THE DEVELOPMENTS OF THE PAST 12 to 18 months regarding commercial ratings have certainly been fascinating to follow. With continued pressure from advertisers for increased accountability, it's encouraging to see the industry getting closer to more precise and relevant data. And regardless of its quality at this early stage, the data has been laid in front of buyers and sellers who must now come to terms with it and figure out how to shape and work on its evolution. However, in the industry's mad dash to obtain average commercial ratings, I believe a crucial issue has been overlooked. Specifically, how compliant are members of Nielsen households in "punching out" on their People Meters when they leave the room during commercial breaks?
... While average commercial audiences are reported to be lower than program minutes, advertisers/agencies may find themselves still paying for some viewers who aren't actually present. That's why it's important for continued industry pressure on Nielsen for deployment of a passive meter, perhaps similar to the one being used by Arbitron in Houston and Philadelphia as well as in the test of Project Apollo (the joint venture between Arbitron & Nielsen). This meter picks up imbedded audio signals (for radio and some in-store media as well) and dispenses with button punching by respondents.
... Analyses of minute-by-minute data show that during the average commercial minute, audiences are generally 5% to 10% lower than during the average program minute. But does this reflect the reality of TV viewing? 20 years ago a company named RD Percy tested a passive heat meter for TV measurement. I recall reviewing a presentation that showed audience levels dropped significantly, 25% to 40% during some commercial breaks. That strongly suggests that Nielsen overreports audience levels during commercial breaks. It would be helpful if Arbitron had minute-by-minute data to share from Houston & Philadelphia that might confirm (or refute) that audiences indeed drop more than what the People Meter reports. [ MORE ]
Comments