The Coalition for Innovative Media Measurement today published a paper arguing that sports audience measurement has stopped being a reporting function and started behaving like market infrastructure. Its most concrete number is not about sports at all: as much as 20-30% of local television viewing may go uncounted in current syndicated systems, a gap the coalition describes as a threat to the local broadcasting model for sports.

The paper, titled "Keeping the Score: Industry Perspectives on Innovations in Sports Media Measurement," was released to the industry on August 25, 2026. It identifies nine persistent measurement challenges and catalogues the technologies and governance arrangements emerging to address them. The document synthesises two CIMM Sports Measurement Innovation Showcases, the first held virtually in the second quarter of 2025 and the second an in-person gathering at Google headquarters in New York in the fourth quarter of 2025, supplemented by stakeholder interviews.

CIMM also confirmed that it will convene members on November 12 for its second annual Sports Media and Measurement Summit, an in-person event at the Empire State Building in New York, held in the building's private basketball court.

"The challenge is no longer simply counting viewers," said Jon Watts, Managing Director at CIMM, in the announcement. Watts added that "sports is driving the future of the entire ecosystem," because financial stakes are high and distribution is fragmented.

Nine gaps, three structural tensions

According to CIMM, the operational problems facing sports measurement reduce to three deeper tensions: fragmented audiences, a broader range of measurable engagement, and greater commercial accountability. The first covers cross-platform aggregation and deduplication, unauthorised streaming, out-of-home and co-viewing, local and regional viewing, and multicultural representation. The second covers second screens, social amplification, sponsorship and new ad formats. The third covers business outcomes and next-day decision-making.

The nine challenges listed in the paper's Exhibit 6 are: aggregating viewing in a fragmented ecosystem; unauthorised streaming and audience visibility; measuring second screen usage; capturing out-of-home and co-viewing consumption; assessing the value of sports sponsorships; measuring mid and lower funnel outcomes; measuring local and regional sports viewing; next-day ratings delivery; and multicultural audience measurement.

Two of those challenges dominate the commercial argument. Incomplete cross-platform measurement, according to CIMM, risks creating what the paper calls a fragmentation discount, where uncertainty about total reach and duplication pushes down advertising pricing. Low confidence in metrics separately produces a measurement risk premium, which skews how rights and long-term revenue streams are valued.

What the rights numbers do to measurement error

The economics section sets out why small errors carry large consequences. United States television and streaming sports media rights payments grew from $14.6 billion in 2015 to $29.3 billion in 2025 and are projected to exceed $37 billion by 2030, according to S&P Global Market Intelligence figures cited in the paper. CIMM argues that a doubling of value inside a decade means minor measurement inaccuracies can translate into millions of dollars of misallocated capital.

The individual deals reinforce the scale. The NFL holds media rights agreements totalling roughly $110 billion across the eleven years from 2023 through 2033. The NBA distributes games across ABC/ESPN, NBC/Peacock and Amazon from the 2025-26 season through 2035-36 for approximately $76 billion. Major League Baseball has explored consolidating local rights under a single streaming partner.

Inflation is not universal. Ampere Analysis reports that the UK sports rights market has broadly plateaued since 2021, reflecting pay-TV budget pressure, flat advertising revenue and limited new-buyer competition. Co-exclusive arrangements have accelerated there, with more than half of listed events now available through more than one broadcaster or service.

"Sports is becoming an asset class, not just an entertainment property," said Jay Prasad, Chief Executive Officer at Relo Metrics.

Where the audience actually sits

Seven out of ten sports viewers say they use streaming to watch sports, according to the paper. Nielsen data from April 2025 cited in the document shows how sharply that splits by age. Among NBA viewers aged two and over, streaming accounted for 21.9% of consumption, cable 30.1%, broadcast 37.3% and other sources 10.7%. Among viewers aged two to 34, streaming rose to 41.8% while broadcast fell to 9.3%. Among viewers 35 and over, streaming sat at 18.7% and cable at 39.3%.

Nielsen also reported that Super Bowl LX drew more than 125 million viewers, which the paper uses to argue marquee properties still aggregate simultaneous mass audiences.

Deduplication remains the practical obstacle. "Deduplication is a real headache for our measurement team," said Nicolas Grand, Executive Director of Research and Investment Analytics at WPP Media, in the report. A single fan moving between ESPN, ESPN+ and Hulu Live can register two or three times.

Attention, co-viewing and the spot-level gap

The paper leans heavily on vendor datasets to argue that sports inventory is not uniform. TVision attention index figures from August 2025 place the WNBA at 127, Major League Baseball at 120, NASCAR at 114 and the NFL at 109, against a streaming programming average of 92.

Co-viewing rates from TVision in September 2025 show football at 68.2%, hockey at 64.0%, soccer at 62.0%, baseball and softball at 61.0%, basketball at 59.9%, tennis at 58.0% and golf at 56.1%, against 52.8% for non-sports content. Nielsen separately estimates that out-of-home viewing may account for 10-15% of total viewership for many sporting events, and a higher share for tentpoles such as the Super Bowl and Thanksgiving NFL games.

Mediaprobe data indicates emotional engagement declines through a broadcast. Indexed to 100 at game start, American football falls to 94 after sixty minutes and 86 after 120 minutes. Soccer falls furthest, to 78.

The sharpest commercial illustration comes from iSpot. Four of the five most-active advertisers in the 2024 NFL regular season saw spot-level impressions under-deliver against Average Commercial Minute figures by a combined 381 million. One insurance brand accounted for 189,589,462 of that shortfall. A fifth advertiser, also in insurance, over-delivered by 82,730,142. CIMM presents this as evidence that timing and context of exposure may increasingly influence inventory valuation, since average audience measures obscure variation inside the break.

"Gone are the days of passive viewing," said Lily Tong, Director of Measurement at Amazon Ads, in the report.

The local measurement hole

The local chapter carries the paper's most quantified claim. Nielsen directly measures over-the-air households via panels in 208 measured markets, while Comscore estimates over-the-air viewing using survey inputs, neighbour tuning and television on/off validation. Digital simulcasts are credited only when vendor-specific requirements are met, including software development kit deployment, source identifier tagging and ad-load alignment. Streams lacking those elements are not credited.

Station apps, mobile and browser streams, FAST channels, YouTube TV and YouTube's ad-supported app are often excluded from local designated market area reporting, or routed into generic categories without market-level or source-level attribution. Return-path data misses households reaching cable subscriptions through an operator streaming app rather than a hardwired set-top box.

Taken together, CIMM estimates the scale of undermeasurement at potentially as much as 20-30% of total viewing, drawing on its own November 2025 study of local television and video measurement gaps. That study covered overall local viewing rather than sports specifically. The paper argues the consequence lands hardest on sports because a high proportion of games reach audiences through local broadcasters, and because persistent gaps understate the economic value of local sports properties during rights renewals, depressing carriage fees and advertising revenue.

The distributional point is explicit. Broadcasters and regional sports networks face the risk that incomplete measurement weakens their negotiating position. Streaming platforms, holding first-party viewing and behavioural data, are better placed to demonstrate engagement inside their own environments.

Sponsorship moves toward performance inventory

The global sports sponsorship market is projected to grow from approximately $73.8 billion in 2022 to more than $150 billion by 2032, according to Allied Market Research figures in the paper. Sponsorship assets have historically been valued through estimated media equivalency rather than measured exposure.

Computer vision and AI-driven logo recognition are changing that. Vendors including Trajektory, Relo Metrics, Samba TV and Hive now measure logo prominence, exposure duration, screen share, contextual placement and downstream behaviour. A Samba TV case study in the paper examined an energy company's Formula 1 Australian Grand Prix sponsorship and found the brand earned 4.7 times the visibility of other sponsors in its category, held the most prominent on-screen logo position for 90% of impressions, and appeared as the only logo on screen for 13% of its total exposure time.

Ownership fragmentation complicates pricing. In professional basketball, teams sell jersey logos, the league sells court signage, stadiums sell in-venue banners and broadcasters sell sponsored moments including tip-off, timeouts, halftime and the scorebug. The paper notes that industry executives complain no standard benchmarks exist to price sponsorships.

"The next era of sports sponsorships won't be defined by more metrics," said Alex Kerr, Chief Executive Officer at Trajektory.

First-party data and the comparability problem

Sports has become an early test case for pulling platform data into syndicated currencies. Amazon integrated self-reported Thursday Night Football audience data into Nielsen's outputs. Netflix began providing first-party data to Nielsen ONE in late 2023, and its Christmas Day NFL games in 2024 demonstrated how platform figures alter the visibility of streaming sports inside syndicated reporting.

The governance cost is comparability. If one platform contributes title-level census-like data and another does not, or if platforms define streams, minimum viewing thresholds and eligible devices differently, apparent comparability misleads. "First party data brings enormous opportunity, but it also raises the bar for governance," said Flora Kelly, Senior Vice President of ESPN Research.

"As publishers shift to first-party measurement solutions, year-over-year comparisons become increasingly difficult," said Kym Frank, Senior Vice President of Research and Data at FOX.

Second-screen behaviour compounds the counting problem. Comscore reported that Duke generated 5.4 million social media engagements during March Madness despite not winning the tournament, against 2.7 million for the Florida Gators, and tracked a 96% increase in NFL TikTok engagements between 2024 and 2025. Separately, one 2025 study cited in the paper reported that 86% of Formula 1 fans use a second screen during races.

"Watching live games is just one piece of the puzzle," said Michelle Auguste, Vice President of Global Media Insights at the NBA.

Discrepancies inside the document

The paper carries several internal inconsistencies that bear on how its figures line up against each other.

The report cover is dated July 2026, while the release to the industry carries an August 25, 2026 dateline. The announcement describes the November 12 event as the Sports Media and Measurement Summit in one passage and as the Sports Measurement Innovation Showcase in another.

The nine challenges listed in Exhibit 6 do not match the nine rows in Exhibit 9, which maps challenges to emerging solutions. Unauthorised streaming appears in the first list but not the second. Exhibit 9 instead introduces a row covering attention, engagement intensity and spot-level audience variation, and reframes multicultural measurement as potential panel bias.

On second-screen reach, the body text states that Monday Night Football on ESPN achieved an incremental reach lift of 31% in December 2024 once ESPN Fantasy Mobile App users were incorporated. The accompanying Exhibit 10 reports a 27% incremental reach lift for December 2025, based on 28.7 million Monday Night Football series viewers and 7.8 million incremental app viewers reaching a cross-platform total of 36.6 million.

The Samba TV case study text refers to Exhibit 10 while the exhibit itself is numbered Exhibit 11. The Mediaprobe emotional engagement chart labels two separate columns "Baseball" with different values.

Two of the executives quoted in the report, Nicolas Grand of WPP Media and Inderpreet Sandhu of Google, do not appear in the acknowledgements list of Showcase participants.

Three futures for the currency

CIMM sets out three forward-looking scenarios rather than a forecast. Unified currency convergence would see stakeholders align around a limited number of interoperable cross-platform frameworks. Platform metric dominancewould see major digital platforms promote proprietary measurement as the primary trading metric, with first-party viewing data, clean-room attribution and retail transaction signals becoming central inside platform ecosystems. Hybrid negotiated currency systems would see multiple currencies coexist and be applied selectively by campaign objective or commercial negotiation.

The paper points to joint industry committees in other markets as a counterpoint to the fragmented United States environment, naming BARB in the United Kingdom, OzTAM in Australia and Mediametrie in France.

Artificial intelligence sits across all three scenarios. The document raises the prospect of synthetic audience currencies that combine directly observed behaviour with modelled estimates, and warns that such systems could reduce methodological transparency by making reported estimates dependent on proprietary modelling assumptions rather than observable exposure data.

Why this matters for the marketing community

The timing places the paper against an unusually active measurement calendar. Nielsen has committed to seven methodological changes to its National Big Data + Panel product, with deployment planned for August 31, 2026, including an updated co-viewing process that integrates panellists fitted with wearable meters. That wearable programme was piloted at Super Bowl LX on February 8, 2026 and extended through the first half of the year, with the stated goal of feeding co-viewing data into currency measurement during the 2026-2027 season. Given co-viewing rates above 68% for football, any methodology change in that area moves reported sports audiences directly.

The commercial pressure on measurement suppliers is visible in their own results. Comscore announced up to $25 million in annual cost reductions in August 2026 as second-quarter revenue fell 11.3%, with workforce reduction charges of $7 million to $9 million running into the third quarter of 2027. The paper's argument that independent validation matters arrives while at least one independent validator is cutting cost.

For buyers, the outcomes layer the paper describes is already shipping. EDO launched an autonomous campaign optimisation suite for convergent TV on June 4, 2026, covering frequency capping, creative rotation and audience targeting against measured outcomes. LiveRamp research published in July 2026 put the cost of identity errors at up to 70% of campaign ROI, splitting the problem into missing exposures and exposures attributed to the wrong person. Both sit directly on the deduplication and attribution gaps CIMM catalogues.

The attention argument also has an independent evidence base. A VAB report produced with TVision in February 2026 found premium streaming outperformed YouTube across every attention and engagement metric measured, using second-by-second person-level data from thousands of households. Women's sports, which the paper singles out as a case where engagement diverges from scale-based valuation, reached 46 billion viewing minutes in 2025 according to Nielsen.

What the paper does not offer is a resolution. CIMM is a convening body, not a standards-setter with enforcement power, and the document is explicit that its interpretations are its own rather than positions held by participating companies. The 20-30% local gap, the 381 million impression shortfall and the nine challenges are diagnoses. Whether they translate into changed carriage fees, rights valuations or advertising rates depends on decisions the paper does not control and the November 12 summit will not settle.

Timeline

Summary

Who: The Coalition for Innovative Media Measurement, a non-partisan pan-industry association, published the paper. Jon Watts, Managing Director at CIMM, provided the announcement commentary. Executives quoted inside the report include Greg Dale of Comscore, Jay Prasad of Relo Metrics, Flora Kelly of ESPN, Kym Frank of FOX, Michelle Auguste of the NBA, Lily Tong of Amazon Ads, Alex Kerr of Trajektory and Nicolas Grand of WPP Media.

What: A paper titled "Keeping the Score: Industry Perspectives on Innovations in Sports Media Measurement," identifying nine measurement challenges and the technologies and governance frameworks emerging to address them. Central figures include a potential 20-30% local viewing measurement gap, a 381 million impression shortfall against Average Commercial Minute delivery for four of the top five 2024 NFL advertisers, and co-viewing rates of 68.2% for football.

When: Released on August 25, 2026, though the report cover carries a July 2026 date. It synthesises showcases held in the second and fourth quarters of 2025. CIMM will hold its second annual Sports Media and Measurement Summit on November 12, 2026.

Where: The United States market is the primary focus, with comparative references to the UK, Australia and France. The November 12 summit takes place at the Empire State Building in New York.

Why: United States sports media rights payments rose from $14.6 billion in 2015 to $29.3 billion in 2025 and are projected to exceed $37 billion by 2030. CIMM argues that at that scale, measurement systems no longer merely report audiences but determine how rights are priced, how advertising inventory is traded and how capital is allocated across premium video, making the accuracy and comparability of those systems a commercial variable rather than a technical one.