The Coalition for Innovative Media Measurement on November 12, 2026 published a study finding that 62% of surveyed programmatic connected TV buyers move spend to another supply path when the signals they require are missing or unmet, and that only 8% say those signals are always put to work by the buying platforms once a campaign goes live.

In Short

A buyer survey from CIMM, co-sponsored by FreeWheel and OpenX, found that more CTV buyers respond to missing or unreliable auction data by taking their money down a different route than by simply paying less. That matters to streaming publishers, whose inventory can be passed over, and to advertisers, whose campaigns get harder to deliver when platforms do not use the data they asked for. CIMM's proposed fix is a voluntary minimum set of information for every trade, plus reports showing what each platform actually received and used, tested first in an industry pilot.

What 50 buyers said

The paper is titled Behind the Bidstream: The Signals that Matter: What Buyers and Sellers Need From the CTV Bidstream. It looks at the information that travels with each ad opportunity through the bidstream - the stream of auction requests that publishers and their intermediaries send to buying platforms - and asks a narrow question. When the data a buyer considers essential is absent, what happens to the money?

The answer, according to CIMM, is that it moves. Of the 50 qualified programmatic CTV buyers surveyed, 62% reported shifting spend to an alternative supply path when their signal requirements were not met. Another 54% said they broadened their targeting criteria, and 46% said they lowered their bids.

Those three figures do not add up to a single choice. The question was multi-select, according to the paper, so a buyer could report several responses at once. The most common answer, chosen by 66%, was to work with the seller or platform to resolve the problem. Some 30% said they proceed as planned and monitor performance after launch. Buyers, in other words, tend to combine responses, and fixing the problem at the source is the most frequent of them. The survey does not show the order in which they try each one.

The second headline number goes to the plumbing on the buy side. Only 8% of buyers said their required signals were always operationalized by their demand-side platforms (DSPs) once campaigns went live. According to the paper, 58% answered "often" to the same question. The gap between what a buyer specifies and what a DSP acts on is precisely the gap the study is trying to name.

Path first, data second

One finding cuts against a common assumption in the market, which holds that richer audience and content data is what wins bids. Buyers were asked to rate, on a scale of one to seven, how likely they would be to bid in four scenarios that combined detailed or limited signals with a verified or unverified supply path. A verified path meant a direct route with a documented, verified supply chain. An unverified one meant multiple intermediaries, an incomplete chain and an unverified reseller.

According to the paper, 88% of buyers rated their bid likelihood at five or above when detailed signals came through a verified path. With limited signals on a verified path, the share fell to 66%. Detailed signals on an unverified path drew just 36%. Rich data, in other words, did not compensate for a supply chain buyers could not check. The paper itself cautions that the comparison does not isolate the effect of verification or of the number of intermediaries.

The ranking question pointed the same way. Asked to rank eight factors, 48% of buyers placed supply-path transparency in their top three, ahead of scale and reach at 46%, according to the paper. And under pacing pressure, when a campaign is behind on delivery, 27% of the 48 buyers who answered the follow-up said they kept the same signal requirements regardless.

How the study was built

The survey is small, at 50 respondents. It was conducted by Christine Pierce of Ozona Insights and fielded by OvationMR during May and June 2026. The sample covered media agency, holding company, independent agency, managed service and in-house advertiser professionals with direct budget authority or significant influence over recommendations.

According to CIMM's announcement, the survey sits alongside in-depth interviews with 28 senior executives across the publisher and platform ecosystem, as well as bid-log analyses supplied by FreeWheel and OpenX. The paper describes those platform analyses as a way to identify associations between the availability of selected signals and bidding or pricing measures within each company's own dataset. It states that they do not establish how widespread the problems are across the market, nor the causal effect of adding any particular signal.

That caveat matters for reading the release. CIMM's announcement says missing or inconsistent information "can materially change" where buyers spend. The announcement does not put a dollar figure on revenue lost by publishers.

The paper was co-authored by Emily Palmer, Christine Pierce and Jon Watts, according to CIMM. Palmer chairs CIMM's Programmatic Working Group. Watts is managing director of CIMM, which describes itself as a non-partisan, pan-industry association focused on measurement, currency development and data collaboration.

The co-sponsors are not neutral bystanders. FreeWheel runs ad serving and programmatic technology used by many TV and streaming publishers, and OpenX is an independent supply-side platform. Both sell into the supply path the paper examines, and both stand to benefit if buyers place more trust in sell-side infrastructure.

Two proposals: a baseline and a receipt

CIMM's paper proposes two complementary steps.

The first is a minimum trading baseline. This would be a voluntary set of information covering inventory identity and selling authority - in plain terms, what the inventory is and who is entitled to sell it. According to CIMM, the baseline would sit apart from both the core technical requirements of a trade and the additional signals negotiated for individual campaigns. Each element would come with defined purposes, limitations and acceptable alternatives, so that participants do not throw out legitimate inventory simply because one field arrived in an unexpected form.

Selling authority is not new territory. The industry already relies on IAB Tech Lab files such as app-ads.txt and sellers.json, and on the SupplyChain object in bid requests, all of which the paper discusses. What the paper adds is an attempt to agree which pieces are the floor for trading, rather than leaving every buyer and every DSP to draw that line differently.

The second step is clearer platform reporting. Supply-side and demand-side platforms would publish comparable capability statements describing what they support. Alongside them would come reconciliation reports showing whether agreed signals were received, forwarded, valid and available for use - and, where it can be observed, whether they were actually used. A buyer who pays for a deal built on specific content or audience data would, in effect, get a receipt.

"Publishers need to understand what buyers require to assess their inventory, and buyers need confidence that platforms can support those requirements," said Jon Watts, managing director of CIMM. "When those expectations do not match, suitable inventory can miss out on demand and campaigns can become harder to deliver. This paper sets out practical ways to make requirements explicit and check what happens to agreed signals across the supply path."

A pilot judged on rejections, not price

The paper recommends testing both ideas through a voluntary industry pilot. It would bring together willing buyers, publishers, distributors, SSPs and DSPs across a limited set of biddable deals.

According to CIMM, the pilot would evaluate candidate trading baselines, compare platforms' capability statements against sampled transmission records, and look at unfilled opportunities alongside delivered inventory. That last point is easy to miss. Most programmatic reporting describes what cleared. Looking at what did not clear, and why, is where a missing signal leaves its trace.

Success would be measured through operational metrics rather than pricing alone. CIMM lists unexplained rejection rates, the speed at which problems are resolved and implementation costs as examples. No timetable, participant list or start date for the pilot was given in the announcement.

The paper also sets out practical actions for buyers and sellers to reduce uncertainty: making signal requirements explicit before trading, and checking afterwards whether delivery matched those agreements.

What the sponsors say

FreeWheel's Alex Strickland, head of programmatic supply relationships, framed the problem as one of reliability rather than volume. "In CTV, more data doesn't always mean more clarity. What matters is that buyers know which signals they can rely on, how those signals are handled across the supply path, and whether they can effectively transact against them," he said. "When supply is accurately represented and the right signals are readily available, publishers are better positioned to demonstrate the value of their inventory and compete more effectively for demand."

OpenX took a firmer line on where responsibility lies. "This research confirms what buyers tell us every day: Rich signals only matter when the path underneath them can be trusted. The sell side's job is to get that right upstream, before the auction, so buyers aren't left correcting for it downstream," said Erika Loberg, VP, global CTV and enterprise solutions at OpenX. "As more buying becomes automated, a clear, transparent foundation matters even more. Automation scales whatever it's built on, so the supply path has to be trustworthy before agents start negotiating on top of it."

The reference to agents is not idle. Agent-to-agent trading is moving from slides into live auctions. In June, a DataBeat measurement found that agentic buyers took part in 86% fewer auctions than conventional demand and cleared at a $6.13 CPM against $6.95. IAB Tech Lab has since counted 13 overlapping functions between the two rival agentic protocols, AAMP and the Ad Context Protocol. Loberg's argument is that automation multiplies whatever sits underneath it, including bid requests whose contents nobody can reconcile.

Why the signal gap keeps coming back

CIMM's study lands in a market that has spent most of 2026 trying to patch the same hole from different directions.

Content data is the most visible part of it. Gracenote's own 2026 research found that 91% of programmatic traders say missing show-level data limits their CTV campaigns, which PPC Land reported when The Trade Desk became the first DSP to take Gracenote's content IDs in August. A Peer39 figure from March 2026, cited repeatedly in PPC Land's coverage, puts the share of CTV bid requests that carry usable program-level signals at 40%.

Vendors have rushed to fill that space. FreeWheel itself launched a series-level Video Content Report with seven publishers in July, citing research that more than half of CTV advertisers would shift spend toward publishers that disclose show and series data - a finding that rhymes with CIMM's 62%. Wurl began writing content signals into CTV bid requests in September and claims more than 140 billion signal-rich impressions a month. Pixalate released a pre-bid API that adds show titles, genre, rating and channel into the Content object of OpenRTB 2.6, keyed off the app bundle and device user agent. Nano Interactive's Screen Graph merges partial records from different publishers to fill in missing episode and cast data.

Each of those products adds data to the request. None of them, on its own, tells a buyer whether the DSP read the field, kept it, or acted on it. That is the gap CIMM's reconciliation reports are designed to close, which puts the 8% figure at the center of the proposal.

Identity and authority

The baseline's focus on inventory identity and selling authority has a fraud dimension as well. In July, HUMAN Security disclosed that the NewsJunkie scheme produced up to nearly two billion invalid CTV bid requests per day per seller, spoofing device, app and IP data through server-side ad insertion. A buyer that cannot verify what an impression is, or who is authorized to sell it, has reason to discount it - which is consistent with the 36% bid-likelihood score for detailed signals on an unverified path.

Standards bodies have been tightening definitions in parallel. IAB Tech Lab published final programmatic auction definitions in June, setting out a 12-step OpenRTB workflow and 15 defined terms. In July it finalized how six CTV ad formats are signaled in bid requests, including pause, overlay and screensaver placements. On the buy side, IAS in January automated supply path optimization exclusions through its Total Visibility product.

Where the money already goes

Sell-side packaging of data has become a large share of CTV trading. OpenX's own figures show supply-side targeting rose from 9% of its US CTV deal spend in 2022 to 51% in 2025, with content objects - title, genre, episode and rating - driving much of the quality-based spend. FreeWheel's first-half 2026 data showed programmatic streaming ad views up 25% in the US and 44% in Europe, with app bundle and genre among the most common custom metadata fields in enriched supply.

When the sell side packages more of the data, the question of whether that data survives the trip through the DSP becomes more pressing. A deal sold on the strength of show-level content is only worth what the buyer's platform does with the content fields.

FreeWheel's own position is also shifting. In October, a trade newsletter reported that Comcast plans to move FreeWheel and Universal Ads to NBCUniversal as part of its spin-off. Comcast has not confirmed that plan. If it goes ahead, the ad server that competing publishers rely on would sit inside one of those publishers' parent companies, which makes neutral, comparable reporting on signal handling a more pointed issue.

CIMM's widening programmatic agenda

The bidstream paper is the latest in a run of CIMM research this year that ties measurement inputs to buying outcomes. In July, CIMM published a report finding that poor presence-of-children data wastes $590,000 of every $1 million campaign at the accuracy levels it modelled. Its Models as Masters? paper, co-written by Watts, set out six steps to stop AI-driven buying from distorting budgets, including engagement with agentic protocol developers. In August, a sports measurement paper put the local TV measurement gap at 20% to 30% of total viewing.

The common thread is that the inputs to automated systems - audience data, model outputs, bid request fields - deserve the same scrutiny as the outputs. The bidstream paper applies that logic to the most granular input of all.

What remains open

Several questions sit outside what the announcement answers. The survey's 50 respondents describe buyer behavior, not market-wide supply; the paper is explicit that its bid-log analyses show associations rather than causes. The DSP side of the market is not among the sponsors, and the reconciliation reports would require DSPs to disclose how they treat the fields they receive. And the baseline is voluntary. Would a buyer actually stop excluding inventory because a pilot says a field's alternative is acceptable?

The pilot, if it goes ahead, would be the first place to find out. Its chosen metrics - unexplained rejection rates, resolution speed, implementation cost - are deliberately unglamorous. They are also the numbers that would show whether a missing signal is a technical nuisance or a revenue problem.

Timeline

Summary

Who: The Coalition for Innovative Media Measurement (CIMM), with co-sponsors FreeWheel and OpenX. The paper was co-authored by Emily Palmer, Christine Pierce and Jon Watts, and draws on 50 programmatic CTV buyers, interviews with 28 senior executives, and bid-log analyses from the two sponsors.

What: A study titled Behind the Bidstream: The Signals that Matter found that 62% of surveyed buyers shift spend to another supply path when their signal requirements are unmet, 54% broaden targeting and 46% lower bids, while only 8% say DSPs always operationalize the signals they require. CIMM proposes a voluntary minimum trading baseline for inventory identity and selling authority, comparable platform capability statements, reconciliation reports and a voluntary industry pilot.

When: CIMM published the paper on November 12, 2026. The buyer survey was fielded in May and June 2026.

Where: The research covers programmatic connected TV trading; CIMM issued the announcement from New York, and the paper is available on CIMM's website.

Why: Buyers and sellers lack a shared view of which signals are required, which ones reach the buyer, and which ones DSPs actually use. According to CIMM, that mismatch can leave suitable inventory without demand and make campaigns harder to deliver, and the proposed baseline and reporting are meant to make requirements explicit and checkable before more of the trading is handed to automated agents.