OMG’s Kelly Metz explains the nuances of YouTube measurement in AdAge’s latest look at the currency debate
Behind the struggle at the root of much of TV’s measurement quandary
April 23, 2023 | By Jack Neff
There is a measurement “cat” fight brewing between YouTube and the TV networks.
Hard as it may be to believe, at the heart of the ongoing war in media measurement are cat videos.
TV executives have long dismissed YouTube’s vast inventory of user-generated content as “cat videos.” But increasingly those videos have become a larger share of the content being streamed to connected TVs.
With Google-owned YouTube now more firmly squaring off with more traditional networks in the battle over connected TV, network executives are concerned that ads on social video—particularly YouTube—will be measured and paid in the same way as ads in premium network programming, driving down prices.
Of course, YouTube executives see no reason why their ad inventory should be treated differently than linear or any other streaming TV.
And it looks like YouTube executives could soon get their wish as it becomes increasingly likely the two sides will live in the same measurement world whether they like it or not.
Nielsen plans to do away with conventional ratings built around linear TV schedules in favor of an impression-based standard like that which long has been used for digital media. Nielsen’s primary competitors are also gearing up for impression-based measurement to replace older C3 and C7 ratings, which measure the viewership of commercials in the three and seven days after a program airs.
The battle came to the fore recently as the newly minted U.S. Joint Industry Committee—dominated by network TV players—issued currency measurement standards that, among other things, incorporate a content quality measure into currency between buyers and sellers. Such a measure would likely prop up relative prices for ads in network programming.
YouTube shot back with its own set of proposed standards calling for video impressions used as currency based on the lowest possible common denominator, the Media Rating Council’s two-second viewability standard, and no content quality metric. Both sides cite agency and marketer buy-side players backing their efforts—albeit without those players endorsing specifics of either side’s positions.
Nielsen, which was invited to participate in the JIC’s request for information, sent a letter last week declining to participate for a number of reasons, including lack of participation by YouTube and other media players, also including Disney, Netflix and Amazon. Among the issues Karthik Rao, CEO of measurement at Nielsen cited, was the JIC’s call for the content quality standard opposed by Nielsen.
Even before the dueling standards debate, anyone attempting to develop cross-platform video measurement has run into the network vs. YouTube battle. Association of National Advertisers executives, as they began a painstaking journey to begin cross-media measurement pilots three years ago, said they quickly heard from the networks’ trade group the Video Advertising Bureau, regarding concerns of an impression standard that would undercut the value of premium programming.
Dave Morgan, CEO and founder of TV and video analytics company Simulmedia, said he’s heard from multiple people in the industry that network concerns about Nielsen providing measurement, that would let YouTube and other digital players undercut their pricing, prompted much of the pushback against the measurement giant and embrace of its rivals in recent years.
Morgan acknowledged that Nielsen brought on its own problems with methodology errors that led to its Media Rating Council accreditation being suspended for 19 months. But the battle of the incumbents against the newcomers is at the core of the controversy, he said.
“This isn’t new in the history of media,” Morgan said, pointing to parallels going back to the 20th century, when daily newspapers tried to exclude weekly newspapers from Audit Bureau of Circulations coverage, or when broadcast networks pushed back against cable networks getting similar ratings coverage.
Today, YouTube poses the most direct threat to TV network pricing among digital platforms, because it operates increasingly on the big screen as the networks do. Meta and TikTok offer plenty of ad-supported user-generated videos, but don’t have the same significant presence on TV screens that YouTube does thanks to its CTV app.
Are impressions equal?
Network executives see the idea, of measuring all video content with a single impression metric that counts reach without accounting for production quality, as a race to the bottom.
“Content that companies like ours produce shape society,” said Kelly Abcarian, executive VP of measurement and impact at NBCUniversal during an appearance at the Paramount Measurement Now event last month. “If we leave the quality of content behind, I really do fear for the society that we’re also starting to leave behind if none of us can afford to invest and produce great content, because it’s a race to the bottom on price, and it’s not even a fair valuation of price.”
Some measurement systems “are focused exclusively on reach as some kind of common aggregator with an impression being an impression being an impression,” Abcarian said, “whether it’s a critically acclaimed drama, a sporting event, or a UGC cat viral video.”
TV networks are spending $10 billion or $20 billion a year on content, said Paramount Advertising President John Halley. “There’s a reason for that. These things work. It relates to engagement.”
NBCU has put forward a Content Quality Index that aims to separate the “cat videos” from the premium network fare. It could become the basis for the JIC’s currency quality modifier, given that no other JIC member has offered an alternative. In an NBCU measurement event in February, Abcarian said the CQI gave her network’s content a 70 rating, compared to 50 for user-generated content.
Whether a CQI based heavily on ad recall, is fair is a matter of some debate. Research executives note that how familiar a consumer already is with a brand is the biggest underlying driver of ad recall scores, regardless of the creative or the media placement. So TV networks, with a more limited range of well-established brands, are likely to score higher on such a measure than social platforms with millions of smaller advertisers, if the comparison is done on a random sample of ads from each medium.
Abcarian in a statement defended CQI as “completely content and platform agnostic.” She added that “We strongly believe there is a lot of good premium content on YouTube and other social platforms that will score well for advertising resonance. It really comes down to how engaging the content is and how good the platform and advertising experiences are together.”
YouTube’s principles released last month are squarely against using any kind of content quality modifier as deal currency.
“Advertisers have expressed the need to measure reach across all platforms,” Kate Alessi, managing director, YouTube/Video Global Solutions at Google, said in an interview. “Regardless of the content creator or the video length or the camera quality, they just want to understand where their audiences are and how can they reach them in the most effective manner. And so that’s what we’re trying to do with our proposal.”
Alessi acknowledged that the MRC, besides the two-second viewability standard, does also have a duration-weighted cross-platform impression standard that can be used by buyers and sellers to modify deals based on how much of a video was watched and will provide duration signals in measurement solutions too.
But in her blog post, Alessi points to research from critics of the duration-weighted standard and to Google-commissioned research from 2018 indicating that duration doesn’t affect ad recall much beyond 10 seconds of viewing, regardless of the ad length. For this reason, she argued that duration should be considered separately from impressions.
Alessi also pointed to YouTube-commissioned research from Eye Square last year on content that ran on the platform. Eye Square found ads running in creator-produced content do as well or better than broadcast quality studio-produced content at generating ad recall and brand preference lifts. That research also found ads in premium YouTube Select inventory performed about as well as those in the bigger, lower-cost pool of YouTube auction content.
YouTube already recognizes distinctions
In reality, some on the buy side say the differences between networks and YouTube may not be as great as either side makes them look.
YouTube already segments its own inventory into a premium YouTube Select tier sold through upfront guarantees and other direct buys that value it on par or often higher than linear TV inventory. Networks long have sold remnant linear inventory at bargain-basement direct response prices that don’t rely on audience ratings. And the JIC, while formed without input from YouTube or other pure-play digital players, has invited YouTube to join, and the company is in talks with the network data sharing consortium behind the JIC, OpenAP, to do so.
The base two-second viewable impression should not be the foundation of cross-platform video currency, Kelly Metz, managing director of advanced advertising solutions for Omnicom Media Group argued at the Paramount event.
Research OMG released last year points to one reason why: YouTube itself doesn’t even go to market that way. YouTube sells inventory, priced differently, at virtually every quality tier of video. That includes YouTube Select, which represents the top 5% of its inventory as ranked by the Google Preference Score algorithm, which is based on such things as likes, watch duration and brand suitability signals.
Like ads in premium network content, YouTube Select is priced on a rate card, sold via guarantees or by private marketplace bidding, and includes human review of 100% of the inventory to ensure no subject matter unwanted by advertisers slips through the cracks, according to OMG.
And YouTube Select already sells in the range of $25 CPMs, or the cost to reach 1,000 viewers, comparable to many CTV CPMs. OMG found the market already operates in a clear content-quality continuum, with CTV and in-stream online video at the top and in-banner and small-player video at the bottom—regardless of there being no official industry standard content quality measure.
Some of YouTube’s premium inventory is even network content, such as clips from NBCU’s “Saturday Night Live,” Metz pointed out in an interview. “But YouTube is getting the brand lift and the credit because it’s on their platform,” Metz said. “So that’s another reason we need to get to the program level [on quality measurement] because you can’t just look at the platform. It’s a very nuanced area. And that’s why every time I see my friends at Google, I say join the JIC. Please just get in the game.”
That’s possible. Alessi said she’s been in touch with OpenAP CEO David Levy, who’s administering the JIC as well, about participating, though no decision is final.
Alessi also said YouTube’s position has been misconstrued in some ways. “Our stance is not that all impressions should be treated equally,” she said. “It’s really that all impressions should be counted, and then measurement can be layered on top.”
For that matter, TV networks themselves have a similar range of quality, and the market makes the distinction pretty readily without an official content quality measure in place, Morgan said.
“We measure the Super Bowl the same way we measure reruns of ‘America’s Funniest Home Videos’ from 10 years ago,” Morgan said. “But the price of one is about 1,000 times higher than the other. We don’t have any problem finding an appropriate price.”
More than cat videos
While “cat videos” have become a meme for networks, it’s an unfair slight against the many YouTube creators whose high-quality work increasingly shows up on TV screens, said Martin Cass, chief marketing officer of BBTV, the biggest network of YouTube creators.
BBTV creators produce less than 1% of the videos uploaded to YouTube each month, but 30% of the platform’s reach, per Comscore, Cass said. And 38% of BBTV creators’ video views are streamed on TV screens, he said.
Cass pointed to BBTV creator Sam Zien, aka “Sam the Cooking Guy,” whose YouTube channel has 3.5 million subscribers. He started his show on the platform after getting fed up doing a cooking show on a San Diego TV station and deciding to do it on his own.
“If YouTube didn’t exist, where would Sam be?” Cass asked. “The answer is he’d be a superstar on The Food Network.”
Nielsen in a sense is caught in the crossfire, but it isn’t seen as an unbiased referee by some network executives.
Some were rankled by Nielsen co-presenting a study with YouTube last year, which was used in a Coalition for Innovative Media Measurement conference and Google blog post, that said YouTube CTV’s return on investment was 1.6 times higher than that of linear TV. This was based on a meta-analysis of marketing mix model studies over several years.
Network executives, who spoke on the condition of anonymity, took issue with Nielsen, supposedly a neutral currency arbiter, putting its weight behind a YouTube sales pitch. They also saw the data as hand-picked and flawed, since it compared YouTube only to linear TV, not to other ad-supported CTV. And the study didn’t indicate whether the higher ROI was simply driven by lower cost per thousand (CPMs) rather than better top line results.
A Nielsen spokeswoman said the company does custom research with many clients and stands by the research and its methodology.
Network executives also called attention to LinkedIn posts by Nielsen employees last year that appeared to offer YouTube audience measurement, or at least upgraded aspects of it, for free to agencies—an inducement no other media company appears to benefit from. While agencies pay far less for Nielsen data than the networks do, they tend to pay extra for most every new feature Nielsen adds, agency executives said.
The Nielsen spokeswoman said the company doesn’t give away YouTube measurement free, but did not explain the apparent free offer in a Nielsen employee’s LinkedIn post.
In a panel during a side-conference hosted by Mediaocean during CES in January, Abcarian said Nielsen also appears to be putting its stamp of approval on co-viewing data for YouTube CTV audiences that YouTube processes for itself. YouTube calculates the number of people per household watching its videos based on pop-up surveys that ask how many people are watching.
A YouTube spokeswoman said those survey-based co-viewing estimates are presented to advertisers directly for planning purposes, not as currency through Nielsen. And the measurement firm’s spokeswoman said the company calculates YouTube co-viewing using its own methodology, though she declined to specify the details.
One agency executive, who spoke on the condition of anonymity, said that while the agency grudgingly uses YouTube’s co-viewing data in planning, he doesn’t trust it as currency. But an executive with a Nielsen rival said YouTube’s pop-up surveys—provided they’re served without bias and audited—are probably a good way to measure co-viewing.
Making the money equal
Even if the Nielsen measurement is unbiased, it’s harder to argue that the financial arrangements are.
Based on reports from several industry executives and from Nielsen executives prior to the company going private last year, each of the five leading TV-centric network companies pays Nielsen a nine-figure annual fee that’s about as much or higher than the company gets from YouTube and all other digital native competitors combined.
Those network contracts were based historically on a costlier-to-manage household panel that produces the C3 and C7 ratings for TV networks versus less costly, more passively collected digital viewership data. But with the advent of its new Nielsen One service, the company has said the C3 and C7 ratings are going away in favor of a uniform, second-by-second impression measure.
The household panel behind those ratings will remain to calibrate data from set-top-boxes and smart TVs. But the new impression-based measurement that panel will now support will cover networks, YouTube and all other video platforms. That would leave Nielsen charging networks a lot more than YouTube or other digital players for the same data.
It’s a business model that an executive close to Nielsen acknowledged will have to change – though it’s not clear how. The same executive acknowledged that while Nielsen’s plan is to do away with C3 and C7 TV ratings by 2024, it’s not clear the industry will actually accept that. In turn, that could provide some rationale for Nielsen keeping those lucrative network contracts in place longer.
A Nielsen spokesman declined to comment on the pricing, business model and ratings issues.
TV ratings may be hard to kill
However, as the battle between networks and YouTube turns out, winning marketers over to a common currency could be harder. Morgan said replacing TV ratings with impression-based numbers won’t be easy for marketers who want comparable historical data.
“Cat videos” or not, those YouTube ads billed on two-second MRC viewable impressions do get seen. Applying the same standard to high-quality, studio-produced network shows is likely to turn up quite a few that aren’t. That could be an unpleasant change for everyone.
Getting rid of TV ratings—which are based on average viewership across six-minute segments—will make about 15% of ad inventory disappear on most shows, Morgan said. These are ads people aren’t actually watching but get credit by being lumped into the average audience. Under a new impression-based system, advertisers may not get charged for ads that don’t get seen, he said, but they’ll probably have to pay more for the ads that do.
“If it all changes tomorrow, a lot of people are going to be a bit lost,” Morgan said. “While the traders and buyers will be able to find ways to transact, it will create a lot of changes for the marketers, who don’t expect that much volatility and won’t necessarily understand the dynamics. No one wants to tell the CEO of the company his media cost just went up 15% because the definition behind it changed.”