2019 it was true.

Here is the part nobody has assembled, and it is the whole explanation.

December 20, 2021. Green writes to church president Russell M. Nelson formally resigning from the Church of Jesus Christ of Latter-day Saints. He is worth roughly five billion dollars. The stock is near its peak. The beat streak is intact. He picks the absolute apex of his life to sever the institution that made him. Eleven family members resign alongside him. He donates $600,000 to LGBTQ groups the same week. The letter is precisely constructed, morally framed, publicly released. It says the church is actively and currently doing harm.

September 29, 2023. Dave Pickles leaves. Co-founder. Fourteen years. The engineer who built the architecture underneath the mission. Framed as retirement, long-term advisor, gracious language, warm statement.

Pickles is the part of this story that never gets told correctly, because founder mythology only has room for one name. He was the other half of the theology. He said, back when the company was small enough to still be honest with itself, that if you talk the talk about it but then you don't do it, that's actually the worst, because everybody knows.

He built UID2 open source, with independent governance, specifically so that in his own words neither us nor anyone else could turn it into a proprietary land grab. And then, almost tossed off: you're going to be subject to audits. That's a good thing.

Three years later FirmDecisions asked The Trade Desk for the billing data. The Trade Desk said no.

Pickles also once described what Green gave him in the brutal early years, when the technology didn't work and the dark stretches were frequent. His partner pushed him to throw things in the garbage and told him to stop and think about it differently. Just having somebody tell you to stop, he said, was super helpful.

By 2024 there was nobody left in the building to return the favor.

And in roughly the same window, the marriage ended.

Now read those three losses together, because separately they are biography and together they are a mechanism. The church, which had bishops and covenants and the explicit threat of divine judgment for dishonesty in your dealings. The co-founder, who had fourteen years of standing and the technical credibility to ask whether the thing did what they said it did. The marriage, which is the last structure on earth with an unqualified claim on who you are.

All three, gone, inside about twenty-four months.

And then, on September 18, 2025, shareholders voted to cement Jeff Green's 48 percent voting control.

A man with 48 percent of the vote who is also chairman of the board cannot be removed by institutional shareholders. He cannot be outvoted. He cannot be managed. Every external accountability structure came out, and the internal one got welded shut behind him.

The tradition he was formed in had a bishop.

The company that replaced it had LinkedIn.

Gary Mittman, who hired him at twenty-three at NAMI Media and watched the next thirty years from a respectful distance, described the young Jeff Green with real affection and one qualifier. He was focused. He was determined. There was, Mittman allowed, a certain element of arrogance, a little bit. He was headstrong about what he believed was the right way to do things.

"He was headstrong," Mittman said, "but he was usually right."

That sentence is a compliment in 2001. It is a diagnosis in 2026, in a building with no bishop, no co-founder, and 48 percent of the vote.

Which brings us to this spring, and to a sentence that should have been the biggest story in this industry and instead sank without a ripple.

Asked about transparency on a podcast, Green volunteered that The Trade Desk has at times been dogmatic about it. Way too open, he said. Past the point of usefulness.

His example was the invoice. Facebook sends one line. The Trade Desk sends thirty pages. It's like finishing a meal, he said, and being handed a list of every ingredient you consumed. And then you debate the price of the croutons.

What you're actually trying to create, he explained, is understanding. He has a name for the new approach now.

Practical transparency.

Now put the calendar next to it, because the calendar is unkind.

FirmDecisions, the independent auditor Publicis retained, found that the DSP fee had been charged on top of other fees. Not instead of. On top of. And that clients had been auto-enrolled into fee-bearing tools they never requested and never authorized. When the auditor asked for the data proving media costs were invoiced at cost without markup, as the contract required, The Trade Desk declined, citing confidentiality. TTD proposed alternatives. Publicis reviewed every one and concluded, in language that will outlive everyone involved in this dispute:

"None of the options proposed by The Trade Desk resolved the issues raised by the audit."

The day before that memo went to clients, Green posted on LinkedIn that the company had never failed an audit, ever, and that it would not disclose the bills of all its clients and partners to one of them merely because they assert ambiguous audit rights.

The S-1 said clear box. The LinkedIn post said we will not disclose the bills. Both documents are public. Both are permanent. The distance between them is not ambiguous.

And here is where the easy version of this story is wrong, and where I want to be careful, because the easy version is that Green is a hypocrite who got caught.

He is doing something considerably more interesting than that.

Look at what the survival path for this company actually requires, and it is not a secret, it is in every strategic analysis anyone has written in the past year. To outlive the take-rate model, The Trade Desk has to own identity and first-party data infrastructure. It has to own AI decisioning and measurement. It has to buy the pieces it cannot build. And it has to move off percentage-of-spend billing toward outcome-based or license-style pricing before the percentage collapses on its own.

You do not itemize a license fee.

The thirty-page invoice is an artifact of charging a percentage of media. It is the receipt for a toll booth. If you are migrating to a model where clients pay for AI decisioning and identity infrastructure and measurement, the thirty pages stop being proof and start being a liability, because every line on them invites a conversation about a business you are trying to exit.

So he is not abandoning transparency. He is abandoning the billing model that transparency was attached to. The covenant was collateral damage.

Which, if you squint, is the correct strategic call. The toll booth is closing. He may well have seen it before anyone else.

It is also, executed in this exact order, completely indistinguishable from a company quietly getting itself ready to be sold. Nobody outside that boardroom can tell which one is happening.

Possibly including the board.

Seven people who mattered have left this company in thirty months.

Dave Pickles, co-founder, fourteen years. Jed Dederick, Chief Revenue Officer, thirteen years. Laura Schenkein, Chief Financial Officer, ten years. Gokul Rajaram, board member, gone two weeks before the memo. Ian Colley, Chief Marketing Officer, seven years, who defended the combative posture publicly on the day the memo dropped and was gone within three weeks. Melinda Zurich, top communications executive. Matthew Henick, who ran Ventura and turned up as CEO of a drug-induced immersive art company in Santa Fe, which is a career move I frankly respect.

Fifty to sixty years of combined institutional memory. People who sat in the rooms where the billing architecture got decided.

Every one of them has said nothing publicly.

Not a farewell post with a subtweet in it. Not an anonymous quote that everybody could identify. Not a podcast where they get seventy percent of the way to the thing and stop. Nothing. Colley did not even announce his own departure; he confirmed he was no longer CMO after journalists kept asking.

In the tradition Jeff Green was formed in, the person who leaves and says nothing is not the person with nothing to say. They are the person who has made a decision about what belongs to them to disclose.

Once is a person. Twice is a coincidence.

Seven times is a pattern, and a pattern is information.

In January 2026, at the top of the year, seventy days before FirmDecisions' findings reached Publicis clients, Jeff Green was asked about Google's antitrust collapse.

Here is what he said, word for word.

"After decades of winning and having a winning hand, they cheated and got caught. And now they have to blow it up. So unless we're stupid, we wouldn't replicate that."

He was talking about Google.

Unless we're stupid.

How we reported this. Litigation, governance, and personnel facts come from SEC filings, court records, and NASDAQ compliance disclosures rather than plaintiff or company press releases. Green's quotes are all on the record: earnings calls, conference remarks, LinkedIn, podcasts.

Insider sale figures are not findings but accusations, and a denied motion to dismiss decides nothing on the merits. Dentsu and WPP's OpenPath exits were reported by Adweek, February 19, 2026.

Amazon performance comparisons come from third-party validated agency tests; results vary by vertical. Take rate, revenue concentration, retention, and principal media figures are industry estimates, not company disclosures. The Trade Desk disputes the FirmDecisions findings and says it has never failed an audit; KPMG's audit for Omnicom found no issues. Nothing in the public record indicates a sell-side advisor has been retained. Biographical material is drawn from The Transparency Sheriff, my forthcoming unauthorized biography, which Green did not cooperate with.

The characterization of his private state of mind is single-sourced, uncorroborated at publication, and was put to the company for comment. Figures as of July 20, 2026.

I've been told a lot lately that I crashed The Trade Desk's stock. I don't own it. I don't short it. I've never traded on my own reporting. The stock is a number. I note it, the way you note a fever. It tells you the body is fighting something. It doesn't tell you what.

I care whether the thing is true.

The response has been almost entirely about Publicis, which is the least interesting thing in the report. Publicis was third. The federal ruling two days before the memo went nearly uncovered. Amazon made the fee look expensive, but the exchanges made it look unnecessary, and only one of those ends a business model. Everyone is quoting retention as a leading indicator when half the revenue is locked into deals signed before any of this happened.

Nobody is modeling the renewals. That's the number.

Behind the paywall is the full report, updated for July 2026. It is probably the most in-depth thing written on this company anywhere, which is why dozens of investment firms pay monthly to read it. I don't write it for them. I write it because the docket was sitting there and nobody else opened it.

ADOTAT+.