• by Joe Mandese , 6 hours ago

This email is how one knowledgeable source sums up one of the weirdest -- and most ironic -- of multiple "enhancements" Nielsen unveiled late last week for measuring the new TV season.

The enhancement, one of seven that will alter Nielsen's national TV advertising currency, effective Aug. 31, is a "latency adjustment" Nielsen is making to an independent source -- the Advertising Research Foundation's DASH surveys -- it agreed to use as the basis for its media universe estimates (UE).

The UEs are the basis for how Nielsen designs, recruits and models the household panel representing how Americans watch TV and how advertisers and agencies pay to buy TV advertising.

For more than a century Nielsen has estimated the size and composition of the media universe based on its own observations of U.S. households when its field force recruits and maintains people in its panel, but Nielsen's own estimates were found to be biased in several ways according to clients, as well as industry ratings watchdog, the Media Rating Council (MRC), and Nielsen agreed to begin using DASH as an independent source earlier this year to reduce that bias and make its ratings more representative of the actual universe.

It even began releasing so-called "impact data" to clients showing how its audience estimates will look like when the changes are implemented with the new TV season so they could negotiate billions of dollars in upfront advertising deals.

Then it decided the objective third-party estimates being supplied by the ARF has own bias -- latency -- because the data lagged by as much as 15 months between when the ARF completed its surveys and Nielsen was able to process them.

Nielsen determined that lag was too long to accurately estimate fast-moving consumer technologies, such as broadband adoption and TV "cord-cutting," so it came up with a solution for adjusting an objective source for its UEs using its own subjective data: what its field force observes while recruiting and maintaining households in its sample (see my source's sum-up in the image above).

"It's not ideal, but for the moment, we're okay with it," MRC CEO and Executive Director George Ivie told MediaPost in an interview explaining why the MRC has signed off on Nielsen's convoluted plan of adjusting what was supposed to be an objective source for adjusting Nielsen.

“To be honest, there’s not really immediately available current data on these subjects," Ivie explained, noting, "You can’t really find this data right now in many other places."

And since nature -- and media audience measurement -- both abhor vacuums, the MRC has blessed this and Nielsen's other 11th hour enhancements though Ivie says it continues to put pressure on Nielsen to improve the process, including accelerating the amount of time it takes to process the ARF's DASH data.

He estimates faster processing time can reduce the DASH data latency "by as much as half," but that Nielsen might still need to continue adjusting the data based on its own observations for the foreseeable future.

"The real issue," he says, "is whether there's a bias in that number. The whole reason why we asked Nielsen to implement an independent source for universe estimates is because Nielsen’s recruitment has fairly low response rates, certain types of people tend to cooperate more than others and biases creep in."

In fact, the biases aren't just in what Nielsen's field workers observe when recruiting new households, but what they "impute" when pre-designated Nielsen households refuse to join its panel.

"A lot of people will just close the door on them, in which case the [Nielsen] rep just estimates it based on what they can observe, like do they have a satellite dish on their house or is there any evidence of another type of reception capability. So there is a small portion of that data that is imputed," Ivie explains, adding that Nielsen's imputing and adjusting isn't done entirely in a vacuum, because the MRC audits Nielsen annually as part of maintaining its MRC accreditation.

As part of those annual audits the MRC and the independent certified public accountants it works with also audit a sample of households recruited by Nielsen's field force.

For its part, Nielsen says not just the MRC, but many of its top clients have also signed off on the DASH latency adjustments, as well as numerous other enhancement announced late last week and that all of it has been factored into the preview data clients received during their upfront planning and buying process, with the exception of one more as-yet-unannounced set of changes.

"Specifically, the last piece that we’re layering in – it’s going to come in about two more weeks – is just a view of The Gauge," Nielsen Senior Vice President-Product Strategy and Thought Leadership Brian Fuhrer said in an interview with MediaPost.

Fuhrer declined to elaborate on exactly what changes will be made to The Gauge over the next couple of weeks, but even those -- ironically -- will have some latency associated with them, because even though they will impact the TV viewing share data for "linear" broadcast and cable TV vs. YouTube and streaming services, they will not be made public until the second or third week of October when Nielsen finally releases its "September Gauge," the first to be recalibrated using Nielsen's Nielsen-adjusted DASH estimates.

Fuhrer was careful to point out that none of that will impact Nielsen's "currency" data that advertisers and agencies use as the basis for billions of dollars in broadcast and cable TV buys, but not YouTube and streaming services, which are not reported as part of Nielsen's Big Data plus panel currency service.

Nielsen has been careful to emphasize that it does not consider the Gauge reports to be currency data -- and even began putting a footnote on the reports explicitly stating that -- but as already noted, media audience measurement abhors vacuums, and the reality is that for many in the industry, the Gauge isn't just a "visualization" of linear TV vs. streaming shares, but the only visible manifestation of it.

Currency, for what its worth, is not a technical industry defined standard, but is in the eye of the beholder and how they use data as the basis of advertising deals or other media industry transactions. And the bottom line is that currency is not what Nielsen says is currency, or even what the MRC accredits, but is the data two sides agree to use as the basis of their deals.

And some believe that Nielsen's UE's to date have over-inflated streaming's share of viewing relative to linear TV and that Nielsen has been stalling a reboot of the Gauge data, because eventually streaming is expected to legitimately surpass linear TV's share of American viewing -- given enough time.

Whatever streaming's share proves to be when Nielsen finally releases the September Gauge -- following two more lame duck versions using its old UEs for July and August -- Fuhrer notes all of Nielsen's reporting, including both its national TV currency and the Gauge, will be based on the same UEs.

"And that’s going to be a hallelujah moment," he predicts, adding, "Everything is going to be consistent. We’ll be reflecting currency in the Gauge. We won’t be living in multiple worlds. And I think that’s going to be a big improvement."