Nielson v. Google - who really knows what we watch
The Los Angeles Times reported a major "glitch" in the software used by its servers lost a significant amount of data collected for the May sweeps - the tracking which sets advertising rates for the upcoming television period. In fact, the article reports that the glitch was really the second significant failing by Nielson during the sweeps. Nielson also reportedly admitted that its users generally do not know how to use the set-top 'people meters' designed to track their viewing patters.
So why rely on Nielson for ratings? Google has already established it can predict the geographic distribution of flu by tracking its search engine data can probably predict the viewership of particular television shows based on audience reaction. There is no question that the data can be used to track very particular responses from viewers.
Of course, general viewership interest is not precisely the same as minute-by-minute viewership tracking. But then again, the 12,000 homes tracked has its own sample problems. In particular, it has long been known that the home patterns miss the university living environments for a large segment of the 18-25 demographic and may paint a distorted picture of the growing online viewership. Rather than people meters, surveys, e-mails, and other less-instrusive tools could be used by the 12,000 volunteers to gain far greater data. (But please, let's be sure those surveyed are participating volunteers.)
As a side note, television news does not even report on its viewership, with a recent study by the University of Pennsylvania's Annenberg School for Communication reported only 22 news reports related to television viewership trends in the past nine years.
Perhaps the two stories have something in common. The data regarding television trends is probably troubling to television executives. Revenue is down, but so is viewership. Perhaps its not the economy ... stupid.
Neilson's technicial glitches only remind us that television is an aging medium trying to keep to its traditions. Perhaps that's why there is such a good musuem dedicated to its history.
So why rely on Nielson for ratings? Google has already established it can predict the geographic distribution of flu by tracking its search engine data can probably predict the viewership of particular television shows based on audience reaction. There is no question that the data can be used to track very particular responses from viewers.
Of course, general viewership interest is not precisely the same as minute-by-minute viewership tracking. But then again, the 12,000 homes tracked has its own sample problems. In particular, it has long been known that the home patterns miss the university living environments for a large segment of the 18-25 demographic and may paint a distorted picture of the growing online viewership. Rather than people meters, surveys, e-mails, and other less-instrusive tools could be used by the 12,000 volunteers to gain far greater data. (But please, let's be sure those surveyed are participating volunteers.)
As a side note, television news does not even report on its viewership, with a recent study by the University of Pennsylvania's Annenberg School for Communication reported only 22 news reports related to television viewership trends in the past nine years.
Perhaps the two stories have something in common. The data regarding television trends is probably troubling to television executives. Revenue is down, but so is viewership. Perhaps its not the economy ... stupid.
Neilson's technicial glitches only remind us that television is an aging medium trying to keep to its traditions. Perhaps that's why there is such a good musuem dedicated to its history.
