Michael Wolff on the Wild West of TV Metrics and the New War on Nielsen

Nielsen Wrestling Illo - H 2015
Illustration by: Kyle Hilton

Nielsen Wrestling Illo - H 2015

The archaic, deeply flawed standard bearer of audience stats in under fire as Rentrak and comScore merge. But one thing is clear: The more data, the less clear the data picture becomes.

A version of this story first appeared in the Oct. 23 issue of The Hollywood Reporter magazine. To receive the magazine, click here to subscribe.

The central media currency is not audience, but trustworthy measurement.

The television business was built on wide adoption of such a single standard. Nielsen's sampling method determined the size and basic demographics of the audience an advertiser reached and therefore what the advertiser paid. But in the competition with digital media, which uses different measurement standards, and in the shift to new screens and platforms without reliable or willing measurement protocols, measurement has come unstuck and the currency is in something of an existential freefall.

Nielsen has so far lagged in its efforts to standardize this new video marketplace. Its evermore disputed measurements helped cause this summer's deep dive in television stocks and also have been the basis of reports by media correspondents and financial analysts that television watching is imperiled.

Now, in a major, if not inevitable, challenge to Nielsen — and in an effort to put the Humpty Dumpty of standardized measurement back together again — Rentrak, the upstart movie and VOD measurement company, is merging with comScore, the leader in digital analytics.

It's an almost comical exercise to try to portray the Wild West nature of current measurement standards. Digital media's basic promise was that, with its total tracking capabilities, it would return absolute, single-user results, with something like full-knowledge of a user's behavior and level of interaction (no more unmeasured bathroom breaks). But in fact, digital has become something of the Volkswagen of media measurement with, by some estimates, advertisers paying for traffic that is 30 percent click fraud and, by a large amount, traffic that has moved off a page before an ad even loads. What's more, without a single standard, almost every website offers a unique data story, with a comScore measure almost always at odds with an internal measure, and with the vast increase in mobile behavior another variant.

A lack of trustworthy digital measurement has been in part responsible for the heavy discounting in digital ad rates. But, curiously, that same lack of standardization and, indeed, industrial-scale hyperbole and fraud, has allowed 100 million-monthly-unique-visitor sites and billion-view videos to become common, helping to fuel the digital bubble.

At the same time, television remains tethered to a Nielsen standard that, over the past 18 months, has tracked a dramatic viewer falloff in traditional scheduled viewing but has not had the wherewithal to capture the surge in equally dramatic new behavior involving delayed and over-the-top viewing.

Viacom, with its younger audience among the most hard hit by new viewing behaviors, has been clobbered by Nielsen numbers but, at the same time, in something of a reality warp, has seen little dent in the per-subscriber fees paid to it by the cable systems that carry its programming — and, also, much higher viewing in the Rentrak profile.

It's a tempest in methodology.

Nielsen, in an approach that has changed little over the years, infers its make-or-break analysis from a sampling of 20,000 homes. The television industry complains that with myriad new behaviors needing to be measured, that is not a mathematically large enough sample — and it asks, not unreasonably in this age of comprehensive date, why not 20 million homes?

And, indeed, Rentrak and comScore are out to offer a total, "census based," data profile — not a sampling, but an exact portrait of behavior on every device, browser, smartphone, set-top box, OTT outlet, smart TV, and those yet to come.

Rentrak started 30 years ago as a video rental store, but seeing the advance of Blockbuster and other chains, moved to support independent stores with a new deal: Distributors would supply videos and get paid on a per-rental basis, a transaction made possible by Rentrak-developed measurement software. Soon Rentrak was competing with Nielsen's survey-based movie measurements and pushing Nielsen out of that business. Then, through deals with a range of cable companies and set-top-box providers, Rentrak extended its measures to VOD ratings, and then to all set-top-box behavior. Now, joining with comScore, it believes it can offer something like a universal picture of who's watching what, where and when at every minute of the day.

Indeed, Rentrak data, overlaid with voter files, car ownership records, other public databases and, now, presumably, comScore's digital profiles, provide the base for the new generation of addressable television — wherein an ad can be served to a single demo- and geo-located viewer — which, by some estimate, will add $2 billion to this presidential campaign cycle's television expenditures (TV stock short-sellers take note).

So, who needs Nielsen?

Alas, the more data, the less clear the data picture seems to become — or the more strenuous the arguments about "data hygiene." And the less reason there is to trust anyone with data, and the further we move from a world ruled by genial data consensus. Hence a reason for more data, and more data methodologies, reaching a new standard not of ultimate authority, but of who is less bad than the others.

Rentrak, as Nielsen might note, failed its review earlier this year by the Media Ratings Council (the MRC pointed out that it isn't unusual for a complex system to fail its first audit). On the other hand, comScore passed its audit with flying colors and now buttresses Rentrak's cred.

Nielsen, itself trying to forge a more comprehensive measurement package in a partnership program with Facebook, will also point out that a household or device measurement, without a more qualitative survey aspect, will fail to distinguish (or at least more or less accurately predict) the critical fact of who in the house is actually using the device.

There is, too, as another variable in the coming measurement wars, TiVo — which has quietly become a sophisticated tech and research company, designed to measure as well as manage content — as player, partner or potential acquisition target (wouldn't Google want to control measurement, too?).

And then there are the larger and larger parts of the content ecosystem that simply don't want to be measured. In the Netflix, Amazon, Apple, Hulu and growing over-the-top video world that pushes beyond advertising, the currency and business model is the opposite of measurement: Keep them guessing.

Which is now pretty much the state of the art.

Michael Wolff's most recent book is Television Is the New Television.