On the one side there’s comScore, the upstart ratings firm that’s making headway using data gleaned from cable and satellite set-top boxes. On the other is the longtime leader Nielsen, which has found it difficult to finally move on from paper diaries, but says that’s about to change. Nielsen execs promise that a new system will be in place by the fall and monthly ratings that stations can sell against will be available in the first half of next year.
Battle Over Local TV Ratings Gets Interesting
Last Thursday, I received — as reporters do — a PDF copy of a May 25 letter in which Sinclair Broadcast Group TV chief Steve Marks informs advertisers and agencies that the mega-station group will be dropping Nielsen and relying “exclusively” on rival comScore for ratings starting next year.
After verifying the letter’s authenticity, I posted a brief article and emailed it to our readers.
The letter is real. What I can’t vouch for is its intent. It could be that Marks was negotiating, trying to cut a better deal on his next Nielsen contract.
Nielsen suggested as much in its response to the letter: “Though we do not comment on contract negotiations publicly, Sinclair is a longtime and enduring client of Nielsen’s and we continue to work closely with them to meet and exceed business needs.”
Marks has not yet returned my call seeking comment.
In any event, the letter attests that Nielsen’s monopoly on measuring TV viewership — one that it has enjoyed since Arbitron abandoned the market in 1993 — is broken, that comScore is a real player with a significant and apparently growing piece of the action.
I checked in with comScore’s Steve Walsh and he told me that the company now serves more than 700 stations, including all of Nexstar’s for which comScore is the only source of numbers.
Its ratings are based on data gleaned from cable and satellite set-top boxes, which is piling up. By the first quarter of next year, it be drawing on data from some 38 million homes. “Passive and massive,” says Walsh.
The inroads comScore has made didn’t happen overnight. ComScore forerunner Rentrak began trying to crack the market in 2010. When comScore acquired Rentrak early last year, it continued pounding on the doors of broadcasters fed up with Nielsen’s prices and dithering.
‘We made the bet the industry would support a competitive environment and, indeed, they have,” Walsh says.
(Walsh wouldn’t comment on the Marks letter other than to say that Sinclair was an early adopter of the comScore/Rentrak numbers and that it has been using them along with Nielsen’s in all of its markets for the past two years.)
To the extent that Nielsen’s grip on local TV measurement market has slipped, it has mostly itself to blame.
As part of its much ballyhooed A2/M2 initiative in 2006, it promised to eliminate the discredited paper diaries and implement a 21st century all-electronic measurement system that would give clients a stable and accurate picture of what every person was watching in every home, and even scoop up out-of-home viewers.
Nielsen has fallen far short of that goal. As of today, it has replaced the diaries in only 70 of the 210 TV markets — local peoplemeters in the top 25, set meters in 31 makets (DMAs 26-56) and code readers in 14 small markets. (Editor’s note: When this story was first posted, it said that Nielsen had shed the diaries in only 39. It was a mistake not to include the 31 set-meter markets.)
To be fair, Nielsen is constrained by economics. Put another way, it’s not going to implement a system that broadcasters aren’t willing to pay for, and broadcasters have made clear they don’t want to pay much more than they are paying today.
Nielsen may no longer enjoy a monopoly, but it’s still the dominant half of a duopoly.
Coincidentally, the day before the Marks letter emerged, I visited two top Nielsen executives — Kelly Abcarian and Barbara McFarland — in their Wall Street offices to get details on Nielsen’s latest plan to trash the diaries and to prevent groups like Sinclair from straying away.
The plan sounds doable and what’s interesting about it is that it now relies primarily on set-top box or so-called return-path data just as comScore does.
Recall that in 2014 Nielsen floated a plan that called for replacing diaries in the 154 markets then without meters with low-cost code readers on a more or less one-for-one basis.
Code readers are about the size of one of those Roku streaming boxes. They sit near TV sets, listen for audio watermarks that identify the programs being watched and periodically upload the info via cell networks. All the participating homeowner has to do is plug the box into an AC outlet.
For demos, age and gender, Nielsen came up its so-called viewer assignment methodology, essentially an algorithm that Nielsen claims produces accurate demos for any given market based on the demos from its other metered national and local panels.
That methodology has been controversial from the get-go. Broadcasters have not yet been able to get their heads around the idea that the demos on their local news can be determined by numbers coming from outside their markets. It has yet to win the Media Ratings Council seal of approval.
Frankly, Abcarian and McFarland couldn’t fully explain it to me, at least not in a way I could understand. I don’t think I am alone.
Following the 2014 plan, Nielsen went ahead and installed the code readers in 14 small markets. Stations in those markets have had ratings based on the readers since last January. Two broadcasters I contacted with stations in code-reader markets seem satisfied with at least the household numbers, if not the demos.
But for reasons not fully clear, most likely economical ones, Nielsen shifted gears about a year ago, saying it would turn to set-top boxes for the basic tuning data in 140 markets without meters of some kind. To that end, it cut deals with DirecTV, Dish and Charter for their data.
So, it seems that Nielsen is simply following the course laid out by comScore.
Far from it, say Abcarian and McFarland. There is set-top box data and then there is set-top box data, they say. Before Nielsen certifies ratings based on the set-top data, it will go to great lengths to, in Abcarian’s words, “clean up” the data in ways that comScore cannot.
Set-top data has real limitations, Abcarian says. Taken alone, it can give viewing credit to the wrong shows, give credit when no viewing is actually happening (as when the set-top is on and the TV set is off) and fail to properly account for time-shifted viewing.
Abcarian also points out that only 40% to 60% of cable homes have set-top boxes capable of generating viewing data, and that there is a big difference between homes that can generate the data and those that don’t. The former, she says, “tend to be higher income, more highly educated and less Hispanic.”
Nielsen cleans up and fills gaps in the set-top data by comparing it with the data it collects from the thousands of metered homes it maintains across the country.
According to Abcarian, Nielsen has detailed knowledge of the people who live in those homes (number, age, sex, ethnicity) and the number and location of their TV sets. With such info, Nielsen can get a pretty good idea of who is watching even in homes with set-meter and code readers that don’t electronically record who is watching as the local peoplemeters do.
To fill two glaring “blind spots” in the set-top box data from the 140 markets, Nielsen will install 15,000 code readers in 7,000 homes. That works out to two or three code readers in 50 homes in each market (on average).
The blind spots are homes that rely solely on over-the-air reception as well as 280 channels in the 140 markets that are not carried on cable and satellite. Those channels include diginets and low-power stations without must-carry rights. Without the code readers, says McFarland, such channels “would be invisible.”
Abcarian points out that the extra 15,000 code readers create yet another panel of homes that Nielsen can use to clean up the set-top box data and bolster its demo algorithm.
The Nielsen execs promise that the new system will be in place and producing preliminary “impact” data by this fall and that monthly ratings that stations can sell against will be available in the first half of next year. The diaries and the four-months-a-years sweeps will be TV history.
If Nielsen’s plans have a weak spot, it is the viewer assignment methodology. As I said, a lot of broadcasters have a lot of problems with it and broadcasters in the 14 code-readers markets will tell stories about how the demos Nielsen attach to some show defy common sense.
But at least Nielsen has demos. Walsh acknowledges that demos are a work in progress, and hints that comScore may have some solution in the market next year.
So, we have a real battle going on in local TV measurement, which is a good thing for broadcasters as it keeps both combatants working on improving their services and prevents them from over-changing. Nobody likes a monopoly except the Parker Bros.