Recently, Nielsen caused quite a stir in the ad industry by formally announcing plans to implement several key enhancements to its television measurement and reporting process. While these enhancements impact both its national and local measurement services, I’d like to focus on just the impending changes to the local service.
There were four major initiatives outlined in Nielsen’s announcement, with several of these plans impacting a number of existing measurement techniques or market types. The first and the most extensive of the planned enhancements is what Nielsen has termed a ratings stabilization effort. Per Nielsen’s official release, the ratings stabilization is a fairly standard statistical technique that it hopes will mitigate ratings fluctuations resulting from panel variability, rather than indicate real changes in viewer behavior. This enhancement is scheduled to be applied in both the top 25 local people meter (LPM) measured markets, and the 31 set meter markets. Since I don’t own an advanced degree in statistics, I can’t offer detailed comments on accuracy of Nielsen’s approach, but a high-level overview of the stabilization technique sounded reasonable to this non-statistician.
The second major initiative Nielsen announced was the introduction of electronic measurement to 14 additional markets currently relying on printed diaries via a new code reader technology. The code reader is similar to set meter measurement in that it passively captures and reports household TV tuning. However, the code reader employs audio-based tracking and is therefore much easier and affordable to install and maintain. This first step toward the eventual replacement of paper diary based measurement of television viewing on the local market level should be welcomed and embraced by all.
Demographics for these new code reader markets, as well as the 31 existing set meter markets will be generated by the third, and certainly the most controversial change disclosed — an initiative Nielsen has dubbed viewer assignment. This approach will use a statistical model to assign the household viewing captured by code readers and set meters to individual members within that same home. The basis of the model is actual demographic viewing drawn from people meter homes within similar regional areas and time zones.
While undoubtedly a major shift in the way Nielsen has generated its TV audience estimates, I’m not sure what all the fuss is about, or what is fueling such negative reactions by some in the industry. Similar types of modeling and data integrations are all the rage among other video audience data providers including Rentrak and Simulmedia, and different types of modeling is widely accepted within the world of digital and programmatic media as well. Therefore, to have one long-time Nielsen client characterize the move as “ill-conceived and certain to do more harm than good” in a recent MediaPost article, seems more than a bit over-reactionary. In contrast, I believe that Nielsen has been mostly transparent about its plans during their years in development, gathering input from various parties. The only major concern I have with Nielsen’s roll-out plans is that the schedule outlined for sharing the impact data (information that demonstrates the potential effect on ratings) does not provide sufficient time for evaluation and planning prior to the implementation in October 2015.
The fourth and final enhancement that Nielsen plans to introduce is a technique to deal with “unidentified audience” in LPM homes. As stated in Nielsen’s release, “Unidentified Audience occurs when the TV is on but no one checked himself or herself in as a viewer” using an LPM-assigned persons’ button. Currently, when this occurs within the LPM home, unidentified viewing is excluded from official reporting. Moving forward, Nielsen plans to use its “Viewer Assignment” model to complete the missing demographic information and include this viewing to the total in-tab reporting. The one comment I’ll make about this “enhancement” is that there should be little doubt that Nielsen’s TV station clients were the driving force behind this effort to increase the TV audience numbers being reported.
All in all, I view Nielsen’s announced enhancements to local ratings measurement and reporting in a positive light. Moving a total of 45 markets (31 set meter and 14 diary) away from diary-based measurement is certainly a step in the right direction. And while an audience modeling approach may not be perfect, it does permit improvements to take place in a more timely and more cost-efficient manner. Furthermore, throughout the development process Nielsen has received and incorporated input from all sides of the advertising industry – stations, advertisers, agencies, and various oversight groups including the MRC. And balancing the different and sometime conflicting wants, needs and requirements of these various stakeholders is no easy task. Just ask Nielsen.