If someone goes to the bathroom while a beautifully shot commercial plays full-screen on their TV, was it actually viewable?
Not so much, said Luke McGuinness, president and COO of TVision Insights, a TV analytics company that helps brands measure whether people are actually paying attention to their ads.
TVision, founded in 2014, started out by measuring attention on linear TV and moved into OTT more recently. It gathers its insights through a panel of 5,000 US homes, which represents roughly 14,000 people.
“Measuring viewability in the digital context is about whether an ad shows up in the field of view so that someone has the opportunity to see it,” McGuinness said. “By and large, TV ads are showing up on the screen – but if no one is in the room to see them, they can’t have an impact.”
TVision clients include AB InBev, MARS, Microsoft, PepsiCo, Duracell, Google and Nestlé.
AdExchanger spoke with McGuinness.
AdExchanger: TVision in a nutshell is … finish that sentence.
LUKE MCGUINNESS: We’re a data and analytics company that measures how people actually watch TV. If you’re an advertiser spending $100 million a year on TV advertising and people are only in the room, say, 30% of the time and only a fraction of those people are actually paying attention – well, there’s a tremendous opportunity there for advertisers to better optimize and allocate their TV dollars.
How are you like Nielsen, and how are you not like Nielsen?
We are like Nielsen in that we use a panel methodology, which is 100% opt-in, but we are unlike Nielsen in that we are creating a very different metric, although our data sets are actually complementary to ratings. But ratings only tell part of the story.
We found that ratings do not correlate at all to whether people are paying attention. We’ve seen highly rated shows with a fairly low attention rate, and plenty of shows with a relatively small but very engaged audience and low ratings. If you’re trying to optimize your advertising schedule on TV, you need to understand both. It’s not just about the size of the audience but whether they are present and engaging with ads when they air.
How does your technology work?
Our panelists put our device in their homes next to their TV. It’s about the size of an Apple TV and it does three key things. First, we use ACR [automatic content recognition] to determine what someone is watching on the TV. Is it “The Voice,” “Stranger Things,” a specific commercial?
Then we detect how the content got to the screen, whether the person is watching through live cable, the Hulu app, a Roku device, the NBCU app on Chromecast, whatever it happens to be. And then, third and most critically, our device has a camera. The technology only processes images – no video – to determine if there is anyone in the room and, if so, who. It recognizes the specific person, associates their demographics and can tell if they’re paying attention to the TV or not.
How do you define attention?
Attention is our metric for engagement. We look to see whether the TV is on, which is akin to every other TV data provider out there. But then we also look at viewability, which means that the TV is on and there is at least one person in the room. Attention goes one step further: The TV is on, someone is in the room and they’re actually looking at the TV. We collect this second by second but we use the MRC standard for digital video, which is two continuous seconds.
Is TVision accredited by the Media Rating Council?
We may at some point pursue accreditation, although we’re not currently in the process. But we did partner with a company called Neutronian, which recently launched as a neutral auditor for data quality. They conducted a thorough audit of what we do and certified our practices as sound.
Your device relies on ACR technology. How do you deal with privacy concerns?
Our panel is 100% opt-in. Our panelists are signing nine-page contracts in order to opt in and we’re very forthright in explaining how the technology works when we recruit them. We also architected our technology so that all of the data is processed on the device in the panelist’s home. We’re not pulling any audio or video into the cloud. Only summarized data comes back to us from the device.
How can advertisers use your data?
One way is as a complement to their existing ratings and CPM data as they head into upfront planning to optimize their TV investment for people who actually see and engage with their ads.
Some use our data to see how well attention correlates with the way they measure the outcomes of their advertising. For example, we have clients that look at brand lift on a weekly basis to measure brand health and awareness, but they never had a good way of understanding what is driving increases or decreases in awareness. Turns out when people pay attention to ads, brand lift tends to go up, and if they don’t, it tends to go down. Seems logical, but they had no way to measure it.
The next logical step is to optimize for attention, for whichever networks, dayparts or specific shows their audience tends to be more engaged with. It’s about allocating dollars and negotiating more effectively.
Will we have a truly data-driven upfront season this year?
Over the last couple of years, there has been the start of an evolution toward advertisers using different data sets to measure the effectiveness of their TV buys, including the amount of attention people pay to their ads. Changes won’t happen overnight, but there has been an acceleration in the pace of change, which has to do, at least in part, with consumer adoption of streaming.
We’re seeing more flexibility and willingness on both the buy side and the sell side to work with each other and with new data sets.
This interview has been edited and condensed.