Why Nielsen’s TV Panel is a Bad Joke
May 19th, 2012Nielsen TV panels cover 20,000 households in US. In a Nielsen marketing video titled “Nielsen Ratings 101: Designing the Sample”, which describes sampling for their TV panels, Arthur Nielsen Jr. says:
“I try to explain how sampling works. Next time you go to a doctor and he wants to take a little bit of blood, tell him you don’t believe in sampling, take it all.”
Sampling 101 anyone?
A blood sample drawn from a single person is homogenous, whereas TV-audience watching behavior across multiple individuals is not. Any 1 cc volume of blood represents every other cc of blood in that person. If I’m watching TV at home, I might not represent at all any other 35-year-old Caucasian male in my income bracket. There is no science or even old wives’ tale that suggests otherwise. Dropping statistical jargon in the mix doesn’t change this and it’s actually not that hard to figure all of this out just by looking at the numbers.
US households receive an average of 119 channels. There are 114 million households. 20,000 is 0.02% of that; that is. Nielsen captures data from 1 out of every 5,000 households in the US. Depending on how you look at it, an average American watches live TV for about 4 hours per day. If we assume that each channel has at least 10 programs per day, an individual has a total of 1,190 choices to make during one day, out of which the 4 hours represents a max of 10 programs—less than 1% of all available programs. So let me ask you, if there are 100 people watching TV at the same time, and you know what 10 of them are watching, does it make sense to you that you can use those 10 to guess the remaining 90? 10 out of 100 is 10%, and not the 0.02% that Nielsen’s TV panel is.
Dear Nielsen marketing team, I’m saddened by how you underestimate the intelligence of your customers. I’m also saddened by how you give the whole industry a bad name with your institutionalized snake oil.
If you’re an advertiser, I highly recommend looking to alternative methods of measuring TV audiences. Methods that are not from the 50s. Maybe something like Bluefin Labs.
It’s time to ask yourself if you want to be a follower or a leader. Using something just because it’s what has always been used is what followers do, and you can’t lead by following.
