UPDATE at 3:40 pm PT: Fast nationals for Monday 12.28.2015 scheduled for release at 5:00 pm PT (8:00 pm ET), an unusually late hour for a data release. But this has been a most unusual day.
UPDATE at 2:25 pm PT: Next update at 3:30 pm PT. Around that time, Nielsen will either announce a data release schedule (some time between 4 and 6 pm PT) or they will close the lid for the day at 6 pm PT (9 pm ET) and deal with the Monday 12.28.2015 fast nationals tomorrow.
UPDATE at 2:20 2:10 pm PT: Still awaiting the next update. Refresh for latest.
UPDATE at 12:10 pm PT: Nielsen has pushed the next update on the 12.28.2015 fast nationals to 2:00 pm PT (5:00 pm ET).
UPDATE at 10:15 am PT: Nielsen continues to process the data for Monday and will provide an updated release time at 12:00 pm PT (3:00 pm ET).
UPDATE at 10:05 am PT: No data yet for Monday 12.28.2015 fast nationals. Awaiting revised release time.
Fast affiliate ratings for Monday 12.28.2015 broadcast prime will be released a few hours later than normal, currently scheduled for 10 am PT / 1 pm ET. This “planned delay” by Nielsen on the surface feels like what happens the day after a switch to Daylight Saving Time or back to Standard Time, a pause to make sure all the clocks and data in the system are properly synched (the thousands of clocks in sample homes embedded in TVs and DVRs and PeopleMeters, as well as all the computers along the way).
But this delay has nothing to do with anything as mundane as a semi-annual time change. Rather, the planned delay is designed to “allow for additional quality checks and validations due to the implementation of Nielsen sample expansion.” Normally, a sample expansion would be cause for celebration: more homes and people in the sample would mean more measured viewing and higher sample sizes, particularly important for lightly-viewed networks, dayparts or demographics. But here’s the rub: on 12.28.2015 Nielsen started including thousands of new homes but no new viewers in the national sample.
How is that possible? Homes with meters that measure only the channel tuned are being added to the national sample, but the demographic viewing within those homes is being ESTIMATED through a series of algorithms and data modeling. The national and local PeopleMeters as of 12.27.2015 remain in the sample (with human beings in these homes registering their TV and DVR viewing in real time in an actual tabulation), while that data is being supplemented by the tuning-only meters with the associated person viewing completely ascribed.
This is the biggest change in U.S. ratings methodology since the introduction of the PeopleMeter in 1987, a system that replaced the old method of people filling out paper diaries to supplement tuning-only household meters. In a perverse back to the future moment, tuning-only meters are being brought back to the national sample except the diaries are being replaced by a computer algorithm, at least for this new portion of the sample.
And yet very few in the industry (at the broadcast or cable networks, at advertising agencies, in the business/ trade press) have been talking about this sea change. Further, almost no one (either inside or outside Nielsen) truly understands how this new system will actually work. It’s really like this: The change is coming and we will figure out what happens as the results come in. It’s going to be great. Trust us.
This I know: It is going to take more than two hours to validate this change and begin to understand what we are measuring or virtually measuring. Fasten your seat belts for weeks and months (at least) of data gyrations and a lot of industry gnashing of teeth if the numbers drop dramatically. Then again, with all the data weighting and the myriad assumptions in the new algorithms, there are plenty of opportunities for the ratings to “stabilize” or “regress to the mean” (either on their own or with a guiding finger on the scale). To fully understand what is going on, the ratings really should be broken down and tracked a number of ways: the PeopleMeter ratings (the pre-12.28.2015 sample), the Expanded Tuning-Only ratings (the new portion of the sample), and the Combined ratings (the weighted combination of the two sub-groups). There is much analytic work ahead for the industry to deconstruct this shift.
An alarm is sounding, like a severe weather siren trying to shake the industry out of its slumber. We will soon see if that tornado on the horizon passes relatively harmlessly or turns the audience measurement world upside down.
###
