The open internet 2025, part 2: Buyer innovations to change the ad market for good

Any healthy, competitive market requires price discovery, the point at which the demand and supply curves cross in an efficient market with transparent information. This requires sellers to describe their products well, which means providing sufficient color and sufficient honesty.

Our work to improve the digital advertising supply chain has recently accelerated, in part driven by major advances in AI.

We have made more progress cleaning up the digital ad supply chain in 2025 than in the previous five years combined.

AI has rapidly accelerated our supply path optimization, valuation, fraud detection and predictive clearing efforts. At the same time, though, these advances have hurt some resellers as spend moves from one reseller to a different reseller, seller or publisher.

OpenPath has accelerated this reallocation, which is good for the open internet but hurts the revenue of some resellers — especially those who deploy duplication and obfuscation strategies.

Disappointing those who duplicate and obfuscate is unavoidable to do the greatest good and to serve buyers.

OpenPath is a great example. As I have said repeatedly on our public quarterly earnings calls — OpenPath is both a canary in the coal mine and a stalking horse. We would prefer a world in which supply-side platforms (SSPs) optimize yield for publishers by leaving the world of optimizing to priority (aka P1, P2, etc.) and resuming the job that header bidding and Prebid had started to enable about a decade ago — advanced yield management through transparency and competition.

OpenPath has been a massive success. Even though it has had little net effect on our profitability, we essentially run it at cost, as a service to the industry. It is a success because it has measurably improved the quality and efficiency of the open internet. It provides the market with a clearer view of inventory quality and supply chain health. OpenPath volume has grown many hundreds of percentage points this year alone, even though it has met resistance in some quarters, even from publishers who stand to benefit the most from it. Until early 2024, I didn’t fully appreciate the adage that a wounded animal is more dangerous than a healthy one — at least regarding SSPs and other resellers.

Of course, there are many sellers and publishers still traumatized by the way Google and others treated them over the years. The whims of walled garden monetization have impacted hundreds of thousands of jobs and kept the open internet from thriving the way it likely will now.

Yet even so, some resellers or SSPs still insist The Trade Desk must continue to buy through their pipes, even if suboptimal, in order to sustain the status quo of today’s ecosystem. However, most ad inventory can be bought through hundreds or thousands of different combinations of resellers and sometimes directly from the seller or the publisher.

In a further effort to compare paths, we supported and promoted transaction ID (TID), just like the many industry standards I mentioned before. TID enables buyers and all demand-side platforms (DSPs) to consider and compare the viability of paths.

Up until a few weeks ago, TID was nearly ubiquitous. At our last check, it was on 59% of browser-based ads on the open internet — and a much higher percentage of what we bought, since it’s also used on CTV and audio.

This was good for buyers. This was good for premium publishers. This was very good for publishers who had tighter supply chains. This ensured we bought via paths that had reliable auctions and sent most of the proceeds to the publishers.

But a few resellers whose primary revenue comes from what would otherwise be less purchased destinations or the road not traveled at all — aka an inefficient supply path — decided to modify Prebid’s implementation of TID with a breaking change — a change that is no longer compatible with the OpenRTB specification.

Instead of TID being a unique identifier per impression, it is now a unique identifier per impression per SSP (aka reseller). Publishers were not given the option to choose between the two methods — the decision was made entirely for them. Many of the sellers and publishers have no technological way of stopping this change. Publishers have reached out to ask variations of, “How do I maintain the existing functionality and not use this new unconventional and noncompliant version of TID?”

This has generally been covered well in recent podcasts and articles. However, there are two things I’ve heard repeatedly that I view as key misunderstandings in the coverage of recent changes to TID — and I fear these positions are widely held.

First, it’s something like, “If DSPs really want this, it’s probably going to reduce revenue in some way for publishers.” And then secondly, some have even explicitly asserted, “This [the change to the definition of TID] is a middle finger to The Trade Desk.”

To clarify, TID is an industry standard that is consumed by all DSPs. It was a great addition to the OpenRTB specification, and the Prebid implementation was working perfectly. We only thrive if there is a healthy, transparent marketplace for publishers too. TID should be a key tenet of this, rather than another way to obscure it. Other buyers and even the IAB Tech Lab were surprised and shocked by this change — a change that was implemented without discussion among the community.

And this isn’t a middle finger to TTD. This is a ploy to obfuscate the supply chain, allowing sellers and publishers who focus on supply chain shenanigans to monetize better than those who focus on driving demand for quality content. This is a temporary downgrade for the open internet. But it is especially bad for the publishers who would better monetize with a quality supply chain. This isn’t about TTD. This is about publishers having the option to describe their quality content well and, in doing so, having the best chance of attracting the demand that their quality content warrants.

Premium publishers are again going to be hurt because of supposed supply-side advocates. This shift betrays the work that so many have done to make Prebid great. It betrays the OpenRTB spec. It’s not even what the best publishers want. This issue or conflict isn’t about buyers versus sellers. It’s more about resellers versus publishers.

This is a battle between different constituents of the sell side — those companies that obfuscate and duplicate versus those that do not.

It’s the latest example of how premium publishers who might otherwise receive warranted spend will instead see spend diverted to inferior publishers.

Nevertheless, markets are functioning. One of the benefits of competition and our obsession with supply chain optimization is that we’re always advancing optionality for players that also want to compete on a fair playing field.

First, we are accelerating efforts to mitigate the bad behaviors of resellers such as request duplication, illicit bid caching, secret hops, multiple hops, breaking frequency caps, ID bridging and similar tactics. Our Sincera acquisition is accelerating this work.

Second, we’re doing what we do best — building. To start, we’ve branched Prebid and created an upgraded auction (or more accurately an auction container, aka wrapper) that reflects the mechanics and standards of Prebid prior to the TID change. Going forward, we’ll maintain this branch of Prebid and we will consider merging future changes from the main or other Prebid branches. Of course, the founders of Prebid prepared for a moment like this by open sourcing the code. We’re grateful for all the work and vision of so many who have contributed to Prebid over the years. This strategic option wouldn’t be possible without their work and foresight.

On top of this Prebid foundation, we’re building an entirely new product. This product is designed to have the most fair and transparent auction the industry has ever seen. Our only aim is a fair auction.

This product will be known as OpenAds.

OpenPath is the connection to fair auctions. OpenAds is a high-integrity, transparent auction. It will come in two versions:

• It can be used by large sellers and publishers (a server-to-server enterprise solution).
• It can be used by smaller publishers who want an easier implementation — often just to put tags on pages or in their CMS (a simplified OpenAds solution).

Third, we’ll use this new branch in connection with OpenPath. The combination of OpenPath and OpenAds will be available to all OpenPath enterprise partners in October. The simplified version for smaller publishers will be available in the coming months.

Fourth, in the spirit of both Prebid and transparency we will open source the underlying auction code allowing anyone to inspect the mechanics of the OpenAds auction.

Fifth, we will allow others to bid into OpenAds with controls and technology that ensure demand is provably direct and transparent. This does not represent TTD getting into yield management or the supply side of ad tech. We’re giving data to sellers and publishers and some resellers so they can do yield management. It is possible that more yield management will be done in the future via API rather than managing uniquely modified auctions or QPS.

Sixth, we are scaling our feedback to publishers. Historically, this was a laborious process with Excel files and QBRs. We are automating all these insights and bundling this into a brand-new product called PubDesk.

PubDesk is a destination for any publisher — large or small — to get information about advertiser demand. What signals do we value? What can you do to improve revenue? Many top publishers already have access to PubDesk, and the early feedback has been fantastic. We will share more details at the Prebid Summit on Oct. 14. 

In all of this, as always, sellers and publishers don’t have to adopt anything. With these innovations, sellers and publishers will have more choices than ever, as will buyers. At the same time, buyers don’t have to buy it. We just want to maintain the power of choice and not let price discovery be hijacked by resellers who have built their business models around obfuscation and duplication.

To discuss these innovations and policy decisions, I’ll be appearing on an episode of the AdTech AdTalk podcast shortly.

More to come. I’ll try to write to this group more often.

But as I said, this is an important moment. The ecosystem needs a fair auction to be credible, and we’re willing to create that and make it available for others. The resellers who duplicate and obfuscate (and flat-out lie) will be hurt more and more as the market bends toward efficiency. Buyers are counting on this innovation, and the movement toward a more competitive, transparent market is inevitable.

Back to innovating — to strive, to seek, to find, and not to yield.

The Current is owned and operated by The Trade Desk Inc.