Nielsen’s Code Reader Getting Mixed Reviews

The new audience measurement system Nielsen has devised to replace the old paper diaries is drawing complaints from affected broadcasters, even as an October implementation deadline looms and the ratings company awaits a crucial decision on accreditation.

The grumbling centers on the method Nielsen has come up with for producing demographic data in the absence of the diaries — the so-called “viewer assignment” methodology in which Nielsen extrapolates the demos based on data collected in local people meter (LPM) households in distant markets.

In an extended trial, Nielsen has been generating ratings based on the new system in an estimated 44 markets, some since late last year.  

The test markets include 14 DMAs where stations have had their audiences measured by diaries only. The other test markets — estimated by sources at around 30 (Nielsen declined to provide this information to TVNewsCheck) — are markets where ratings have been produced by a combination of diaries and passive, electronic set meters.

By now, broadcasters in all the markets have received test data — also called “impact data” — and have been crunching the numbers and comparing them to diary data produced concurrently in the same sweeps months (starting last November and including last February and May).

“A lot of stations are upset about the whole methodology and want Nielsen to keep pushing it off, at least until we understand a little more about it,” said Billy McDowell, VP of research at Raycom Media, which has stations in a number of the test markets. “I have other stations that have lower ratings across all their newscasts.”

In the 14 diary-only test markets, which range from DMA 40 to DMA 123, household viewing is being measured by a new type of passive electronic device — known as a “code reader” — that tracks viewing by “listening” for a Nielsen “audio watermark” embedded in the signals by the stations.

Since the new passive code readers are incapable of determining who in the household is watching, Nielsen uses the viewer-assignment data from LPM markets in the region. The viewer-assignment methodology is also being tested in 30 set-meter markets.

Nielsen maintains that the makeup of households in the LPM markets is sufficiently similar to that of the companion code-reader markets to make the “remote modeling” possible and valid.

But some station execs are alarmed  by what they’re seeing. “November [impact data] came back and it basically cut 20% of the market value of our ratings out of our station and flipped us with another station as far as news is concerned,” said Lawson Fox, general manager of KTVN, a CBS affiliate in Reno, Nev. (DMA 107). “So I’m asking: Why is that?

“If you look at our situation, we’re in Reno, which is the 107^th market, and the markets that are being used for the viewer-assignment ‘donor’ homes are [in] California, Oregon. Reno is pretty much a four-wheel drive, SUV, pickup truck town and, you know, San Francisco’s a Prius town — [there’s] kind of a difference there, what people watch [and] the way they live.”

As an example of how some of KTVN’s ratings have been turned upside down, Fox cites the performance of Ellen, Warner Bros. popular syndicated talk show. “In the code reader demo estimates, Ellen has half the female viewership as previously measured by the diary survey, and has a balanced demo composition of women and men,” he said.

That goes against everything he knows about the audience for afternoon television in general and Ellen in particular, he said.

“We have written Nielsen and laid out our case for Ellen and told them that in our view the viewer assignment data is bogus and for them to release it — to make it currency effective in October in the Reno market at least — is wrong.”

Fox is not alone.  “Our reaction hasn’t been positive,” said Todd Brown, VP-GM of KVVU, a Fox affiliate owned by Meredith in Las Vegas (DMA 41), one of the set-meter markets. “We’ve got some concerns about the process and the methodology that they’re using.”

Brown and managers at other Meredith stations in markets involved in the changeover were so concerned that they convened a meeting with Nielsen execs on Aug. 13 in Las Vegas.

“It was just us basically expressing our concerns and getting some answers in the process,” said Brown, who declined to go into specific details about the meeting. “What we were wanting [was a] more in-depth [explanation of] the methodology that they’re using and how they’re using it and how they’re coming up with the demos for the marketplace. That was our biggest question.”

The concerns have filtered their way up to the New York-based Media Rating Council. “There are a lot of station groups that are worried about their audience profiles,” acknowledged CEO George Ivie.

According to Ivie, the MRC is now in the midst of conducting its own comprehensive audit of Nielsen’s viewer-assignment methodology and will share the audit’s results with the ratings company by Oct. 1, which is the date Nielsen is aiming for to “commercialize” the new system.

He said the MRC will make a determination before Oct. 1 whether to accredit the new method, deny accreditation or defer a decision with suggestions for improvements.

Ivie said the MRC shares the concerns about the demo methodology. “I believe that’s the truth in certain marketplace situations, where it’s hard to model the audience, and you see changes in profiles,” he said. “I think [the variabilities] can be significant and they are concerning to us.”

The early returns on the new system are not all negative. “I’ve got stations that are excited about [the changeover] because all of their newscasts are up big and they are expecting big gains [when the system is up and running],” Raycom’s McDowell said.

John Humphries, GM of Hearst’s WGAL, an NBC affiliate in Harrisburg, Pa. (DMA 45), is among them. “There was no real displacement here,” he said. “Shares change. HUTs and PUTs change. Some of the ratings change — up a little bit here, down a little bit there,” he said. “But by and large, the market hierarchy remained the same.

“We had some gains and we had some losses in newscasts. We had some gains and some losses in some syndicated shows. But it fell within what I felt was a comfortable margin of error. I’m not saying those folks [other broadcasters] are wrong. It just didn’t seem to play out that way here.”

Part of Nielsen’s problem may be communications. Brown said issues raised by the Meredith stations have not yet been addressed. And KTVN’s Fox said Nielsen hasn’t answered his questions about Ellen.

TVNewsCheck requested an interview with a Nielsen official to address broadcasters’ concerns. This reporter also asked for some help in basic fact-checking. Both requests were denied.

Instead, a company press rep provided a prepared statement, which, not surprisingly, articulated Nielsen’s view that its viewer-assignment methodology was sound. “Viewer Assignment is a proven statistical technique that we have been fine-tuning for more than two years,” the statement said. “This methodology will bring significant improvements to our ratings service, including increasing stability in the demographic ratings.”