Mike Shields asked the question on LinkedIn this week and it is the only correct question.

Why did The Trade Desk name a product after a fake Hawaiian island, and why does the sequel appear to be named after a wine cooler alternative that died in 2008.

Both answers are worse than the question.

Kokai is not a place. No Kokai in Hawaii. No Kokai beach in California. No Kokai anywhere. It only reads as Hawaiian because the company already had Koa, its AI, named after a tree, since 2018. Your ear got pre-loaded. Kokai is a vibe with a trademark filing.

The rest of the naming scheme is not mysterious at all. It is one guy with a surfboard and a map of the 101.

Solimar. Real beach. Ventura County. Ten minutes from HQ.
Ventura. Literally HQ. Also a macOS release, which apparently nobody in the room flagged.
Zuma. Malibu. Forty minutes down PCH.
Koa. Tree.
Kokai. Sounds nice.

And no, Zuma is not Zima, the Coors clear malt launched in 1993 for people who found wine coolers too much of a commitment, discontinued in 2008, briefly resurrected as a joke about itself.

It just sounds exactly like it.

Which, considering what Zuma actually is, may be the most accurate branding decision the company has made in three years.

Straight from The Trade Desk's own materials. Zuma makes the periodic table view optional rather than the default.

Take a beat with that.

The periodic table. The Programmatic Table. The signature feature. The thing that made Kokai look like a different product in a screenshot. The centerpiece of what Jeff Green called the most advanced product launch in company history.

It is now a checkbox.

Unchecked.

When VP of Product Bill Simmons walked out in March 2025, he defended it on the way through the door. Critics were only seeing the surface. Kokai, he said, was a bold, fundamental reinvention of what a DSP can be.

Seventeen months later the bold fundamental reinvention is a preference pane.

June 6, 2023. Kokai launches. Most advanced product launch in company history.

November 2024. Ventura announced. A television operating system. Named after the town the company sits in. Confidence.

December 4, 2024. TTD closes at $139.51. All-time high. That same month, the largest reorganization in company history. Those two facts share a calendar page and nobody has ever explained why.

February 12, 2025. First revenue miss ever. On the call, Green concedes Kokai rolled out slower than expected, then adds that in some cases the slowness was deliberate.

Of course it was. The rollout was slow on purpose. The stock fell thirty three percent on purpose too, presumably.

$122.23 to $81.92. One session.

July 2025. AdExchanger reports Solimar is not being deprecated any time soon, because Kokai adoption is behind and several Kokai features are not actually finished.

August 2025. Users report they can no longer create new campaigns in Solimar.

If they will not walk through your door, weld the other one shut.

August 26, 2026. Zuma ships. Periodic table optional.

There is a consolidated securities class action. Cohen Milstein got lead counsel on June 4, 2025, representing Arkansas PERS and Mississippi PERS.

Class period: November 15, 2023 through August 8, 2025.

Named defendants: Green, former CFO Laura Schenkein, Chief Strategy Officer Samantha Jacobson.

The complaint alleges billions in shareholder value gone between December 2024 and August 2025, and hundreds of millions in insider stock sales at prices plaintiffs say the Kokai story inflated.

Allegations. Unproven. Disputed by the company. Noted, lawyers.

TTD closed at $13.42 the day Zuma launched.

From $139.51.

That is ninety percent in twenty months.

At a company still throwing off a 77 percent gross margin on a 15.7 P/E. The machine works. Nothing in the engine seized. The market just stopped believing the story bolted to the outside of it.

Zuma arrives with exactly one hard figure: a 32 percent average improvement in CPA performance.

Improvement against what.

Over what window.

How many campaigns.

Which verticals.

Whose attribution model.

Compared to Kokai before Zuma, or to Solimar, or to a coin.

The release does not say. No methodology. No sample. No third party. No denominator.

A vendor graded its own forecasting engine using its own outcome data and published the mean.

Every DSP does this and it is normally unremarkable. It is slightly more remarkable when it is the exact species of claim named in your securities litigation, deployed in the release whose entire premise is that lessons have been absorbed.

Ask Koa. A conversational front door routing buyers to specialized agents. Audience building, campaign creation, insights, troubleshooting. Alpha to closed beta.

Announced in a press release and simultaneously leaked out of an internal roadmap to Adweek on purpose, that stated what their public memo said. Huge news.

The message discipline that held the Kokai narrative together for two years did not survive forty eight hours on this one.

Here is the part that made me put my coffee down.

The executive who ran Ventura TV OS was Matthew Henick, SVP of Consumer Products. Ventura was the consumer bet. An operating system. For televisions. Announced November 2024, promised for the first half of 2025, pitched as a cleaner supply chain for streaming.

The Sonos hardware tie-up went nowhere. Nothing.

In April 2026, Henick left to become the CEO of Meow Wolf.

Meow Wolf builds immersive art installations.

Its flagship is Omega Mart, a fake supermarket in Las Vegas where the products are absurdist gags and a refrigerator door opens into another dimension. It’s very Vegas, and around the corner is the largest marijuana shop in existence, and a swingers bar.

Before The Trade Desk he ran Deep Voodoo, the AI effects studio founded by the South Park guys.

To be clear, this is not a shot at the man. Running Omega Mart is a real job at a real company with real buildings and real humans buying real tickets. LA opens this year. New York in 2027. He went where the portals are.

It is the vector that is funny.

He left the television operating system for the fake grocery store, and the fake grocery store is the one with customers.

Renaming is the cheapest change money can buy. No engineering. No reorg. No admission.

Solimar became Kokai.

Kokai became Kokai Zuma.

Koa became Ask Koa.

Ventura became a guy's LinkedIn history.

Every name shows up with a beach and a promise, and every one gets graded against the previous name instead of against anything a media buyer can independently verify.

Killing the periodic table is the right call.

Listening to three years of traders telling you the thing was unusable is correct.

Late beats never.

But it is a retreat. And in literally every other industry on the planet, retreats do not get beaches named after them.

The next release is going to be called Point Dume and I want it in the record that I called it first.