The Trade Desk's Bet Against the Walled Gardens

Context

There are two ways to buy digital advertising. You go inside a walled garden, where one company owns the audience, the inventory, the auction, and the measurement, and you trust that the company’s numbers are honest. Or you buy the open internet, the everything-else of websites, apps, streaming, and audio that isn’t owned by Meta, Google, or Amazon. The Trade Desk is the largest independent demand-side platform for that second world, and its entire strategy is a wager that the second world is worth defending.

What I noticed

TTD says, loudly and constantly, that it does not own media and does not compete with its customers. That is the single most repeated line in their positioning, and it is not a slogan. It is the product.

Compare the conflict structure. When an agency buys through Google’s DV360, it is renting a buying tool from the company that also owns YouTube, owns the largest ad exchange, and runs the auction it is bidding into. Google sits on both sides of the table. The buyer can never fully know whether the tool is steering spend toward Google’s own inventory or shading the auction in Google’s favour, because the same company controls the buy side, the sell side, and the referee. I work on the data behind a DSP and an exchange, and the structural temptation to favour your own supply when you own both ends is real and constant. The only way to credibly not do it is to not own the other end.

That is TTD’s whole shape. By owning no media, it can sit purely on the buyer’s side and route spend to wherever performs, with no incentive to self-deal. Agencies and brands prefer this for the same reason advertisers prefer a neutral measurement partner: a buying tool run by a publisher is a buying tool with a thumb on the scale.

The product thinking

Neutrality is the growth engine. Every dollar that an advertiser is nervous about handing to a Google-owned or Amazon-owned buying stack is a dollar TTD can credibly ask for. The open internet is also genuinely large: connected TV alone is a structural tailwind, because streaming is fragmenting away from a single owner and CTV inventory needs an independent buyer that spans Disney, Roku, Netflix, and the long tail without favouring any of them. TTD positioned itself as the default buyer for the part of the market that is too fragmented for any one walled garden to own. The more fragmented the open web stays, the more indispensable an aggregating, neutral buy-side becomes.

But the open web has a fatal weakness the walled gardens don’t, and TTD’s leadership understands it better than anyone. The walled gardens have logged-in identity. Meta and Google know exactly who you are across sessions and devices because you sign in. The open web ran on third-party cookies, and cookies are dying. Without a durable way to recognise and target users, the open internet’s targeting and measurement degrade to noise, and advertisers flee to the walled gardens by default, not because the gardens are better but because they are the only place left where identity still works.

The UID2 gamble

This is why TTD built Unified ID 2.0 and then did the unusual thing of giving it away to an independent governance body rather than keeping it proprietary. UID2 is a shared, consented, email-based identifier the open web can use in place of the cookie. The strategic logic is sharp: TTD cannot win the open internet alone, because an identity standard only works if publishers, other DSPs, and SSPs all adopt it. A proprietary TTD identifier would have been just another walled garden, smaller than the real ones. An open standard has a shot at becoming the open web’s collective answer to logged-in identity.

So TTD did something counterintuitive. It seeded a public good it doesn’t fully control, because the value to TTD comes from the open internet remaining a viable place to spend, and the open internet is only viable if it has identity. The company’s fortunes are downstream of an industry standard succeeding, not a TTD feature succeeding. That is a rare and risky place for a public company to stand.

The risk

Here is the existential part. UID2 depends on users consenting to share a hashed email and on publishers having logged-in relationships to collect those emails in the first place. Most of the open web does not have logged-in users. The walled gardens do. So the identity solution for the open internet structurally depends on the open internet acquiring the very thing that made the walled gardens dominant, and it has to do so against browser makers (one of whom, Google, is also a competitor) who can throttle or deprecate the signals UID2 leans on.

If UID2 fails to reach critical scale, the failure mode is not dramatic. It is quiet erosion. Targeting and measurement on the open web degrade, performance-minded budgets drift to Meta, Google, Amazon, and the CTV platforms that have native identity, and TTD’s addressable market shrinks not because TTD did anything wrong but because the terrain it stands on got smaller. The walled gardens don’t have to beat TTD. They win by default if the open web can’t keep the lights on.

The PM lesson

TTD’s neutrality is both its sharpest weapon and its deepest dependency. Refusing to own media is what makes the buy side trust it, but it also means TTD cannot manufacture the one thing the open internet is missing. The most strategic product decision the company ever made was to give away the standard it needed, betting that an open coalition could out-build the closed giants. It is the right bet. It is also a bet on other people’s behaviour, which is the only kind of bet a neutral platform is ever allowed to make.