What’s up with the sudden panel palooza?
Broadcasters have long been dissatisfied with Nielsen’s panel-based approach.
And Nielsen lost accreditation for its local and national TV ratings last year because it couldn’t service its people meters, which caused its panel to be less accurate and less representative.
But now panels are making headlines as a fresh way to measure streaming and TV consumption.
Earlier this month, the Video Advertising Bureau (VAB) said it’s building a Nielsen-size panel for cross-platform measurement. The following day, Nielsen revealed the size of its “streaming meter panel” for the first time. (For anyone who’s counting, Nielsen’s panel clocks in at 42,000.)
The Association of National Advertisers (ANA) is also building a panel of its own for cross-media measurement, and, heck, even Google’s making one, albeit for online measurement.
Don’t confuse what the ANA and VAB are building with audience panels, though. These are calibration panels.
What’s a calibration panel?
Alternative TV measurement providers, such as VideoAmp and iSpot, have made it their mission to compile large first-party data sets.
But panels can help validate census-level information using device-level data from set-top boxes and smart TVs to verify impressions and deduplicate audiences. This can help determine who’s watching and, in turn, control frequency.
Calibration panels serve the specific purpose of validating big data and identifying audiences, as opposed to an audience panel like what Nielsen offers, said Sean Cunningham, CEO of the VAB.
In a word, calibration is about “personification,” Cunningham added. “Calibration panels have a much narrower application than audience panels and can help produce more stable, identity-based measurement,” he said.
But although the VAB is creating a panel that will be similar to Nielsen size-wise, Cunningham stressed it’s not meant to compete with Nielsen’s panel.
“That’s not the idea at all,” he said. “The idea is to create a calibration panel with a narrower application that can be used as a common resource for qualified measurement companies.”
ANA vs. VAB
The ANA and the VAB are collaborating to some extent, but both are creating TV audience calibration panels of their own, and they’re taking slightly different approaches.
The VAB plans to enlist a panel provider, such as Kantar, TVision or HyphaMetrics, to get census-level data, which the VAB will then “personify” using automatic content recognition and set-top box data to “reveal who’s watching,” Cunningham said.
The ANA, meanwhile, is still busy chipping away at its Cross-Media Measurement (CMM) Initiative, which the VAB signed on to last year.
The CMM initiative is built on what the ANA calls its Virtual ID, designed to solve for cross-media reach and frequency by personifying TV household audience data, the ANA Group EVP Bill Tucker told AdExchanger.
The ANA is building and scaling its Virtual ID through proof-of-concept testing and data integrations with alt measurement currencies, such as Comscore and VideoAmp, to deduplicate and identify audiences.
It’s run three tests so far this year and is in the process of vetting panel providers, said Tucker, who noted the next step for the ANA’s initiative will be to launch a calibration panel. The ANA’s version will be a little different from the VAB’s, though.
While the VAB is focusing more on TV, Tucker said, the ANA is looking at “cross-media,” which includes over-the-top video that might not be viewed on a big screen. “And we’re focusing on use cases that require [significantly] smaller panel sizes than what the VAB is looking for,” he said.
The ANA’s panel will focus specifically on measuring deduplicated reach and frequency to “help solve for ad wastage,” Tucker said, whereas the VAB is looking at audience personification more broadly.
In other words, similar goal, slightly different approach, and both are in the early stages.
Cunningham said to expect more announcements by the end of September.
Panels: Yay or nay?
In the meantime, the industry continues to debate the value of panel-based measurement.
Identity, for example, is always a challenge. Without actual, device-level data, it’s difficult to determine who a viewer is.
But “calibration panels allow non-media data scientists to course-correct,” said Alice Sylvester, a partner at Sequent Partners. “They can bring certainty to numbers that the industry has been lacking amidst the alt currency duel.”
Panels can also help mitigate bias.
Big, census-level data sets might exclude people who aren’t accounted for in the census, potentially for immigration reasons. But compared with big data sets, it’s easier to spot – and root out – bias in a well-maintained panel, said Paul Donato, CRO of the Advertising Research Foundation, because audiences are more identifiable.
Even so, “panels aren’t perfect – they have their own biases,” Donato said.
For example, panels generally consist of device-level data from set-top boxes and smart TVs, which means households watching over-the-air broadcast TV are underrepresented.
So, nothing is perfect – not to mention creating a good panel isn’t cheap.
“We need high-quality panels, but they’re very expensive,” Sylvester said. The most valuable panels are “built precisely for an intended purpose,” she said, and “have to be micro-managed or they fall apart.”
“I’m skeptical the industry has the patience and wherewithal to create and nurture a panel of the magnitude we’re talking about,” she added.
And regardless, it could be a while.
If all those sexy alt currencies the industry is fawning over haven’t even hit scale yet, be ready to wait for any viable new panel solution.
