Avatar
I'm Drew Breunig and I obsess about technology, media, language, and culture. I live in New York, studied anthropology, and work in advertising technology.

These are reactions to things I feel are important.

Follow me on Twitter.

How Your Nielsen Ratings Sausage Is Made 

kateoplis:

sectionfive:

According to this article, 25,000 people are being polled for the ratings. That’s 1/4 of 1% of the 100,000,000 TV Households.

That’s not an accurate sampling. It sounds like a joke. (via NewTeeVee)

So is the MPAA rating system. How many board members does it have now? 10? 12?

Listen,  TV/Movie market research has always been smoke and mirrors; there’s really no such thing as an “accurate sampling”. It has no basis in fact either for that 1% or for the extrapolation of that data into the greater audience. In fact, market research has shown that companies who don’t use market research are more successful.

Nielsen’s reasoning for not pulling data from cable boxes boils down to inaccuracy from losing the demographics, but that’s precisely what we should hope for: that one day we can all be lumped into the demographic of ¨people¨ (or more likely ¨widgets¨ or ¨units¨), instead of these made up racist, agist, and sexist categories ad agencies are so salacious for.

Transitioning away from the demographic categories which have lasted for the last half century will be a sea change for the ad industry. Beyond more appropriate (and more human) targeting, I think we’re going to discover that the idea of a type of individual is rather worthless in our digital age. The problem with targeting a type, is that often types don’t run in packs made up of entirely the same type. So then your message might resonate with one, then die.

I hope targeting will embrace the social graph. What media companies will package will be cultures, not simply people. Understanding how your message lives in that ecosystem dynamic will be key.

Nielsen, and many measurement companies, have golden handcuffs: they hesitate to change because they’re ‘metrics.’ And as long as they’re ‘metrics’ no one will question them or scratch below the surface because agencies and other businesses are dependent on the illusion of certainty they provide.

(via ericmortensen)

Source: sectionfive

98 notesShowHide

  1. thoughtsickles reblogged this from soupsoup
  2. alohanico reblogged this from soupsoup
  3. shaydonarmstrong reblogged this from soupsoup
  4. silas216 reblogged this from soupsoup
  5. pstek reblogged this from mikehudack
  6. jryu reblogged this from mikehudack and added:
    The math involved here is really interesting and non-intuitive, as evidenced in the dichotomy of the comments/notes. The...
  7. christinebeardsell reblogged this from mikehudack and added:
    Amazing. We could even take that metaphor as far as the book The Jungle!
  8. defenestratorx reblogged this from evangotlib and added:
    These comments are wildly off base. With a population of 100,000,000, you can poll fewer than 2,000 people to get...
  9. matthew said: That’s statistically more than a large enough sample size.
  10. zombiecuddle reblogged this from barthel and added:
    Pet peeve. Puce ribbon and all that. The accuracy of a sampling is based on population distribution and the size of the...
  11. deepomega reblogged this from barthel and added:
    Yeah I didn’t see any structural issues with the sampling (although I’m totally willing to believe there might be!) and...
  12. zigziggityzoo reblogged this from mikehudack and added:
    A truly random sample of 1000 is enough to get an accurate reading.
  13. barthel reblogged this from soupsoup and added:
    It’s called the Law of Large Numbers, and it’s the foundation of every nationally representative survey you’ve ever...
  14. justin reblogged this from mikehudack and added:
    Too bad their isn’t a more accurate medium where you could spend your advertising dollars :p
  15. ohhleary reblogged this from mikehudack and added:
    I’ll only agree with this if you also adhere to the opinion that every single political poll ever released is a joke....
  16. evangotlib reblogged this from mikehudack
  17. mikehudack reblogged this from soupsoup