Comscore has a new monthly scoreboard that says linear TV holds 77.7% of measured TV ad time. It landed two months after streaming took most of the primetime upfront money. Its own chart contradicts its own press release. And the endorsement in the release comes from a broadcaster.

On October 21, Comscore published the first monthly edition of The Score Report, a free ranking of who gets the most ad exposure time on U.S. television. Comscore bills it as "an independent view of advertising exposure." The August numbers: broadcast 43.2%, cable 34.5%, streaming 22.3%. Add broadcast and cable together and linear holds 77.7% of measured TV ad exposure time. Comscore hasn't said whether that time is counted in people, households or devices.

That is the number broadcasters have wanted all year. Nielsen's Gauge had streaming at 49.6% of all TV viewing that same month. Comscore's argument, delivered by interim chief revenue officer Kevin Burns, is that viewing time has stopped being a reliable stand-in for the audience an advertiser can actually reach with an ad. The outside endorsement in the release came from Dan Lanzano, president of national ad sales at Nexstar.

And they need it now. For the first time, streaming took more of the 2026-27 primetime upfront than linear: $17.2 billion against $16.6 billion, by Media Dynamics' estimate, while cable CPMs fell 8.5%. Back in March, Nielsen's own upfront planning guide put streaming at 66.7% of ad-supported TV time for adults 18-49. Comscore's chart arrives two months after the money moved, with a number that says it shouldn't have.

Two scoreboards. Two answers. Both free. Neither is currency. Each one comes out the way the side it favors would want.

One rule in the report looks like a trick and isn't. Comscore's FAQ credits ads seen through YouTube TV, Sling and Hulu Live to the network or station where the content originated, so an NBC ad watched on YouTube TV counts as NBC. Nielsen's Gauge has handled virtual pay-TV viewing the same way since 2023. That's the industry convention, not a Comscore invention.

What the convention costs is real, though. Virtual pay TV now reaches 21.4 million homes, about 35% of all pay TV, per MoffettNathanson. A breakdown attributed to Nielsen's January Gauge put the big virtual services at roughly a tenth of all broadcast and cable viewing. Count that time as streaming and, by Unmeasured's rough estimate, streaming's 22.3% lands closer to 30%. And phones, tablets and laptops aren't in the report at all.

Frank Kowalczyk: I did two tours at comScore. Second tour, a guy in a quarter-zip explained to me that a stream is not a stream if the channel is old enough. I asked if that went for me. He said I was "legacy inventory." He wasn't wrong. He was just early.

1. DoubleVerify votes Thursday. Shareholders vote October 29 on Nielsen's $13.60-a-share buyout. Under the proxy, a shareholder who doesn't vote counts the same as a no. Funds affiliated with Providence Equity, holding about 11.8% of the shares, have already agreed to vote yes. Unmeasured's October 14 issue covered the clause that lets Nielsen refuse a regulator's firewall.

2. Nielsen takes Cumulus to the Supreme Court. Nielsen has asked the Court to review the Second Circuit ruling that upheld an injunction barring it from tying national radio ratings to local ones. By Nielsen's own account in the petition, the gap that left Cumulus with no real choice was about $1.2 million a year.

3. Same campaign, two counts. Warner Bros. Discovery's October white paper on five measurement vendors found estimates that nearly doubled from one vendor to the next. One example: a Q4 cross-platform financial campaign where Comscore counted 43.5 million impressions among adults 25-54 and iSpot counted 24.6 million. WBD's verdict was that cross-platform reach measurement is still immature.

Frank Kowalczyk: Forty-three and a half against twenty-four six. I've seen tighter spreads at a church raffle. The raffle had a guy with a clipboard. Nobody here has the clipboard.

Comscore's release says streaming's recent rise is shifting from cable. Its own trend chart says something else.

Between June and August, streaming rose from 21.1% to 22.3%. Cable didn't fall. Cable rose, from 34.2% to 34.5%. Broadcast fell, from 44.7% to 43.2%. From July to August alone, cable gained 0.4 points and broadcast lost 0.8.

On Comscore's own bars, streaming's summer gain came out of broadcast. That is not the sentence a report endorsed by a broadcaster would want to lead with. The release says cable.

The release also says streaming sat on a plateau around 19% for the five months before February. The chart shows 16.1% to 18.5% for September through January, an average of about 17%. None of those five months reaches 19. Then streaming jumps 4.5 points from January to February. Comscore has pointed to February's sports calendar, the Super Bowl, the Winter Olympics and the NBA All-Star Game, but hasn't said how much of the jump came from programming, ad loads, coverage or method.

Over the full year, the release is right about direction: broadcast gave up 3.4 points, cable 2.8, and streaming gained 6.2. It's the recent story, the one meant for this fall's buyers, that the chart doesn't support.

Frank Kowalczyk: "Around nineteen." Chart says seventeen. In Rolling Meadows that's not around. That's a different exit. I put the Kowalczyk Ratio on it and got two numbers, which is one more than the ratio is supposed to give you.

The phones aren't counted. The totals aren't published. The meter is built on cable boxes.

Below the line: what "streaming" means in this report, what data it's built on, who's standing next to it, why Comscore needed a scoreboard this quarter, and why the 2027 upfront is about to be argued with two free charts that aren't currency.

Enter your email to keep reading. It's free.