TV measurement is hamstrung by household-level data.
Despite the scale of set-top box and smart TV data, there’s no easy way to connect ad exposure on a television to a specific person within a household.
“And if you don’t break TV data down to the person, it’s very difficult to do cross-media measurement or run any kind of real demographic analysis, said Yan Liu, CEO and co-founder of TVision, a TV data and analytics company that uses eye-tracking technology to measure attention.
On Monday, TVision (not to be confused with the recently launched T-Mobile streaming service of the same name) released a solution called Advanced Audience Projections that helps measurement companies track TV ad reach at the person-level.
Oracle’s Moat, WarnerMedia’s Xandr and iSpot are already using the tool to augment their measurement capabilities, and there are five or six other TV measurement providers in the pipeline, Lui said.
TVision maintains a panel of 5,000 homes which translates to somewhere between 13,000 and 14,000 people. With its eye-tracking technology, TVision gathers second-by-second viewership data from the living room.
“Think of it as a probabilistic model that we can apply to any big set of data, and then we give it a score,” Liu said.
Using this data, the new Advanced Audience Projections platform performs three main functions.
Companies can combine TVision viewing data and demographic propensity data to model which viewer within a house is watching TV.
Household and person-level session data can be mixed with set-top box data to gather insights about specific viewing sessions.
Lastly, advertisers can develop lookalike models to predict individual ad exposure behaviors across larger household-level data sets, such as Vizio’s smart TV data and set-top box data from DirecTV.
“Brands are looking for cross-platform reach and frequency tools to help them understand ad exposure across platforms and what is the duplication there,” Liu said. “It’s the most basic question but, unfortunately, there is no holistic product for doing that.”
Xandr is layering TVision’s person-level viewership and attention measurement data with its own household-level TV viewership assets so that it can help customers “properly reach the people in their target audience,” said Jason Burke, VP of strategy at Xandr.
“This critical enablement means marketers and media companies can realize a level of audience targeting and measurement parity with the person-level tactics of the $70 billion TV advertising and the $7 billion digital video ad markets in the US,” Burke said.
Oracle Data Cloud uses TVision data in a slightly different way, which is to understand the overlap between TV and digital campaigns and to measure incremental reach.
ODC measures OTT campaigns through existing Moat integrations, and also partners with both TVision and iSpot so that it can match a household with a digital impression and then create person-level reach and frequency tools that help customers with cross-platform measurement.
"You can finally understand across digital and TV whether ad impressions truly reached your audience or just played to empty rooms or loaded below the fold," said Kevin Whitcher, VP of new product development for Moat by Oracle Data Cloud.
The future of TV advertising, though, will be a hybrid of household-level targeting and person-level measurement, Burke said. Both tactics need to coexist to drive the most value for advertisers.
“Marketers strive to gain mindshare and drive business outcomes for their brands by reaching their target audience at a person-level vs. delivering their message to households in which those people might reside,” he said. “[But] household-level targeting within television and connected TV … remains an important element within media and advertising.”
TVision, which raised a $16 million venture round in early November, is planning to pursue accreditation for its Advanced Audience Projections solution based on the Media Rating Council’s cross-platform audience standard.