Why This Matters: In a dynamic media environment, Nielsen says its tried-and-true methodology is needed to make sure no group is counted out.
Facing competition from rivals and challenges from clients, Nielsen is defending its dominance in the media measurement business by insisting its methods more accurately count minorities and preserve diversity in media in a way big-data providers don’t.
Many of Nielsen’s would-be rivals rely on return-path data from set-top boxes or smart TV sets to provide viewership information from millions of households, which they claim is more accurate than Nielsen’s old-fashioned approach based on a sample panel of 40,000 Nielsen families.
Nielsen maintains that return-path data significantly undercounts minority audiences, and that discrepancy hurts not only viewers, but programmers, talent and a $70 billion advertising business that counts on having accurate information about who is seeing commercials.
“If you move to an alternative way of measurement, there’s a really good chance that you’re going to compromise diversity and inclusion both in terms of measurement, but also in terms of advertising and subsequently on screen media talent, Kelly Abcarian, senior VP for product leadership at Nielsen, said.
“We believe the industry would suffer a ripple effect and would have no methodologically sound way of identifying what diverse audiences are watching,” Abcarian said. “Marketers looking to reach those unique audiences would become hard-pressed to do so, effectively discriminating against diverse audiences and leading to less-informed programming decisions.
Nielsen looked at return path data and found that nationally, it underrepresented Asian viewers by 35%, Hispanic viewers by 31% and African-American viewers by 22%, compared to Nielsen’s national TV panel. White/non-Hispanic viewers were overrepresented by 12%.
The results varied in local markets. In Chicago, RPD data undercounted Asians and African-Americans by 27%, while over-counting HIspanics by 1% and non-Hispanic white viewers by 10%.
Nielsen began talking about the diversity measurement issue just before the CBS deal was made. Abcarian wouldn’t say if the issue came up during negotiations.
“I won’t comment specifically on the contract discussions,” she said. “But what I will tell you is that we have talked about this to all of our clients, CBS included, about the challenges of big data.”
CBS declined to comment on whether diversity was a factor in signing a new contract with Nielsen.
Adonis Hoffman, former general counsel for the American Association of Advertising Agencies and former chief of staff for the FCC, said Nielsen has good data, but has not done a great job of explaining how important it is that it accurately counts minority viewers.
“They’ve got the data about these consumers and it matters, it should increasingly matter to the marketplace,” Hoffman said.
Hoffman, who serves on Nielsen’s external advisory panel, said Nielsen was forced by government and civic leaders to focus on measuring diverse audience when it introduced Local People Meters in 2009.
“By virtue of the pressure that was brought upon it. Nielsen made a corporate decision to look at that and to go deep into those data pools,” he said. “Other companies, from what I understand, not only have they not scratched the surface, they just haven’t even acknowledged the importance of that [minority] data.”
“In any market or medium that is highly fragmented — as is modern TV — much larger samples are needed to capture behavior that is fragmented across hundreds or thousands of potential content options,” Comscore said in a statement. “Comscore TV and crossplatform measurement are based on massive data assets sourced from set-top boxes and our digital census network, calibrated with our consumer panels. Our TV services alone collect viewing behavior from more than 65 million TVs in more than 30 million homes nationwide.
“Since minority programming has, by definition, smaller audiences than general market programming, the measurement scale provided by nationwide set top box data assures that such programming is fairly measured even on the most niche of networks and stations,” Comscore said.
But other data experts said return-path data alone is not necessarily the answer. RPD data measures devices, not humans, and that creates a gap in properly identifying groups such as Spanish-speaking households and households with children, said Clypd chief research officer Pete Doe, a former Nielsen employee.
“There’s undoubtedly value in these new solutions,” Doe said. “But everybody needs to approach any data set with eyes open and understand what the pros and cons are.”
Former Turner chief research officer Howard Shimmmel, another Nielsen alum, said NIelsen had a point, but that set-top box data could be weighted to make it reflective of minority audiences.
“I think Comscore will need to have their TV data accredited at a certain point, and that will be one of the things that I’m sure the accreditation process will look at,” Shimmel said.