Measurement
State of Play
Revised Nov 1, 2026The moment. The count changed what it counts. Comscore’s new Score Report ranks TV and streaming by share of ad-exposure time rather than viewing, and its first reading puts broadcast at 43.2%, cable at 34.5% and streaming at 22.3%. VideoWeek has NBCUniversal and Versant leading the August table at 17.5%, with a merged Paramount-WBD set to top it at 23.9%. iSpot’s Q3 report has streaming splitting ad revenue evenly with linear even though linear still delivers the vast majority of impressions, which is the same gap seen from the money side. Political money is the other pressure on the count: forecasters now project $12 billion for the 2025–26 midterm cycle, with local broadcast holding 50% of the $11.038 billion spent cycle-to-date against CTV’s 21%, and eMarketer has that spending crowding holiday marketers out of TV and CTV inventory.
The tension. Buyers want more precision from a surface that shows them less. Nielsen found 55% of marketers trust AI-generated recommendations only when a human verifies them, and says the share prioritizing measurement precision has doubled year over year. AdExchanger’s guest column argues B2B buying has moved into AI assistants, private messages and dark social that ad tech cannot observe. Extreme Reach’s John Batter says the industry still has no creative ID to tie ad variants to masters or production cost to outcome, and AI is multiplying the variants. The platforms are supplying the proof themselves: OpenAI’s Ads Manager now imports 30 days of aggregate Google Analytics 4 data so advertisers can run attribution comparisons inside it, and a six-month WPP Media study with Amazon Ads is the evidence offered that streaming moves sales. TVREV’s Sofie Sue Rutgeerts argues attention is a middle step, not the outcome. The sell side keeps answering the measurement question with its own numbers.
Who’s moving. Comscore, counting ad exposure instead of viewing. Comscore again, reporting Q3 results on November 3. iSpot and TVREV’s Jason Damata, on revenue parity at a fraction of the impressions. Extreme Reach’s John Batter, on the missing creative ID. OpenAI, wiring GA4 into its own ads manager. MediaPost’s Joe Mandese, on $12 billion and CTV’s 21%. eMarketer’s Marisa Jones, on political crowding out the holidays. Truthset and Aristotle, verifying identity on political CTV buys. Databricks, pitching prediction over impressions. Nielsen, with streaming at 58% of minutes for the top animated series. eMarketer’s Asa Hiken, on CTV conversion rates rising. Nielsen again, confirming a WNBA Finals Game 1 record of 1.76 million viewers.
What the desk is watching. Whether buyers start pricing on Comscore’s exposure share rather than viewing share, and whether the November 3 call puts revenue behind the new report.
‘Resident Evil’ reboot opens to franchise-record $60 million
Sony’s “Resident Evil” opened to $60 million domestically, beating projections of $40 million to $50 million and more than doubling 2010’s “Afterlife.” Sony spent just $6.4 million on national TV ads, concentrated on NFL and college football broadcasts.
Tech brands triple NFL ad share as Starlink debuts on linear
The Measure’s Jon Lafayette reports that iSpot’s Week 1 NFL scoreboard shows tech, AI and cloud brands hit 6.21% share of voice, up from 2.20% a year ago. Starlink, absent from linear TV last season, ranked 27th of 149 brands by impressions.
Podcast industry presses IAB for measurement overhaul as ad spend lags listening
At the IAB’s Podcast Upfront, executives praised rising investment in the channel but warned that ad spend still trails actual consumption. Speakers called for modernized, standardized metrics to close that gap and better reflect podcasting’s growth.
Why time spent no longer equals value in media
Doug Shapiro argues time-spent metrics, long treated as a proxy for value, are breaking down as monetization per hour diverges widely across media, and AI’s cheap content and compression tools will widen the gap further.
DoubleVerify adds conversational AI layer to its measurement platform
ExchangeWire reports that DoubleVerify launched Ask Neura, a natural-language interface built into DV Pinnacle that lets advertisers query fraud, brand suitability, viewability, and attention data instead of relying only on predefined reports.
Prediction markets drive sports betting’s TV ad surge
The Measure’s John Cassillo reports that betting brands spent $269.8 million on national linear TV through August 31, up 40% year over year, per iSpot data shared with AdAge.
Forrester: “Recommendation share” will replace impressions as top media KPI
MediaPost’s Ray Schultz reports that Forrester’s “Predictions 2027: Media And Advertising” report argues that as consumers delegate research to AI engines, marketers will care less about ad views and more about whether engines recommend their brand.
NFL opening week viewership falls 13% to 19.5 million despite extra games
MediaPost’s Wayne Friedman reports that NFL opening week averaged 19.5 million viewers per Nielsen, down 13% from last year’s 22.3 million even with two extra games added. Stripping out four early Sunday afternoon windows, the week actually rose 2% to 21.9 million.
PeakMetrics launches GEO tool tying AI brand answers to news and social narratives
markets.businessinsider.com reports that PeakMetrics debuted AI Perceptions, a platform letting comms teams track how brands appear across ChatGPT, Gemini, Claude, Grok, and Perplexity.
Puck podcast: Paramount’s red-state gambit, ESPN’s talent math
Puck reports that Julia Alexander and Dylan Byers unpack Paramount’s Nashville-or-Texas relocation brinksmanship with Rob Bonta, the Bari Weiss ratings debate and Nielsen’s shifting methodology. They close by weighing ESPN’s talent-pay math on Joe Buck, Troy Aikman, and Tom Brady.
Study: 5% of CTV ad impressions are mislabeled, costing buyers $2 billion
ExchangeWire reports that 51Degrees analyzed 11.03 million OpenRTB impressions from May 15-20, 2026, finding 4.92% labeled as CTV were actually lower-value, non-CTV inventory, versus 0.18% mislabeled in reverse. The firm ties the mismatch to roughly $2 billion in global overpayment by advertisers.
Nielsen speeds up streaming charts as Reacher, Outer Banks lead summer
TheWrap’s Loree Seitz reports that Nielsen sped up its top-10 streaming reports from a four-week to a two-week delay, releasing three weeks of data at once covering Aug. 17 through Sept. 6.
Automotive TV spending falls 15.4% in August, sports still dominant
MediaPost’s Tanya Gazdik reports that Automakers spent $72.5 million on national linear TV in August, down from $85.7 million a year ago, per iSpot.tv. Household impressions fell 20.2%, while NFL preseason alone drew over 22% of auto industry outlay.
Nielsen study: pairing digital audio with CTV nearly doubles brand recall
Nielsen reports that a Nielsen and Katz Digital brand-lift study of 600 D.C.-area adults found unaided recall of 37 percent for audio-only and 37 percent for CTV-only ads, versus 65 percent when both formats reached the same person.
Nielsen halves the lag on its Streaming Top 10, to two weeks
Variety’s Selome Hailu reports that Nielsen cut the reporting delay on its weekly Streaming Top 10 charts from four weeks to two, citing improved analytics capabilities. The company calls it the most significant enhancement to Streaming Content Ratings since the metric launched in 2017.
SAGE data: Wendy’s value messaging beats humor and sports tie-ins on likeability
iSpot’s SAGE AI analyzed Wendy’s 2026 TV ads, finding value/price messaging the top likeability and purchase-intent driver, ahead of humor and sports tie-ins.
Nielsen moves streaming ratings to daily delivery, matching linear TV speed
Nielsen is shifting Streaming Content Ratings from weekly to daily reporting, cutting delivery time by up to 80%. Clients get data each business day for viewing two days prior, and Streaming Top 10 reports move up by two weeks.
Brands can claim log-level campaign data as their own first-party asset
MediaPost’s Kym Insana reports that a commentary argues brands can request log-level campaign data, including hashed viewer data, free from buying platforms. Loaded into a CRM, it lets brands find who watched most of a CTV ad and retarget them on Facebook or Instagram.
FreeWheel adds attention and emotion signals to CTV inventory packaging
FreeWheel’s Premium Signal Packaging now folds in data from Mediaprobe, Adelaide and Magid, letting CTV buyers select inventory by attention, quality and emotional resonance. A FreeWheel-Mediaprobe study found six-minute hourly ad loads drove 89% higher recall than twelve-minute loads.
Small TV markets outdraw top 10 for NFL and college football viewing
The Measure reports that Madhive data shows the 160 smallest of 210 U.S. TV markets deliver more total audience than the top 10 markets combined for both NFL and college football broadcasts this season, per new findings.
Horizon Media pipes Roku device-level TV data into its Blu platform
Roku is bringing its TV data into HorizonOS’s Blu marketing intelligence platform, giving advertisers device-level viewership data instead of panel-based estimates. Full integration is planned for this autumn, with A+E Global Media cited as a client already using Blu.
New Meta data: high-click video ads deliver the worst ad returns
MediaPost’s Wayne Friedman reports that Billo’s analysis of 88,329 Meta video ads found the highest-click categories, health/beauty and arts/entertainment, posted ROAS of just 1.82 and 2.04, while low-click baby/toddler ads hit 4.99 against a 2.17 average.
Study: MVPD deals deliver 2.5x the reach of programmatic CTV
TVREV’s Michael Guth reports that a Spectrum Reach and EMARKETER study of 239 CTV ad buyers found MVPD campaigns delivered 61% sustained accurate reach over 90 days versus 24% for programmatic CTV.
Nielsen adds data enrichment tools to Marketing Cloud
TV Tech’s George Winslow reports that Nielsen has launched new data enrichment capabilities inside its Marketing Cloud platform, aimed at helping advertisers and agencies append richer audience and behavioral data to their existing customer records.
Brands scramble for AI visibility as no standard measurement exists
Digiday’s Seb Joseph reports that marketers now fight over whether AI chatbots mention them at all, not how they rank, but citation patterns vary wildly by platform.