Internet Ratings For The Big Boys (pt 2)
We left off last time talking about Nielsen Media Research and how they use the People Meter system to report TV ratings.
The People Meter sample is about 5000 television viewing households at any given time (households rotate in and out on a certain schedule). And, it is as close to unbiased as Nielsen can get it. How do they achieve this? With the most phenomenal sales force I have ever seen in my life!
Nielsen periodically selects random street addresses where new People Meters should be installed to replace the households that are rotating out of the sample. For each of these “primary” addresses, they also select one or two “alternate” addresses. That’s so if the residents of the primary addresses refuse to be metered, Nielsen can still fill the sample using the alternates.
But, for the sample to be statistically valid, a high percentage (I think it’s 75% or 80%) of the primary addresses have to be used. Using too many alternates skews the sample.
That means that a Nielsen rep has to visit the primary addresses, convince the majority of the residents to let him take apart all of their video equipment, and install a People Meter so Nielsen can “watch” what they’re watching 24/7! Oh, and by the way, sample participants aren’t paid (that would skew the sample, remember?).
Bit of a difficult sale, wouldn’t you agree? But somehow the Nielsen reps do it and that’s why I consider them to be the greatest sales force around.
But, what does any of this have to do with Internet ratings? Everything!
Nielsen isn’t just for TV anymore. Now there’s Nielsen//Netratings and guess how they track what web sites their sample members visit? With something very much like a People Meter. And because the sample is built in a manner similar to what Nielsen does for TV ratings, it’s as close to unbiased as you’ll likely get.
You can go to the Netratings site and pick up some interesting free information. But if you want current, detailed data, you’d better be willing to pay. Maintaining the sample isn’t cheap and Nielsen is in it to make money, no bones about it.
But what exactly does Netratings track and why should we care. For the small operators (like me) the answers are “lots of stuff” and “maybe we shouldn’t.”
The Netratings service tracks page visits by it’s sample members. And it links these data with the demographics of the sample members. So Nielsen can tell you which site is most popular with 2 income households making a combined $60,000 a year who have 1 dog, 2 cats and a goldfish. With these numbers, a website owner can set advertising rates on his site -AND- attract the specific advertisers most interested in his targeted audience.
If you’re a website owner, you should know this about your visitors already (have you asked?). And, if you’re an advertiser looking to buy space on someone else’s site, they’ll already have the demographic data and traffic stats for you (especially if they bought it from Netratings).
Netratings also monitors exposures for different advertising creatives on behalf of clients. But, from what I can tell, it doesn’t track conversion ratios (how many people bought X when exposed to adversitement Y) or help clients work out the ROI of their advertising campaigns.
There is a whole lot more I could add to this topic, but this has turned into a very long article. And you may be asking why I’ve gone on for so long on about ratings. First, it’s something I know more than a little bit about (and I’m really a show-off at heart). And second, I wanted to contrast the difference between ratings derived from a carefully constructed sample (Netratings) and those derived from an arbitrary sample (Alexa).
Now that you know the limitations of Alexa and how it compares to more sophisticated ratings services, you’re in a better position to interpret the numbers you see reported by Alexa.
Now you know, and knowing is half the battle!
- Daniel Joseph Moran
