Last year, the human race “took back the web” from bots defrauding advertisers.
Desperate to hide any hint of online fraud, Self proclaimed “digital” marketers have of late appeared engulfed with the contention that TV’s ratings are inaccurate because Nielsen’s sample is too small is not valid. So if you’re one of them, here’s some clarity.
You don’t need a 100 million home panel to get a reasonably precise fix on national TV viewing preferences. So long as the sample is well designed and the methodology employs a fair process for obtaining the data, you are on safe ground.
The networks and advertisers have monitored Nielsen closely over the years and most other surveys, including some with much larger samples, confirm Nielsen’s general validity.
I, myself, have done some of these comparisons.
As for Nielsen not including all of the audience—meaning digital venues—that situation is being dealt with, however the digital component of the audience for most nationally presented TV shows is fairly small—except for some of the broadcast networks’ prime-time fare. Moreover, advertisers are not, as yet, being charged for such “unmeasured” audiences.
For decades, businesses in all areas have used sampling to avoid the outrageous costs of “every thing”. This is one of the defects of big data theory – a nearly religious belief that we miss big things if we don’t count every thing. Except, we don’t miss much. My big data experience is it finds lots of tiny things – not much big.