Choice has never been as abundant across the media landscape as it is today, and consumers are actively engaging with the platforms and channels that most appeal to them. The expanse of choice amplifies the industry’s need for accurate measurement—given that advertisers, publishers and agencies seek to attract, engage and measure engagement regardless of where consumption happens.
Importantly, the consumer remains central amid this boon of choice, which elevates the need for holistic and comprehensive audience measurement; measurement that needs to account for the myriad of new data sources the industry’s evolving platforms and channels have introduced. Those data sources, however, cannot accurately measure audiences by themselves, as they cannot provide a true representation of the U.S. population.
To measure the actual audience, you need real people.
For years, Nielsen’s panels have been the gold standard for television measurement, and they remain integral to providing critical insight into TV audiences that big data from set-top-boxes and smart TVs alone can’t illuminate. But there is tremendous value in these big data sets. They provide exponentially larger audience sizes than traditional panels can provide, but they lack specific audience information. To be holistic and representative, measurement needs to leverage big data in conjunction with panel data.
Importantly, set-top-box and smart TV data was not designed for measurement. For example, the return path data (RPD) from a cable or satellite box can tell you that a TV is on and when the channel is changed, but it can’t tell you who’s in the room or who’s controlling what’s on the screen. The same is true of the automatic content recognition (ACR) data that smart TVs provide. For example, a Nielsen analysis of RPD found that without correcting for when the TV is on when no one is watching, tuning minutes would be overcounted by 145% to 260%, depending on the provider.
Measurement shortcomings aside, big data has a big upside and can play a critical role in the future of audience measurement, especially as device and platform usage increase. The growing size and scale of big data is unrivaled, but it needs to be anchored with representative, person-level data to ensure holistic and accurate insight into actual audiences. For example, a recent Nielsen analysis found that RPD measurement of a primetime program overstated the total U.S. impressions by 69%. That same analysis found that ACR data understated the impressions by 12%.
Nielsen’s nationally representative panels are also critical in measuring the growth of streaming, which has grown to account for more than one-quarter of total TV usage. While streaming presents consumers with a seemingly endless array of content choice, big data lacks the ability to fully account for audiences and engagement. Big data cannot account for over-the-top streaming devices like Roku devices and Amazon Fire Sticks, and many streaming applications block ACR data transmissions while the apps are in use. That’s where partnerships with key OEMs and panel data become paramount, especially as new platforms and channels enter the market.
Insight into actual audiences requires data from real people—data that can be used in conjunction with other sources to dramatically increase sample sizes. Nielsen’s ability to identify and correct for data quality issues with its panels ensures that big data is stable, reliable and consistent for use in audience measurement. When big data is calibrated with person-level measurement, the industry realizes the full potential of set-top-box RPD and smart TV ACR data.