AOL, one of a handful of ad tech players attempting to lead the charge in programmatic TV, on Monday announced it has bolstered its programmatic TV offering by expanding its data licensing agreement with FourthWall Media.
AOL began using FourthWall’s data in 2011, but asserts that the new data includes viewer information from “approximately two million U.S. households across 90 DMAs.” AOL now boasts that its TV targeting uses a more granular pool of data than it did before, and a company representative says AOL has onboarded data from a dozen data providers to accurately match an advertiser’s customer data with TV viewing data.
Both Dan Ackerman, head of programmatic TV at AOL Platforms, and Bill Feininger, president of FourthWall Media’s MassiveData division, claim that the new data FourthWall will be providing AOL is “second-by-second, raw television viewing data.”
There may not be real-time bidding (RTB) in the programmatic TV world, but this expanded partnership suggests that the transmission of data from TV sets to data management platforms is speeding up.
"Programmatic TV” providers are already claiming to adjust campaigns on the fly based on real-time analytics. Last week, for example, jewelry company Ritani noted that it used AOL’s programmatic TV platform to run a campaign over a six-month period, and that the campaign was “continually optimized.”
AOL says it feeds data into its "tRatio" algorithm, which determines where an advertiser should buy TV spots to reach certain audiences. The fresher the data that is fed into the algorithm, the more accurate the targeting will be -- at least in theory.
It’s a lot of sales speak, but the AOL-FourthWall announcement, coupled with news out from Google this morning that it's trial testing "real-time ad tracking capabilities similar to the systems it designed for online advertisements ... with Google Fiber TV subscribers," all point to one thing: The programmatic TV world, while still in its early stages, is becoming more “real-time.”
The data driven reporting and targeting will become increasingly accurate over time. As the data collection grows and behavior (how many sets tuned to what shows, when and where) patterns are revealed, the software applications will learn from the data points. Audience measurement statistics will be thus backed by actual viewing behavior data. I wonder what that will reveal about the accuracy of the statistical methods.
Henry, it is actual tuning data not actual viewing data. Yep, it is 'right' in single-person homes and then relies on 'statistical methods' for all other household types.
Of course, every one realizes that single person households are not even remotely representative of all households. Also, once again, as John noted, we're talking about set usage, not viewing. I'd really be interested to see how they massage the data to derive viewer projections from it.