OpenX's first annual supply-side intelligence report, released on October 15, 2026, puts supply-side targeting at 51% of the connected TV deal spend on its U.S. marketplace in 2025, up from 9% in 2022, and shows third-party audience data falling from 78% to roughly 40% of the money flowing through those targeted deals. The full document goes well beyond the four figures in the company's announcement that day. It also sets out who was interviewed, who carried out the research and how every deal was sorted - and those details decide how far its conclusions travel.

In Short

OpenX runs a marketplace where websites and streaming TV apps sell ad space, and its report says advertisers increasingly use data and filters on the seller's side of that marketplace rather than only inside their own buying software. That matters to agencies, brands and publishers because it moves some decisions about which ads to buy toward the seller's platform, most of all in streaming TV. The numbers come from OpenX's own records, sorted by rules OpenX wrote, so they show what happened on OpenX rather than across the whole market.

From four headline figures to a full data set

When OpenX announced the report on October 15, PPC Land covered the four figures in the announcement: supply-side targeting, or SST, rising from 11% of spend in the fourth quarter of 2023 to 37% two years later; 17% of 2025 deals including real-world outcome measurement; CTV spend using SST growing 500% between the first half of 2022 and the first half of 2025; and SST growing about 11% a quarter in both the U.S. and EMEA since early 2025. That coverage noted the absence of a methodology, a sample size, dollar values and any detail on how interviewees were selected.

The report itself, titled "The State Of Supply-Side Intelligence: An OpenX Annual Trend Report 2026," answers several of those questions. According to OpenX, it draws on "the complete deal-level transaction record of the OpenX marketplace from January 2022 through Dec. 31, 2025," with every U.S. transaction classified under the company's own taxonomy of targeting methods. A separate section covers Europe, the Middle East and Africa with data from that region. The qualitative side rests on eight one-on-one interviews with senior buy-side executives: seven from media agencies, including two of the world's largest, and one from a Fortune 50 brand. According to the report, "all are OpenX clients." W Media Research, LLC carried out the project "in partnership with OpenX."

What the report still does not contain is a single dollar figure. It gives shares of spend and growth rates, never the size of the base.

OpenX grounds its standing to publish in scale. The company says it ranks as the fourth-largest global SSP in Jounce Media's September 2026 Supply Coverage Report, with direct access to more than 150,000 domains and apps, and claims "the market's largest independent supply-side identity graph," covering more than 275 million U.S. individuals and 120 million CTV households. When OpenX made OpenXSelect generally available on July 16, 2025, it put the same graph at 237 million U.S. users and 150 million connected television devices - different units, so the two sets cannot be compared directly.

How OpenX sorts a deal

Every number in the report depends on a classification scheme, and OpenX lays it out in detail.

Supply-side intelligence is the umbrella: capabilities within the SSP that help buyers control inventory quality, improve campaign performance and adapt their workflows. Beneath it sit three top-level buckets. Curation covers inventory-based controls, "including inventory lists and viewability targeting," which the report describes as the practices critics have in mind when they call curation bundling. Supply-side targeting covers four more advanced capabilities applied closer to the source of supply:

  • Audience targeting, using first-party or third-party data on audiences.
  • Algorithmic optimization, which uses predictive or machine learning models to determine which impressions a buyer bids on.
  • Quality, which prioritizes inventory on the strength of the ad experience, including attention and brand safety.
  • Measurement, which uses real-world outcomes such as purchases to build and validate audiences.

A third bucket, Other, holds untargeted run-of-network deals.

The rule that matters most is how mixed deals are counted. Most deals use more than one signal. "Because there is no practical way to determine how much each method contributes to a campaign's effectiveness, we classify a deal with several signals by its most advanced signal," according to OpenX. The company says it ranks methods "from most sophisticated to most basic in a natural progression," but the report does not publish that ranking. Under this rule, a deal built on a site list with a single third-party segment added counts entirely as SST. The report acknowledges that "signal layering prevents us from calculating exact spend shares for each targeting method," while maintaining that the trends remain clear.

OpenX then proposes a change in vocabulary. The industry, according to the report, should use "curation" to mean only "site or inventory bundling," and refer to any targeting method beyond that as SST. That is a narrower reading than others in circulation. Germany's BVDW, in a whitepaper published on February 26, 2026, defined curation as the targeted selection, combination and enrichment of inventory and data signals into tailored packages, a definition broad enough to take in most of what OpenX now files under SST. The IAB Tech Lab's own Curation Framework likewise bundles Seller-Defined Audiences, taxonomies, data transparency and the SupplyChain object under the curation label.

The marketplace numbers, and two versions of 2025

In 2022, nearly 90% of spend transacted through OpenX used curation alone, while SST accounted for a mid-single-digit share, according to the report. Little changed through mid-2023. From late 2023, the share of spend using SST climbed every year: 11% in Q4 2023, 19% in Q4 2024 and 37% in Q4 2025. Curation-only deals ended 2025 at 54%.

The report does not give one consistent number for the end of 2025. Its executive summary says SST reached a "32% share of deal spend," the trend chapter gives 37% for Q4 2025, and the closing section says "nearly 40%." The four quarterly values in the report's chart for 2025 average about 32%, by PPC Land's reading, which suggests the summary uses a full-year figure and the trend chapter a fourth-quarter one. The report does not say so. It also describes the Q4 split of 54% to 37% - a 17-point gap - as curation holding "a slight edge."

Absolute spend, not only share

The October 15 coverage pointed out that a doubling in share and a doubling in spend are not the same thing. The full report settles part of that. Curation-only spending kept growing in absolute terms every year, according to OpenX, while SST deals "simply grew an order of magnitude faster: 151% in 2023, 242% in 2024, and 233% in 2025."

Compounded, those rates put SST deal spend in 2025 at roughly 28.6 times its 2022 level, by PPC Land's calculation. So the "~2x annually" in the title of the report's first trend chapter fits the growth in SST's share more closely than its spend, which more than tripled in each of the last two years. The report's indexed chart, with Q1 2022 set at 100, shows SST near 5,900 by Q4 2025 and curation-only deals near 320. Combined with the starting shares, those readings imply total U.S. deal spend on OpenX roughly five times higher in Q4 2025 than in Q1 2022 - a rough estimate, inflated by the fourth quarter's usual seasonal strength.

Why the turn in late 2023? The report points to Google disabling third-party cookies for an initial 1% of Chrome users in January 2024, Apple's App Tracking Transparency framework and privacy laws in several U.S. states and the European Union. "Buyers read the writing on the wall: signal loss would only get worse, not better," the report states. The supply side, closer to publishers' first-party and context data, was the logical place to look, according to OpenX, and "all major SSPs, sensing an opportunity, shipped a curation or activation product during that period."

A mid-2026 crossover the report cannot show

"If current trendlines hold, the share of spending on deals using SST will exceed the share of deals using only curation by mid-2026," the report states. It was published in October 2026, after that point had passed, yet it contains no 2026 figures to show whether the crossover happened.

The arithmetic is demanding. Assuming other deals held at about 9% of spend, SST would have needed to rise from 37% to above 45% within two quarters - more than eight points - for the forecast to come true, by PPC Land's calculation. Between Q3 and Q4 2025, the share rose by roughly two points, according to the report's chart.

CTV: from 9% to 51%

Connected TV is where the report's numbers move most. In 2022, curation accounted for 76% of CTV spend on OpenX and SST for 9%. By 2025, according to the report, "those positions had reversed: SST rose to 51% of CTV spend while curation fell to 41%."

Here, too, the report carries two figures. A later passage says the CTV share roughly doubled "from 14% in 2023 to 29% in 2024, then jumping to 53% in 2025." The 51% appears three times in the report, the 53% once. The 2025 quarterly values in the report's CTV chart average about 51%, by PPC Land's reading.

That chart also shows something the text does not discuss. SST's share of CTV spend peaked at roughly 56% in Q3 2025 and fell to about 47% in Q4 2025, while curation's share rose from roughly 38% to about 44%. Non-CTV deals kept moving the other way, their SST share rising to about 30% by year-end. The report offers no explanation for the fourth-quarter reversal in CTV.

CTV's weight explains why it pulled the overall number up. It grew from a third or less of OpenX deal spend in 2022 to more than half for most of 2025, according to the report's chart, and its "higher share of SST relative to curation really drove SST penetration across the marketplace overall," according to OpenX.

Content objects as the engine

The report's explanation for the CTV shift is structural. Streaming has no browser cookies, device identifiers are fragmented and household-level signals vary across TV manufacturers, operating systems, streaming services and intermediaries. Buyers typically knew only which app, publisher or package an impression came from. "Two impressions from the same streaming service can represent radically different environments depending on the show, episode, genre, rating, and other content characteristics," the report states.

Content objects - metadata describing the title, genre, episode and rating of the programming around an impression - filled that gap. Content-object targeting surged in 2024 and accelerated in 2025 as buyers pushed publishers to supply the data, according to OpenX. "Virtually all of that quality spend involved content objects," the report says of CTV, and by 2025 quality accounted for 35% of all U.S. SST spending, "driven overwhelmingly by CTV." OpenX also pitches content objects as a publisher yield tool: the same inventory can be sold unreserved at a discount to price-sensitive buyers and at a premium to those who want guaranteed placement against specific titles, genres or ratings.

The wider market has been building the same plumbing. The report cites the IAB Tech Lab's OpenRTB Content Object and Content Taxonomy 3.1, which the Tech Lab announced in December 2024 with a dedicated subset of CTV genres, and OpenX says it normalizes CTV genre variations into nine categories and 132 subcategories. Gaps remain: Peer39 found in March 2026 that only 40% of CTV bid requests carry usable program-level signals. In September 2026, Wurl began writing content signals into CTV bid requests and Pixalate released a pre-bid API that adds show titles to CTV bids, while on the buy side The Trade Desk gained show-level CTV data from Gracenote on August 27, 2026.

A classification that shapes the result

Here the taxonomy carries real weight. Selecting CTV inventory by show, genre and rating can be read as a finer-grained form of inventory selection - choosing what to buy rather than whom to reach. OpenX's scheme places it in SST, under quality. Had content-object deals been classed as curation, the CTV figure, and a large part of the overall SST growth since 2024, would look very different. The report does not publish its numbers under any alternative classification, so readers cannot test how sensitive the headline figures are to that choice.

Inside SST: what buyers layered on

The mix of methods within SST changed as fast as SST itself, according to the report.

In 2022, third-party audience segments accounted for 78% of SST spend and measurement for about 15%, with quality, algorithmic optimization and first-party data marginal. Third-party segments were "established, well understood, and relatively easy to activate," according to OpenX, and Google's July 2022 decision to push its Chrome cookie deadline out to 2024 gave advertisers little reason to drop them.

In 2023, third-party data slipped to roughly 70% as quality and measurement gained, which OpenX links to outcome-based measurement and supply-path consolidation. "Once an SSP became a strategically selected supply path rather than simply another exchange in the chain, applying targeting and filtering logic at that level became increasingly attractive and practical," the report states.

In 2024, third-party data fell to just over half, quality grew substantially and first-party audience data reached its high-water mark, as buyers tested advertiser-owned identities after Chrome's 1% cookie test began in January.

By 2025, third-party data was down to around 40%, quality had risen to 35% and first-party audience data "fell sharply." The report gives two reasons. Google abandoned its plan to eliminate third-party cookies from Chrome, confirming on April 22, 2025 that it would keep them after first shifting toward a user-choice model in July 2024. And first-party activation on the supply side "was also cumbersome and difficult to scale." Buyers had found easier ways to improve outcomes "without moving sensitive customer data into another activation layer," according to OpenX.

A falling share is not falling spend. Applying the 2022 and 2025 shares to the compounded growth in SST spend, the money flowing through OpenX SST deals classed as third-party audience targeting grew roughly fifteenfold between 2022 and 2025, by PPC Land's calculation. The most-advanced-signal rule adds a further caveat: a deal that carries both a third-party segment and a content object is counted under whichever OpenX ranks higher, so part of the shift from third-party data to quality may reflect precedence in the taxonomy rather than a change in what buyers bought. The report does not say which ranks higher.

Algorithmic optimization remains a small slice of SST spend in the report's charts. OpenX nonetheless says 2026 "will likely be a breakout year for algorithmic optimization and predictive models," a claim it supports with no 2026 data. The 17% measurement figure from the October 15 announcement is not restated in the report's text.

EMEA: behind, and a quarterly figure that does not fit the chart

The shift is "not just a U.S. phenomenon," according to OpenX. SST represented only a small share of EMEA spending in 2022, began accelerating in late 2023 and 2024 and kept gaining share afterwards. "Through 2024, SST accounted for roughly the same share of spending in both regions," the report states, before U.S. adoption accelerated in 2025, leaving EMEA "roughly two to three quarters behind."

The regional chart is choppier than that summary. EMEA's SST share spiked to about 24% in Q4 2023, when the U.S. stood at about 11%, then fell back to roughly 12% by Q2 2024. EMEA ended 2025 at about 23% against 37% in the U.S., with curation still holding around 60% of EMEA spend.

The report also repeats the announcement's claim that SST grew 11% a quarter on average in both regions since Q1 2025. For the U.S., the figure matches the chart: quarterly shares of roughly 27%, 31%, 35% and 37% through 2025 imply average growth of about 11% a quarter, by PPC Land's calculation. That resolves a tension raised in the October 15 coverage, where 11% quarterly growth sat awkwardly beside a doubling of spend - the 11% describes share, not dollars. For EMEA, though, the chart puts SST at about 23% in both Q1 and Q4 2025, close to zero average share growth. The EMEA figure must measure something else, perhaps spend, but the report does not define it.

OpenX attributes Europe's lag to market structure: a high proportion of ad-free public television, stricter data rules in some markets, fragmentation across countries and broadcasters, and premium streaming inventory still sold mostly through direct or broadcaster-controlled channels. Citing Advanced Television, the report says programmatic carried 19% of European video ad delivery in H1 2025 against 30% in the U.S., and only 32% of European programmatic ad views ran on CTV in late 2024, against 88% in the U.S. It cites 44% growth in European programmatic ad views in H1 2025 as broadcasters, FAST services and other publishers opened inventory to automated buying.

Newer data points the same way. FreeWheel's 1H 2026 Video Monetization Report, which PPC Land covered on September 29, 2026, put programmatic at 22% of ad views in Europe against 34% in the U.S., and CTV at 39% of European programmatic impressions against 92% in the U.S., with European programmatic ad views again up 44% year on year. OpenX expects CTV to drive a similar SST acceleration in Europe as content metadata becomes more standardized.

Eight buyers, all OpenX clients

Four of the eight interviewees are named and quoted. According to OpenX, the buyers said they increasingly use SSPs for targeting, either reducing the importance of the demand-side platform or bypassing it altogether. Most were not rejecting it outright: they reserved it for "handling sensitive first-party data, orchestrating budgets, and controlling frequency," while moving more decision-making to the supply side.

Chris Milano, Global VP of Supply at Stagwell Media, whose quote also ran in the October 15 announcement, argued that applying data on the SSP "whenever possible" brings better data fidelity because it works closer to the publisher.

Maikel O'Hanlon, SVP, Managing Director, Programmatic at Horizon Media, called the move from DSP-side to SSP-side targeting a "greater and greater, almost default sensibility," according to the report.

Audrey Elman, Media Director at Polaris, said: "The relationship between advertisers and SSPs [has] ... changed pretty drastically. A lot of the tools that I traditionally would go to the DSP for, I'm now bypassing and going straight to an SSP."

Wade Wilkinson, Director, Platform & Technology Partnerships at Sojern, argued against the term at the center of the debate: "We need to move beyond simply using the word 'curation' because it has become too vague and means too many different things to different people. The important distinction is between simply packaging inventory and using intelligence within that packaging to make better decisions."

All four named buyers appear in the report's acknowledgements, a list of 52 contributors that also includes Tyler Romasco, OpenX's EVP, Commercial, who was quoted in the October 15 announcement. Because every interviewee is an OpenX client, the interviews describe how the company's existing customers see the shift. Buyers who tested SSP-side targeting and walked away, or never started, are not part of the sample.

The research firm

The acknowledgements also name Karsten Weide. PPC Land identified Weide as principal and chief analyst at W Media Research when it quoted him on October 5, 2026 about HUMAN Security's new ad integrity suite for platforms, which launched with OpenX as its first partner. That article noted that neither Weide's post nor his profile stated a relationship with HUMAN or OpenX. Ten days later, OpenX's report credited W Media Research, LLC with carrying out the research "in partnership with OpenX." The report does not say when that work began or what the partnership involved.

The argument OpenX is answering

The report presents itself as a response. "For the past two years, an argument has been making the rounds in the advertising industry, most audibly in the Programmatic Governance Council, that supply-side curation amounts to little more than bundling advertising inventory into resalable packages and therefore SSPs cannot be used for serious targeting and optimization," according to OpenX. "The trouble with this debate is that it rests almost entirely on assertions."

The council in question is younger than that two-year window. The IAB Tech Lab announced the Programmatic Governance Council's formation on April 21, 2026, with 13 founding members: Omnicom, WPP, Dentsu, Disney, Amazon Ads, The Trade Desk, Magnite, PubMatic, Hearst, News Corp, Yahoo, Raptive and Mediavine. OpenX was not among them. The council's first published output, Programmatic Best Practices version 1.0, released on September 16, 2026, set rules to curb duplicate bid requests, with public comment open until October 16 - the day after OpenX released its report. The report does not cite any specific council document or statement on curation.

Its request to standards bodies is direct. "For industry bodies drafting the rules of programmatic trading, the request is simple: Let the definitions follow the data," the report states. "A framework that treats a CTV deal with content-object targeting or algorithmic optimization as interchangeable with a site list accurately describes neither." And: "The market has already drawn the distinction through its budgets. Industry definitions and the policies built on them should catch up."

Whether budgets have drawn that distinction, or OpenX's taxonomy has, is the question the report leaves open. Its definitions are the ones doing the sorting.

Buyers, publishers and the cost question

For buyers, the report's conclusion is that "the supply side is quickly becoming a decisioning layer and should be evaluated as one." The question, according to OpenX, is no longer simply whether to use SST but which signals a targeted deal carries: "Does it rely on inventory selection alone, or layer audience, quality, measurement, and optimization on top? And at what cost, and with what evidence of performance?"

The report itself answers neither of the last two. It contains no data on fees, CPMs, take rates or campaign performance for any method it classifies, so a buyer weighing an SST deal against the same targeting applied in a DSP would find spend shares, but no prices. For publishers, the report argues that content metadata, quality certification and participation in measurement can make inventory more valuable, particularly in CTV, and open opportunities "such as audience extension."

Models and data running inside the SSP rather than downstream are visible elsewhere too. PubMatic launched Decision Fabric on June 1, 2026, letting DSPs, curators and algorithm companies run their own models as containers inside its auction, and Magnite set out two routes for partner AI models on August 6, 2026. These are the algorithmic optimization tools OpenX expects to take off. The report's growth figures also measure, in part, demand for products OpenX sells; the company has marketed itself as "The Intelligent SSP" since April 13, 2026.

OpenX says its second annual report, planned for Q1 2027, will analyze full-year 2026 data and cover agentic buying and algorithmic optimization. "We will publish this report annually so the industry can see whether these trendlines hold," the company states. That edition should show whether SST passed curation by mid-2026, as forecast, and whether the fourth-quarter dip in CTV was a blip or a turn.

Timeline

Summary

Who: OpenX, the independent supply-side platform that markets itself as "The Intelligent SSP," with research carried out by W Media Research, LLC and interviews with eight buy-side executives, all OpenX clients, four of them named: Chris Milano (Stagwell Media), Maikel O'Hanlon (Horizon Media), Audrey Elman (Polaris) and Wade Wilkinson (Sojern).

What: The full State Of Supply-Side Intelligence report. It shows supply-side targeting rising from a mid-single-digit share of OpenX U.S. deal spend in 2022 to 37% in Q4 2025, with SST spend growing 151%, 242% and 233% in 2023, 2024 and 2025; SST reaching 51% of CTV spend in 2025, up from 9%; and third-party audience data falling from 78% to around 40% of SST spend as content-object targeting lifted quality to 35%. The report classifies each deal by its most advanced signal, gives no dollar values, carries two different figures for both the overall and the CTV 2025 shares, and forecasts a mid-2026 crossover it does not document.

When: Released on October 15, 2026, covering transactions from January 2022 through December 31, 2025. A second edition covering 2026 is planned for Q1 2027.

Where: OpenX's U.S. marketplace, with a separate EMEA comparison in which SST's share ended 2025 at about 23% against 37% in the U.S.

Why: OpenX is answering the argument, which it associates with the Programmatic Governance Council, that supply-side curation is mere bundling, and it asks standards bodies to define curation narrowly and treat audience, quality, measurement and optimization signals as a separate category. How much of the shift the report records depends on that same classification, and the report offers no data on the fees or performance buyers should weigh.