“The new measurement is therefore addressing an increasing and significant pain point in measurement, where some smaller rated events are not viewed by people in research samples even though there is an audience in the population,” Doe said in his letter. “This manifestation of sampling error is very effectively addressed.”
It’s progress, said one network executive, who however added that Doe’s letter glossed over the fact TV networks and local stations have for years been unable to sell ads on such zero-rated programs, losing millions if not billions of dollars on the sort of lesser watched, fragmented inventory that digital publishers can trade programmatically.
Nielsen's long and winding big data road
MRC committee meetings to review separate audits of Nielsen’s smart TV data and cable and satellite set-top box data are set for this week and next, and the results of those are uncertain.
Nielsen backtracked on offering use of its big data as currency in April in the face of pushback from TV networks and some agency executives who said it had provided too little data too late about the historical impact the new methodology would have on ratings. That move came just after Nielsen’s household panel-based national TV measurement service had its MRC accreditation restored, and those panel-based ratings remained the predominant currency in this year’s upfront. Nielsen’s move now appears to mostly affect scatter deals until next year’s upfront.
“This year’s upfront was done on panel,” said Sean Cunningham, CEO of TV network trade group the Video Advertising Bureau. “So the idea of mixing and matching two different currency sets is kind of a non-starter for 99% of advertisers. I’m struggling to think of one advertiser I’ve ever worked with that made deals with multiple currencies from the same provider.”
Read more upfront news from Ad Age
It's a positive, however, that buyers and sellers get a chance to get a better idea of how the big data “successor currency” affects audiences, possibly trading on it in scatter, prior to next year's upfronts, Cunningham said.
Technically, the MRC audits to be reviewed over the next two weeks aren’t for the new Nielsen One product. They’re for big data components that ultimately will be part of that product but are currently being considered as part of Nielsen Digital Ad Ratings.
The MRC audit and review process may be moving faster than some people want, but that’s driven by the fact that Nielsen is going to market with big data, so it needs to be examined fast, said MRC’s Ivie.
“It’s a bit of a wishy-washy situation,” Ivie said. “They’re not calling it a currency. In September, when it hits, they’re calling it a new measurement. … I have to audit that, because that’s altering the production system for a national service. So I’m auditing that as hard and as fast as I can, because it’s a huge change.”
Nielsen officially rolled out its Nielsen One next-generation cross-platform measurement, incorporating big data, in January. But the company hasn’t broadly supported use of its big data for trading, just for planning. Facing pushback from some media companies, notably TelevisaUnivision, according to people familiar with the matter, Nielsen did in April say sellers and buyers who wanted to trade on its big data could.
Read more: Nielsen brings next-gen measurement to market
Mixed impact for Spanish-language TV
But TelevisaUnivision has found few if any takers among agencies for trading on Nielsen’s big data, according to people familiar with the matter. One agency executive expressed sympathy for the effort but resistance to trading on two separate Nielsen currencies at a time in the market. Another issue is disparate results for different Spanish-language networks for reasons that are hard to explain, industry executives said.
Representatives of Nielsen and TelevisaUnivision declined to comment.
Shifting from big data to panel alone lifted all Spanish-language networks fairly evenly, by mid-single-digit to low double-digit percentages. But in the key 18-34 demo, Univision gets two to four times the lift of other networks, one of which even showed a sharp decline in young viewers despite a gain in households, according to data obtained by Ad Age.
In his letter, Nielsen’s Doe noted the issue but acknowledged he had no explanation.
“The differences in Spanish language networks may reflect a potential panel recruitment bias for heavier tuning Spanish dominant homes (though we don’t have evidence of that),” Doe wrote. “But it may also reflect measurement bias of some form in the big data sources or the modeling techniques that we have employed. We get to the point with data evaluations where we have ruled out obvious errors and are left with differences that, in the absence of a definitive truth set, are unexplained.”
